D I S A S T E R R I S K M A N A G E M E N T S E R I E S N O . 3 Building Safer Cities The Future of Disaster Risk Edited by Alcira Kreimer, Margaret Arnold, and Anne Carlin The World Bank Disaster Risk Management Series Building Safer Cities: The Future of Disaster Risk Edited by Alcira Kreimer Margaret Arnold Anne Carlin The World Bank Disaster Management Facility 2003 Washington, D.C. © 2003 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. 1 2 3 4 06 05 04 03 The findings, interpretations, and conclusions expressed here are those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is copyrighted. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or inclusion in any information storage and retrieval system, without the prior written permission of the World Bank. The World Bank encourages dissemination of its work and will normally grant permission promptly. For permission to photocopy or reprint, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Photo Credits: Cover: submerged houses in Changsha, China, © Reuters NEWMEDIA INC./CORBIS; page 1: flooding after earthquake interrupts commercial activities in Turkey, © Alcira Kreimer, World Bank; page 90: flooded Manila Street, © Reuters NEWMEDIA INC./CORBIS; page 180: cleanup of church in Honduras after hurricane, © Bernard Bisson/CORBIS SYGMA; page 244: landslide destroys buildings in Venezuela, © PAHO. Library of Congress Cataloging-in-Publication Data has been applied for. ISBN 0-8213-5497-3 Contents Acknowledgments vii Abbreviations viii Contributors xi Editors' Note xiv Part I Globalization and the Economic Impacts of Disasters 1. Disasters, Vulnerability, and the Global Economy 3 Charlotte Benson and Edward Clay 2. Natural Hazard Risk and Privatization 33 Paul K. Freeman 3. Natural Disaster Risk and Cost-Benefit Analysis 45 Reinhard Mechler 4. Globalization and Natural Disasters: An Integrative Risk Management Approach 57 Torben Juul Andersen 5. Urban Disasters and Globalization 75 J. M. Albala-Bertrand 6. Interdependent Disaster Risks: The Need for Public-Private Partnerships 83 Howard Kunreuther Part II Environment, Climate Variability, and Adaptation 7. Cities and Climate Change 91 Anthony G. Bigio 8. The Resilience of Coastal Megacities to Weather-Related Hazards 101 Richard J. T. Klein, Robert J. Nicholls, and Frank Thomalla 9. Flood Management and Vulnerability of Dhaka City 121 Saleemul Huq and Mozaharul Alam iii iv Building Safer Cities: The Future of Disaster Risk 10. Flooding in the Pampean Region of Argentina: The Salado Basin 137 Hilda Herzer 11. Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 149 Hans Günter Brauch 12. Urban Land Markets and Disasters: Floods in Argentina's Cities 165 Nora Clichevsky Part III Social Vulnerability to Disaster Impacts 13. Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 181 Ben Wisner 14. Living with Risk: Toward Effective Disaster Management Training in Africa 197 Prvoslav Marjanovic and Krisno Nimpuno 15. Urban Vulnerability to Disasters in Developing Countries: Managing Risks 211 E. L. Quarantelli 16. Natural Disasters and Urban Cultural Heritage: A Reassessment 233 June Taboroff Part IV Protecting Critical Infrastructure from Disaster Impacts 17. A New Structural Approach for the Study of Domino Effects between Life Support Networks 245 Benoît Robert, Jean-Pierre Sabourin, Mathias Glaus, Frédéric Petit, Marie-Hélène Senay 18. Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 273 Lamine Mili 19. Damage to and Vulnerability of Industrial Facilities in the 1999 Kocaeli, Turkey, Earthquake 289 Mustafa Erdik and Eser Durukal 20. The Behavior of Retrofitted Buildings During Earthquakes: New Technologies 293 Mikayel Melkumyan Figures 1.1 Dominica--Annual fluctuations in agricultural, nonagricultural and total GDP, 1978­99 22 1.2 Bangladesh--real annual fluctuations in GDP, agricultural, and nonagricultural sector product, 1996­2000 23 1.3 Malawi--real annual fluctuations in GDP and agricultural, industrial, and services sector product, 1980­98 25 Contents v 2.1 Economic losses from natural catastrophes in the 20th century 34 2.2 Natural catastrophe trends in the 20th century 35 3.1 Impacts of natural disasters 45 3.2 Project analysis under risk 47 3.3 World Bank post-disaster reconstruction loans in relation to total World Bank lending, 1980­99 48 3.4 Risk management of natural disasters 49 3.5 Important indicators for ability to spread disaster risk for Honduras and the United States 52 3.6 Projection of GDP paths with and without insurance of public assets in Honduras 52 4.1 Development in reported and insured catastrophe losses, 1970­2001 (three-year moving averages) 58 4.2 The relationship between economic growth and catastrophe losses, 1990­2000 59 4.3 Economic growth and changes in tariff rates 61 4.4 Economic losses and export concentration 62 4.5 Commodity price developments, 1990­2000 65 4.6 Elements of the dynamic risk management process 66 9.1 Demarcation between Pre-Mughal and Mughal Dhaka 123 9.2 The buildup area of the Mughal capital 124 9.3 Flood and drainage infrastructure of Dhaka 128 9.4 Water level hydrographs for Turag, Tongi, Buriganga, and Balu Rivers and rainfall in Dhaka during 1998 131 9.5 Water level hydrographs for Turag, Tongi, Buriganga, and Balu Rivers and rainfall in Dhaka during 1988 132 9.6 Existing and proposed flood control and management infrastructure in Dhaka 134 11.1 Worldviews and environmental standpoints 150 11.2 Horizontal and vertical security dimensions 151 11.3 Causes and outcomes of environmental stress 151 17.1 Risk scenario: a linear process 249 17.2 Evaluation of the impacts of a natural hazard and use of the results 250 17.3 Summary of the psychological varieties of unsafe acts 251 17.4 Diagram of the characterization of a life support network 254 17.5 Consequence curve for a municipality 256 17.6 Diagram of the conditions for decreasing the efficiency of a mission 257 17.7 Vulnerability curve for a transformer station 258 17.8 Definition of links 260 17.9 Affected components as a function of the water level rise in the upstream storage basin 261 17.10 Example of a repercussion function: relationship between the water level in a storage basin and at a transformer substation 262 17.11 Schematization of a risk curve 263 18.1 Major regions and river basins in Brazil 281 Tables 2.1 Catastrophe exposure in case study countries 40 2.2 Government financing needs in case study countries 40 2.3 Resource gap in case study countries 41 vi Building Safer Cities: The Future of Disaster Risk 3.1 Qualifications to applicability of risk neutrality­theorem 51 3.2 Disaster losses and availability of resources for spreading risk for the United States and Honduras 51 3.3 Assessment of costs and benefits of insuring public assets in Honduras 53 6.1 Expected outcomes associated with investing and not investing in protection 83 6.2 Illustrative example: Expected costs associated with investing and not investing in protection 84 8.1 The world's largest cities, with projected populations in 2015 exceeding 8 million 103 8.2 Summary of the major weather-related hazards and the occurrence of subsidence during the 20th century for coastal megacities as forecast in 2015 104 8.3 Qualitative overview of direct socioeconomic impacts of weather-related hazards and climate change on a number of sectors in coastal zones 107 8.4 Generic approaches to hazard reduction based on purposeful adjustment 107 9.1 Area and population of Dhaka City, 1600­2001 126 9.2 Flood-affected people in Dhaka City by severity of the flood, 1998 129 9.3 Cost of rehabilitation and replacement of Dhaka Water Supply System (DWASA), March 1999 129 9.4 Flood characteristics of 1988 and 1998 floods in surrounding rivers of Dhaka City 133 10.1 Average rainfall for each region 138 11.1 Population growth of Mediterranean countries, 1850­2050 153 11.2 Changes in the urbanization rates of MENA countries (1950­2030) 154 11.3 Growth of urban centers in the Mediterranean, 1950­2015 (millions) 154 11.4 People reported killed by natural disasters by country, 1975 to 2001 (in thousands) 157 13.1 Megacities at Risk (UNU Study Cities in Italics) 182 13.2 Comparison of four megacities 185 13.3 Groups perceived by disaster management professionals to be highly vulnerable to disasters (Percent officials) 185 13.4 Knowledge of vulnerable groups and planning of programs to reduce vulnerability in Mexico City and Los Angeles (Percent officials) 186 13.5 Groups perceived by disaster management professionals to be highly vulnerable to disasters (Percent officials) 190 13.6 Knowledge of vulnerable groups and planning of programs to reduce vulnerability (Percent officials) 191 17.1 Cases of evaluation of consequences of natural events 250 17.2 Example of essential information relative to life support networks 267 20.1 Short form of the European Microseismic Scale EMS-98 295 20.2 Results of comparative analysis of seismic-(base)-isolated and fixed-base (conventionally designed) buildings 298 Boxes 1.1 Financial fallout from the Montserrat volcanic eruption 8 7.1 Coastal cities and small island states 92 7.2 Natural disasters: what percentage is due to climate change? 93 8.1 Responding to coastal flooding in London, United Kingdom 108 8.2 Responding to human-induced subsidence in Shanghai, China 108 Acknowledgments The papers in this volume were prepared as background Aysan, Tim Campbell, Robert Chen, William Cobbett, materials for the conference on The Future of Disaster Arnaud Guinard, Maritta Koch-Weser, Somik Lall, Risk: Building Safer Cities held at the World Bank from Rodney Lester, Ajay Mathur, Reinhard Mechler,Evavon December 4 to December 6, 2002. Additional presen- Oelreich, David Peppiatt, John Pollner, Jean-Luc Pon- tations and conference proceedings are available on celet, Christoph Pusch, John Redwood, Sergio Sal- the websites of the ProVention Consortium (http://www. daña, Rainer Steckhan, Pablo Suarez, Helena Molin proventionconsortium.org) and the Disaster Manage- Valdés, Koko Warner and Ricardo Zapata. ment Facility (http://www.worldbank.org/dmf). We We would also like to thank the staff of the Disaster would like to thank speakers and managers from the ManagementFacilityfortheirtremendouseffortsinorgan- World Bank who supported this effort: Nemat Talaat izing this conference: Jonathan Agwe, Maria Eugenia Shafik, Michael Klein, Ngozi Okonjo-Iweala, Frannie Quintero, and Zoe Trohanis. Additional advice and sup- Leautier, John Flora, Katherine Marshall and Orsalia port for preparations was provided by George Tharakan, Kalantzopoulos. We would also like to thank keynote Mirtha Araujo, and other staff of the Transport and Urban speaker Martin Palmer, Director of the Alliance for Reli- Development Department. gions and Conservation, and opening panel speakers We are grateful to the authors of the papers in this Cristobal Sequeira, First Executive Secretary for Disas- volume for their contributions and submitting to the ter Mitigation, Government of Nicaragua; and Margaret editorial changes required for publication. Shields, Chairperson, Wellington Regional Council, Support for the conference and publication was pro- New Zealand for their participation in the event. vided under the umbrella of the ProVention Consor- We are also grateful for the participation of and the tium. We would especially like to thank the World Bank, extra efforts made by World Bank staff and colleagues the United Kingdom's Department for International from other organizations working in the field of risk man- Development (DFID), and the Government of the agement who served as moderators, discussants, and Kingdom of Norway (The Royal Ministry of Foreign rapporteurs to make our conference a success: Yasemin Affairs) for their generous support. vii Abbreviations ACDS The African Centre for Disaster DWASA Dhaka Water Supply and Sewerage Studies Authority ADB Asian Development Bank ECLAC Economic Commission for Latin ADPC Asian Disaster Preparedness Center America and the Caribbean AIJ activities implemented jointly EIB European Investment Bank AIUF additional isolated upper floor ENLA Emergency Network Los Angeles ALOS Advanced Land Observing Satellite ENSO El Niño Southern Oscillation AMBA Buenos Aires metropolitan area EOS earth observation satellites ATC Applied Technology Council EPRI Electric Power Research Institute BWDB Bangladesh Water Development EUROMED European-Mediterranean Partnership Board FAO Food and Agriculture Organization CBA cost-benefit analysis of the United Nations CBO community-based organization FDI foreign direct investment CDM clean development mechanism FEMA Federal Emergency Management CENAPRED National Center for Disaster Agency (U.S.) Prevention GAGAN GPS and geo-augmented navigation CEOS Committee on Earth Observation system Satellites GATT General Agreement on Tariffs and CEPA-INDEC Comité Ejecutivo para el Estudio de Trade la Pobreza en Argentina GBM Ganges-Brahmaputr-Meghna CER carbon emission reduction GDIN Global Disaster Information Network CERT citizen emergency response training GDP gross domestic product CPACC Caribbean Planning for Adaptation GEF Global Environment Facility to Climate Change GEO geostationery-earth-orbit CRED Center for Research on the GHG greenhouse gas Epidemiology of Disasters GII global information infrastructure CRID Regional Disaster Information Center GPG Gauteng Provincial Government DMF Disaster Management Facility of the GPS global positioning systems World Bank GR greater Resistencia DERC Disaster and Emergency Reference GSF greater Santa Fe Centre (the Netherlands) IAS innovative anti-seismic DIFPP Dhaka Integrated Flood Protection ICBS International Committee of the Blue Project Shield DIMITRA International Network on Disaster ICCROM International Center for the Study of Management Training in Africa the Preservation and Conservation of DRC Disaster Research Center Monuments viii Abbreviations ix ICOMUS International Council on Monuments NAO North Atlantic Oscillation and Sites NASDA National Space Development Agency IDB Inter-American Development Bank of Japan IDNDR International Decade on Natural NASSCOM National Association of Software and Disaster Reduction Service Companies IDRC International Development Research NATO North Atlantic Treaty Organization Centre NDMC National Disaster Management IEEE Institute of Electrical and Electronics Centre (South Africa) Engineers NDPRC National Disaster Preparedness and INODEX Indian Ocean Experiment Relief Committee (Malawi) IFAD International Fund for Agricultural NERC North American Electric Reliability Development Council IFI international financial institutions NGO nongovernmental organization IFRC International Federation of the Red NOAA National Oceanic and Atmospheric Cross and Red Crescent Societies Administration IIASA International Institute for Applied OAS Organization of American States Systems Analysis OECD Organization for Economic IMF International Monetary Fund Cooperation and Development INDEC Instituto Nactional de Estadísticas y OFDA/USAID Office of Foreign Disaster Assistance Censos of the U.S. Agency for International INEGI National Institute of Statistics, Development Geography, and Information OP operational program INGC National Institute for Disaster OSCE Organization for Security and Management (Mozambique) Co-operation in Europe IOM International Organization for PCF Prototype Carbon Fund Migration PCUP Philippine Commission on Urban IPCC Intergovernmental Panel on Climate Poor Change PICCAP The Pacific Islands Climate Change IPCC CZMS Intergovernmental Panel on Climate Assistance Programme Change, Response Strategies PMU phasor measurement unit Working Group, Coastal Zone PSDS Private Sector Development Strategy Management Subgroup PSI Private Sector Development and IPEC Instituto Provincial de Estadística y Infrastructure Vice-Presidency of the Censos World Bank IRI International Research Institute for PUWC Pico Union Westgate Cluster Climate Prediction PWV Pretoria-Witwatersrand-Vereeninging ISRO Indian Space Research Organization RAJUK Capital Development Authority, JI joint implementation Dhaka JICA Japan International Cooperation RAT rational action theory Agency RMS Risk Management Solutions LDC least developed country RPG Rassemblement du Peuple Guineen LEO low-earth-orbit SADC Southern African Development LME London Metal Exchange Community MBS Montserrat Building Society SIFEM Sistema Federal de Emergencias MIC methyl isocyanate TMG Tokyo Metropolitan Government MFI micro-finance institution UBN unmet basic needs x Building Safer Cities: The Future of Disaster Risk UNCTAD United Nations Conference on Trade UNHCR United Nations High Commissioner and Development for Refugees UNCHS United Nations Human Settlements UN/ISDR United Nations International Programme Strategy for Disaster Reduction UN/DESA United Nations Department of UNOCHA United Nations Office of the Economic and Social Affairs Coordination of Humanitarian Affairs UNDP United Nations Development UNU United Nations University Programme USAID U.S. Agency for International UNDRO United Nations Disaster Relief Office Development UNEP United Nations Environment USGAO U.S. General Accounting Office Programme USGS U.S. Geological Survey UNESCAP United Nations Economic and Social USNRC U.S. National Resource Council Affairs Programme VAP Variability and Adaptation Facility UNESCO United Nations Educational, WCC World Coast Conference Scientific, and Cultural Organization WHO World Health Organization UNFCC United Nations Framework WMO World Meteorological Organization Convention on Climate Change WTO World Trade Organization UNFPA United Nations Population Fund WWF World Wildlife Fund Contributors J. M. Albala-Bertrand Hans G. Brauch Sr. Lecturer Chairman Department of Economics Peace Research and European Security Queen Mary & Westfield College Studies (AFES-PRESS) University of London Free University of Berlin Mile End Road Alte Bergsteige 47 London E1 4NS 74821 Mosbach UK Germany Mozaharul Alam Edward Clay Research Fellow Senior Research Associate Bangladesh Centre for Advanced Studies (BCAS) Overseas Development Institute House No. 23, Road No. 10 A 111 Westminster Bridge Road Dhanmondi R/A, Dhaka London, SE1 7JD Bangladesh UK Torben J. Andersen Nora Clichevsky Professor Urban Development Consultant George Mason University Soldado de la Independencia 1343, 3082 S. Woodrow Street piso 6° "A" Arlington, VA 22206, USA Buenos Aires, Argentina Charlotte Benson Eser Durukal Senior Research Associate Department of Earthquake Engineering Overseas Development Institute Kandilli Observatory and Earthquake 209 Jalan Ara Research Institute Bangsar Baru Department of Earthquake Engineering 59100 Kuala Lumpur 81220 Cengelkoy, Istanbul, Turkey Malaysia Mustafa Erdik Anthony Bigio Professor and Chairman Sr. Urban Specialist, Urban Unit Department of Earthquake Engineering The World Bank Bogazici University 1818 H Street, NW, 81220 Cengelkoy Washington DC 20433 Istanbul 81220 USA Turkey xi xii Building Safer Cities: The Future of Disaster Risk Paul Freeman Mikayel Melkumyan President President Cook Street Holdings Armenian Association for Earthquake Engineering 3402 E. Eastman Ave. 1 Lane of Nansen St., 6, Apt. 20 Denver, CO 80210 Yerevan-56, Armenia USA Lamine Mili Hilda Herzer Professor Professor Alexandria Research Institute Buenos Aires University Virginia Polytechnic Institute & State University Mendoza 3454 206 N. Washington St., Suite 400 (1430) Buenos Aires, Argentina Alexandria, VA 22314 USA Saleemul Huq Director Robert Nicholls International Institute for Environment Professor and Development Flood Hazard Research Center 3 Endsleigh Street Middlesex University London WC1H 0DD, UK Enfield, London EN3 4SF UK Richard Klein Potsdam Institute for Climate Impact Research Krisno Nimpuno P.O. Box 60 12 03 University of the Witwatersrand 14412 Potsdam, Germany P.O. Box 592 Wits 2050 Howard Kunreuther Johannesburg Professor South Africa Risk and Decision Processes Center The Wharton School Enrico Quarantelli University of Pennsylvania Emeritus Professor 1332 Steinberg Hall-Dietrich Hall Disaster Research Center Philadelphia, PA 19104 University of Delaware Newark, DE 19716 Prvoslav Marjanovic USA Professor University of the Witwatersrand Benoit Robert P.O. Box 592 Researcher Wits 2050, Johannesburg Ecole Polytechnique South Africa 2500, Chemin de Polytechnique Montreal, Canada Reinhard Mechler Research Scholar June Taboroff International Institute for Applied Cultural Resource Specialist Systems Analysis (IIASA) 526 Fulham Road Schlossplatz 1 London, SW6 5NR A-2361 Laxenburg, Austria UK Contributors xiii Frank Thomalla Ben Wisner Potsdam Institute for Climate Impact Research Research Fellow P.O. Box 601203 London School of Economics 14412 Potsdam 15 Riverfleet, Birkenhead Street Germany London WC1H 8BJ, England, UK Editors' Note Vulnerability to disaster impacts is one of the most under- Urban areas are particularly vulnerable to disruptions estimatedissuesinurbandevelopment.By2050,theworld from extreme events, especially in developing countries, population is expected to grow by 3 billion people. Most where the combination of structural poverty, decaying of this growth will take place in developing countries-- and substandard infrastructure, high population densi- andwithinthesecountries,incitiesandtowns--morethan ties, and the concentration of economic assets and com- doublingurbanpopulations.Largenumbersofpeoplewill mercial and industrial activities magnify the problem. beconcentratedinmegacitiesandonfragilelands,making reductionofvulnerabilitytodisastersinmetropolitanareas a critical challenge facing development. The Future of Disaster Risk: Building Safer Disaster impacts are increasing in severity. Annual Cities Conference direct losses for weather-related events have increased from $3.9 billion in the 1950s to $63 billion in the In order to increase the awareness of development agen- 1990s. Moreover, a number of ongoing trends have the cies regarding the urgency of addressing urban vulner- potential to cause even more severe and broader disas- ability to hazards, the World Bank's Disaster Management ter impacts than ever before. These include increased Facility and the ProVention Consortium--a coalition environmental degradation, the impacts of climate of international agencies, nongovernmental organiza- change, population growth in cities, and globalization. tions (NGOs), governments, the private sector, and Increasingly, disasters affect communities far beyond academics--hosted a conference from December 4 to the areas of geographic impact as regions are linked in December 6, 2002. new ways. During the 1999 earthquake in Turkey, numer- The conference explored a range of issues related to ous textile factories collapsed, bringing to a standstill disaster vulnerability and identified priorities for devel- Turkey's large demand for African cotton. While the Sep- opment and disaster prevention activities to ensure safer tember 11 terrorist strikes had devastating impacts in cities in the future. Papers to serve as conference back- the United States, the greatest economic and human ground materials were commissioned from experts, dis- impacts may be felt in Sub-Saharan Africa. The World aster management researchers, and development Bank estimates that the resulting global economic slow- practitioners. The papers were complemented by pre- down could kill 20,000­40,000 children, half of them sentations. Discussions revolved around a range of issues in Africa, as poverty worsens. facing urban areas, including: While industrialized countries may register higher · economic impacts and globalization; economic losses following a disaster, there are frequently · adaptation to climate extremes and climate change; systems in place to respond to the event to minimize loss · preventive strategies to reduce disaster risk; of life. Property is often covered by insurance. In devel- · social infrastructure and the vulnerability of the poor; oping countries, by contrast, disasters can cause major · social perception of risk; setbacks to economic and social development, inflict · the impacts of disasters on critical infrastructure link- massive casualties, and cause the diversion of funds ages; and, from development to emergency relief and recovery. · threats to megacities from new types of hazards. xiv Editor's Note xv Conference Volume to divert funding from development to disaster relief, stunting the country's growth even more. Furthermore, The papers in this volume are organized into four sec- Andersen argues, some countries do not take steps to tions: Globalization and the Economic Impacts of Dis- mitigate potential hazards since they expect the inter- asters; Environment, Climate Variability, and Adaptation; national community to bail them out in the event of a Social Vulnerability to Disaster Impacts; and Vulnerabil- large disaster. This bailout, however, results in a "moral ity of Critical Infrastructure to Disasters. Presentations hazard." Although helping a country following a dis- and related conference proceedings are available on the aster seems to be the right action to take, this action dis- websites of the ProVention Consortium (http://www. courages governments from adequately planning for proventionconsortium.org) and the Disaster Manage- disasters. Andersen also discusses the importance of ment Facility (http://www.worldbank.org/dmf ). There countries having diversified economic bases and expert is some divergence of views among authors selected concentrations to help withstand the common shocks for this publication, though we hope their differing view- that disasters can cause. points enrich the debate and highlight the myriad issues Additional positive and negative aspects of global- surrounding disaster management. ization on the economies and social development of developing countries are examined by José Miguel Albala- Bertrand. He suggests that globalization has given rise Globalization and Economic Impacts of Disasters to worldwide economic cycle synchronization that is linked to the performance of industrialized economies. In the first section, Charlotte Benson and Edward Clay This theory proposes that if industrialized countries are explore the relationship between integration in the global in recession, disaster recovery for developing countries economy and sensitivity to natural hazards. They take may take longer, since less assistance will be forthcoming a macroeconomic perspective illustrated by case stud- and a worldwide recession will make any sort of recov- ies on Bangladesh, Dominica, and Malawi--three coun- ery difficult. He focuses on urban disasters and argues tries with varying degrees of economic diversity and that their effect on the macroeconomy is often negligi- vulnerability to natural hazards. More generally they ble because reconstruction and business opportunities note that, with reduced barriers to international trade brought by a disaster provide opportunities and eco- and increased foreign direct investment (FDI), there has nomic stimulation. Disaster management activities, there- been a steady, accelerated movement toward globaliza- fore, should focus on communities and their resilience tion, especially since the late 1980s. They also point out since the economy will either recover on its own or be that this increased global economic activity is resulting subject to greater forces that cannot be controlled. in environmental degradation that in turn increases Several contributors to the volume address the deci- the frequency and intensity of natural disasters, making sionmaking process involved in financing and dealing their impacts more devastating. In the three case stud- with disaster risk. For example, Paul Freeman exam- ies, they explore the complex developmental, economic, ines the consideration of disaster risk in the privatiza- and societal factors that affect a country's vulnerability tion process. As part of a development agenda, countries to natural hazards. often privatize infrastructure, particularly telecommu- Torben Andersen also addresses globalization and nications, electricity distribution, and water pipeline notes that, while the frequency of disaster events has systems. As the provision of goods and services is trans- quadrupled over the past 30 years, reported economic ferred from governments to the private sector, associ- losses have increased by a factor of 2,000 to 3,000 and ated risk must also be allocated. Freeman's paper explores insurance losses have increased by a factor of 1,000. the role that privatization can play in shifting the risk of These economic losses have by far outweighed economic financing post­natural disaster reconstruction from the growth figures for the same period. Andersen notes that government to the private sector. Though governments these losses hit developing countries hardest and those have traditionally been seen as the entity best able to without post-disaster contingency plans were forced cope with risk, Freeman suggests that natural hazard xvi Building Safer Cities: The Future of Disaster Risk risk no longer be placed automatically with a govern- mitigation measures, therefore, he suggests greater public- ment and that this risk be allocated during the privati- private partnerships that encourage individuals to under- zation process. Freeman identifies several complex issues take loss reduction measures and governments to enforce surrounding risk identification and allocation and states regulations and building codes. that the risk should be placed with the entity most capa- ble to deal with it. The power of taxation has tradition- ally made governments best able to cope, but in politically Environment, Climate Variability, and Adaptation unstable countries and those subject to recurrent disas- ters, a resource gap may exist. This gap--identified by The 20th century ushered in a number of trends that studying the likelihood of an event, insurance coverage, affected the environment and altered its natural rhythms. and a country's ability to raise money through taxes-- The industrial revolution, ever-increasing technologi- may mean that a government is unable to sufficiently cal innovations, rapid urbanization resulting from the assume natural disaster risk, and losses might be more mechanization of agricultural production, and the birth efficiently handled by the private market. of megacities have put pressure on natural resources Reinhard Mechler also addresses how to account for and contributed to climate change. disaster risk when making development investment Six authors address urban vulnerability and envi- decisions. Mechler suggests that cost-benefit analysis ronmental issues through papers relating to climate (CBA), used in the economic and financial evaluation change, coastal megacities, flooding, urbanization, and of public investments, is an underutilized tool that could urban land markets. Anthony Bigio's work on climate be better used to account for disaster risk. According change promotes the idea that development-financing to Mechler, using CBA for investment and risk man- institutions such as the World Bank, which invest sig- agement projects in the context of natural disaster risk nificant amounts in urban areas affected by climate vari- improves decisionmaking and the allocation of scarce ability, should incorporate adaptation measures into resources to the most profitable undertakings. This leads project design. He notes that sea-level rise is the phe- to more careful project selection and designs that decrease nomenon exclusively linked with climate change, though vulnerability to hazards and secure project benefits. climate change also increases the risk of wildfires and Though cost-efficiency as measured by CBA should storms, impacts fisheries and agriculture upon which not be the sole criterion for assessing investment in urban areas depend, worsens air pollution, and enhances development and risk management projects, it provides urban heat islands. There are adaptation mechanisms important information for efforts aimed at reducing that include improving infrastructure and strengthen- potential economic impacts due to natural disasters, thus ing defenses, especially in response to sea-level rise, contributing to more robust economic development. though these changes may impact the environment in In a similar vein, Howard Kunreuther discusses such a way that local economic bases are changed. whether individuals and businesses have enough eco- Climate change in the context of coastal megacities nomic incentives to carry out socially appropriate levels is discussed by Richard Klein, Robert Nicholls, and of mitigation for reducing future disaster losses. He pro- Frank Thomalla, who note that much of the projected vides several scenarios of interdependent disaster risk growth in large cities is expected to take place in such illustrated by measures taken (or not taken) by fami- locations. Many of these cities have existed for centuries, lies in adjoining homes to reduce losses from an event though it was only during the 20th century that these such as an earthquake and the possibility of resulting cities expanded rapidly and began to critically impact fires, water leaks, or gas explosions that could spread. natural processes. While weather-related hazards have Kunreuther concludes that such interdependent risks always been greater for coastal locations, these haz- serve as a disincentive for undertaking mitigation meas- ards, combined with human activities and environ- ures since one or a handful of families will not invest mental degradation, lead to greater erosion, storm and in mitigation measures if their home will still be at risk wind damage, flooding, and salinization of surface waters. due to an unprotected neighbor's home. To encourage Though the threat is global, it is thought to be most Editor's Note xvii severe in South and Southeast Asia, Africa, the southern embankments, and pipes to control the flow of water. Mediterranean coasts, and to a lesser extent in East Asia. It is also trying to control the expansion of the city and The authors suggest that, to deal with climate change, has forbidden the filling in of wetlands. With an increas- desirable policy and management goals should include ing population of urban poor and unsuitable construction resiliency and adaptive capacity for weather-related haz- in floodplains, the city still faces numerous challenges. ards. This framework would have the benefits of link- The Río Salado Basin in Argentina covers half of the ing the analysis of present and future hazardous province of Buenos Aires and is subject to regular flood- conditions and enhancing the capacity for disaster pre- ing. Hilda Herzer writes that socioeconomically it is one vention and preparedness with disaster recovery. of the most important areas in Argentina. It comprises Hans Günter Brauch focuses more specifically on the 56 municipalities, and to support its growing popula- potential impacts of climate change in the Mediterranean tion a number of large public works have been carried region, an area of rapid urbanization. He notes that out, including hydraulic systems to modify the basin's disasters in the region have varying economic and social drainage. The basin's primary activities have also shifted impacts that have not been adequately addressed at a from cattle breeding to irrigated farming. As a result, regional level, since the Mediterranean encompasses flood and drought cycles now affect agricultural pro- diverse communities in Southern Europe, North Africa, duction. The farming and cattle breeding that take place and parts of the Middle East. Disasters impact each of in the basin form an important part of the provincial these areas differently, and while Southern Europe may and national economies. Therefore, the impacts of flood- be significantly affected by disasters, the preventive ing and drought cycles are not localized and affect the measures in place generally prevent extensive loss of entire country. life. By contrast, disasters in North Africa and Turkey Nora Clichevsky also looks at flooding in Argentina, often result in greater loss of life and property, which but studies the role of the state as land market regulator may be preventable. Many of these events appear to in urban areas vulnerable to flooding and the impacts of have greater impacts resulting from a combination of flooding and flood defenses on land markets. She dis- environmental degradation and climate change that cusses the competition for desirable urban space in the increases the frequency and severity of flooding, extreme country and the legal and illegal land markets that arise winter weather, and mudslides. Rapid urbanization in out of it. With the high rate of urbanization in Argentina the region also increases the potential for losses in heav- and the increase in populated land prone to flooding, this ily populated areas. is becoming an even greater issue. There is minimal reg- In addition to the rise in sea level, climate change is ulation of the legal land market in Argentina and little also thought to affect rainfall, which in turn could result control of new housing developments in areas prone to in flooding. Three authors examine flooding by study- flooding. Clichevsky points out, however, that despite ing causes in Dhaka City and the Río Salado Basin in flooding in neighborhoods of all income groups, this does Argentina, as well as the impact of flooding on urban not make a large impact on property values. Evidence of land markets in Argentina. flooding is masked to make property marketable, but Flooding in Dhaka is reviewed by Saleemul Huq the factors most affecting the value are location and neigh- and Mozaharul Alam in the context of historical processes. borhood rather than flood versus nonflood zone. Founded 400 years ago by the Mughal Emperor Jahana- gir, Dhaka is surrounded by two major rivers and has experienced flooding for years, including numerous Social Vulnerability to Disaster Impacts floods throughout the 20th century. Dhaka is now a city of more than 10 million people and recurrent flooding Of utmost concern in disaster management are the is a problem for residents. Most of the city's low-lying protection of human life and post-disaster recovery that areas and wetlands have been filled in, upsetting the allows individuals and communities to resume digni- natural water runoff process. The city has imple- fied lives. Industrialized countries often have the resources mented a flood protection program including canals, and the advance warning systems to evacuate thousands xviii Building Safer Cities: The Future of Disaster Risk of people and build disaster-resistant structures, all of It had initially been assumed that these two types of insti- which save lives. Florida, for example, can be hit by tutions could work together and would serve commu- hurricanes that claim few lives and have impacts that nities well. Wisner found it was not that easy. Though destroy only a fraction of local economic growth. Most municipalities had the mandate and some funding to developing countries are not as fortunate. Disasters assist vulnerable groups, they often lacked in-depth still claim tens of thousands of lives each year and destroy knowledge of social groups and did not have their trust. livelihoods in an instant. NGOs possessed detailed knowledge of vulnerable Katherine Marshall, World Bank Director of the Devel- groups and had their trust, but they lacked capacity or opment Dialogue on Values and Ethics, highlighted the a mandate to respond to disaster emergencies. The study importance of religion as an integral part of social infra- highlighted numerous obstacles to their collaboration structure. Conference keynote speaker Martin Palmer, and it was recommended that additional capacity build- Founder and Director of the Alliance of Religions and ing in both types of institutions continue. Conservation, explored this in detail, discussing the Enrico Quarantelli also states that risk is a socially con- unique role that faith-based organizations can play in structed concept that can vary vastly from one society implementing successful disaster risk reduction strate- to another, though he highlights the emergence of new gies. Palmer noted that the 11 major religions of the categories of vulnerables that are a direct result of urban- world control 7 percent of the earth's habitable surface ization and mobility. For example, college students and and operate 54 percent of the schools worldwide. With workers living alone or in quarters, but existing far from this wide reach, and thousands of years of experience families and traditional social support networks, are a in organizing and motivating people, religious organi- group infrequently accounted for in disaster management zations have the potential to influence how people think programs. Notions that slums and squatter communi- about risk. Palmer proposed that with the disaster ties are disorganized are also challenged, as Quarantelli management, development, and environmental con- notes that migrants to cities often live among people of servation challenges of the 21st century, these groups similar ethnic backgrounds and religious beliefs, and are should become more engaged and that the role of reli- able to organize more effectively than one may think. gion move from passive to active; to illustrate this Quarantelli also discusses the profound effects that urban- point, he cited initiatives in countries such as India and ization and the development of new technologies is Thailand that have successfully channeled traditional having on the environment, creating newer and more religious values to motivate communities to protect envi- hazardous technologies with impacts that sometimes are ronmental resources. Palmer concluded that harness- not known for years. Suggestions for dealing with such ing religious values and linking them to development risk include education programs that raise the con- and disaster risk reduction goals is an ancient yet sciousness of government officials and communities to inventivewaytoincreasecommunityinvolvement,reduce understand and mitigate risks. social vulnerability to disaster impacts, and shift per- During the conference, Africa was highlighted as a ceptions of natural disasters from fatalistic to preventive. region in serious need of disaster management initia- Ben Wisner details diversity in culture and risk per- tives. Many African countries are particularly vulnera- ception in two case studies involving four cities: ble when disasters strike urban areas because most Mexico City and Los Angeles; and Manila and Tokyo. countries have only one major city, and many of these He finds that even in cities that might share similarities, are already overburdened. Prvoslav Marjanovic and there are vast differences in perceptions of vulnerability Krisno Nimpuno submit that, while many African lead- and risk. In Mexico City, for example, squatters and chil- ers recognize that disasters pose a major obstacle to the dren were thought by other city dwellers to be the most continent's efforts to achieve sustainable development, vulnerable to disasters. In Los Angeles, by contrast, the a lack of resources and trained professionals hinders elderly and disabled were perceived to be the most managing disaster risk more effectively. Marjanovic and vulnerable. Wisner also looked at agencies providing Nimpuno state that in an attempt to address the shortage disaster assistance, mainly local governments and NGOs. of trained professionals, southern African countries have Editor's Note xix embarked on a number of training initiatives, includ- during an emergency plays an important role in reduc- ing three universities in South Africa now offering degree ing the number of casualties. programs in disaster management. South Africa also Benoit Robert and colleagues discuss critical life sup- adopted a new law in 2002, the National Disaster Man- port networks and the risks faced from various types agement Act, which highlights prevention over response, of failure including technical malfunctions, sabotage, shifting the focus of disaster management activities. and natural events. They also weigh the risk of system Cultural heritage is also an important component of failure against the level of risk acceptable to the com- social infrastructure and quality of life. June Taboroff munity served by that system. In the case of essential addresses the impact of disasters upon urban cultural services, they point out that the failure of one system heritage and cites efforts to save historic buildings and can cause several other critical systems to fail, result- precious works of art. In August 2002, flooding in East- ing in a domino effect. Realization of the interlinkages ern Europe was featured on the front pages of major and possible multisystem failures should be taken into newspapers worldwide. Highlighted were not death tolls account when identifying risks and attempting to mit- and injury statistics, but mourning for the loss of irre- igate them in the disaster management process. placeable treasures and elation at the salvation of Lamine Mili identifies similar issues with respect to others. During disasters in developing countries, cul- critical infrastructure, linkages, and possible failure. He tural heritage is often an afterthought to the emergency focuses on electricity and telecommunications systems response and rarely is it incorporated into disaster and cites examples of massive power failures and their management planning. Several international organiza- impacts in India, Brazil, North America, and Europe. tions, including UNESCO, are in the process of raising The power failures were the result of extreme events-- the profile of cultural heritage and working with coun- a heat surge in India, drought in Brazil, and severe weather tries to introduce legislation to protect it under a range in North America and Europe. He also looks at hidden ofcircumstances.Whilesomecountrieshavefewresources risks that cause system breakdowns, another factor that to devote to preserving cultural heritage, increased aware- must be accounted for in planning, since power failures ness is slowly spreading and governments and com- risk lives and negatively impact the economy. Mili empha- munities may begin to see value in finding ways to protect sizes that the implementation of fault detection, isola- cultural heritage for present and future generations. tion, restoration systems, and plans for survivability of electric power networks following major disturbances is critical to ensure continuously functioning systems. Vulnerability of Critical Infrastructure Mili also highlights advances in telecommunications and to Disasters satellite technologies already being used to monitor severe weather and cites examples of LANSAT-1 ground station Urbancommunitiesaredependentupontheinfrastructure linkages with Brazil, China, India, Iran, and Zaire that that supplies them with essential services such as clean are able to use this technology. water, waste management, electricity, transportation, and Mustafa Erdik highlights the devastating loss of life telecommunications. Basic services such as these are and property that can occur from building failure. The often the main assets of the urban poor, which assist 1999 earthquakes that struck Turkey's industrial belt them to pursue livelihoods and improve their quality of killed 18,000 and injured 50,000, mostly a result of col- life. Thus, it is essential to protect critical infrastructure lapsed buildings. Infrastructure and economic losses from failures in order to prevent families and entire com- ran into the billions of dollars. Erdik states that, munities from slipping further into poverty. though industry losses were better insured than pri- Several authors addressed the issue of critical infra- vate losses of life and property, the earthquake devas- structure, retrofitting existing infrastructure, and what tated tens of thousands of families and altered Turkey's happens when infrastructure fails. Hospitals, fire depart- industrial landscape. A significant number of skilled ments, and emergency service stations are also consid- workers were killed and many of those who survived ered essential infrastructure and their proper functioning would like to move from the area since another, possibly xx Building Safer Cities: The Future of Disaster Risk stronger, earthquake is predicted for the coming years. of developing countries to natural disasters. Natural dis- Unfortunately, the greatest lessons from Turkey's tragedy aster impacts often span geographic boundaries and revolved around the revelation of substandard build- must be understood in local, national, and global terms ing practices and corruption related to building code to ensure that appropriate disaster management pro- enforcement. While building codes had been written grams are in place to mitigate and, where possible, pre- and adherence to them could have saved lives, it is too vent major negative impacts on communities and the late for the victims of the 1999 earthquakes. environment. Armenia, situated in a seismically active zone, expe- This point was reaffirmed during presentations by rienced a similarly devastating earthquake in 1988 the concluding panelists: Maritta Koch-Weser, Presi- that killed 25,000 people and injured 15,000. Since dent, Earth 3000; Eva von Oelreich, Head of Disaster then, there have been improvements in earthquake- Preparedness and Response, International Red Cross resistant technologies, some of them developed locally. and Red Crescent Societies; Helena Molin Valdés, Inter- Mikayel Melkumyan, an Armenian researcher, devel- national Secretariat of Disaster Reduction; and, Jean Luc oped a system for retrofitting buildings using laminated Poncelet, Chief, Program on Emergency Preparedness, rubber bearings. Installation of the bearings does not Pan-American Health Organization. These panelists rec- require building evacuation and costs just 35 percent ommended areas for leadership and urged conference of traditional strengthening materials. The system has participants to rise to meet the challenge of preventing been tested over the past five years, and surveys of res- future disasters. The panelists also recommended risk idents living in retrofitted buildings have revealed that management techniques for moving forward, includ- they no longer feel minor earthquakes. ing: investing in improved data and indicators on dis- aster risk, developing community participation programs, creating new risk transfer and risk reduction mecha- The Way Forward nisms, and reinforcing partnerships among stakehold- ers to reduce communities' vulnerability to risk. In synthesizing the various research papers and discus- Parallel to these important issues, Ngozi Okonjo- sions that took place over the two days of the confer- Iweala, World Bank Vice President and Corporate Sec- ence, two main issues emerge: the urgency of addressing retary, in her opening remarks and conference discussions increasing disaster vulnerability; and the interdepend- iterated the importance of creating innovative approaches ence of systems at the global, regional, and local levels. to disaster risk management as being crucial to assist- Throughout the conference, speakers and participants ing developing countries cope with vulnerability. Sev- proposed priorities and solutions for moving forward. eral ideas were presented at the conference, including: Two common threads appear in the numerous approaches creative risk sharing and transfer mechanisms, low-cost discussed: developing innovative approaches to disaster ways of retrofitting buildings, and techniques for build- risk reduction and changing people's perception of risk. ing effective community participation programs. John Flora, World Bank Director of the Transport Other speakers discussed innovative uses of standard and Urban Development Department, and Orsalia tools, such as cost-benefit analysis, to integrate disas- Kalantzopoulos, World Bank Country Director and ter risk reduction into development planning. Regional Coordinator of Southeast Europe, noted that Along with innovation, a consensus emerged among as urban populations have continued to multiply, nat- conference participants that changing people's percep- ural disasters have become bearers of increasing misery, tion of risk is key to advancing disaster risk reduction. especially for the poor. Additional factors such as climate Frannie Leautier, World Bank Institute Vice President, change,thecreationofnewhazards,environmentaldegra- emphasized that communities must understand that dation, and rising poverty levels are contributing to the they are not helpless in the face of disasters. To empower increase in disaster risk. Furthermore, globalization-- such groups, education and training were detailed in the increasing interconnectedness of economies, cultures, conference discussions and papers as powerful tools and the environment--affects the level of vulnerability to raise awareness of the importance of preparedness Editor's Note xxi programs and natural disaster risk reduction. From and better able to protect their lives, homes, livelihoods, community-level awareness raising and involvement to and assets. We hope that, by bringing together differ- building a professional-level cadre at senior levels of ent stakeholders that do not traditionally interact on government and disaster management organizations, the topic of disaster management, this conference made learning activities were the most frequently cited solu- a contribution to advancing the agenda in disaster risk tion to creating capacity for disaster risk reduction. management and will precipitate future collaboration By applying innovative approaches to disaster risk and research among participants. By publishing this reduction and by empowering people through effec- volume, the editors hope the dialogue that was initi- tive disaster reduction strategies, communities and gov- ated at the conference is expanded and its impact ernments will be more resilient when disaster strikes broadened. Flooding after earthquake interrupts commercial activities in Turkey. PART I GLOBALIZATION AND THE ECONOMIC IMPACTS OF DISASTERS Chapter 1 Disasters, Vulnerability, and the Global Economy Charlotte Benson and Edward J. Clay Two worldwide trends in recent decades are com- environmental catastrophe, social unrest that is unprece- monly noted and sometimes linked in discussing dis- dented, economies of most countries in shambles, an asters. First, the reported global cost of natural disasters increase in poverty, hunger, landlessness, migration and has risen significantly, with a 14-fold increase between social dislocation. The experiment may now be called the 1950s and 1990s (Munich Re 1999). During the a failure."1 1990s, major natural catastrophes are reported to have But what does globalization imply for vulnerability resulted in economic losses averaging an estimated to natural hazards? Rising disaster losses have paralleled US$54 billion per annum (in 1999 prices) (Munich Re increasing globalization. But are the two trends related-- 1999). Record losses of some US$198 billion were and, if so, necessarily? Or are they coincidental but recorded in 1995, the year of the Kobe earthquake-- separate movements? And can differences in the inci- equivalent to 0.7 per cent of global gross domestic dence of occurrence and nature of natural hazards influ- product (GDP) (Munich Re 1999). ence the form and level of integration of a country into Second, there is an apparent steady movement toward the global economy? globalization, with an increasing share of economic activ- This paper seeks to explore the relationship between ity taking place across countries and regions as barriers integration in the global economy and sensitivity to to integration are reduced. Between 1987 and 1997, natural hazards--that is, to events caused by geophys- the share of international trade in total output (defined ical, hydrological, and atmospheric forces. It takes a as exports plus imports relative to GDP) rose from 27 macroeconomic perspective and draws on both the wider to 39 percent for developed countries and from 10 to literature and on evidence accumulated by the authors 17 percent for developing countries (World Bank 2000). in a series of studies of the economic impacts of natu- Global foreign direct investment (FDI) flows more than ral disasters. This research includes, most recently, an tripled between 1988 and 1998 to US$610 billion, and ongoing study on The Economic and Financial Impacts of foreign direct investment is now the largest form of pri- Natural Disasters: An Assessment of Their Effects and Options vate capital flow to developing countries (World Bank for Mitigation undertaken on behalf of the World Bank's 2000). Labor migration and financial remittances to Disaster Management Facility, with the financial support home countries have also been of increasing importance of the U.K.'s Department for International Development.2 to developing countries and poorer regions within them. The paper is organized as follows. First, definitions As the World Bank (2000: 1) comments, "globaliza- of the key concepts concerning disasters and global- tion is one of the most charged issues of the day.... ization employed in the paper are given. The next sec- Extreme opponents charge it with impoverishing the tion then considers the implications of various aspects world's poor, enriching the rich and devastating the and impacts of globalization for forms and nature of environment, while fervent supporters see it as a high- vulnerability to natural hazards. Various aspects of glob- speed elevator to universal peace and prosperity." Or, alization, covering international trade in goods and serv- in the words of the 1998 Siena Declaration, "rather than ices; international financial markets; international labor leading to economic benefits for all people, it (economic mobility; and international research and exchange of globalization) has brought the planet to the brink of information are considered. The domestic impacts of 3 4 Building Safer Cities: The Future of Disaster Risk globalization in certain specific areas--namely, rates of capital, and a reduction in economic activity, such as growth, poverty, food security, and environmental income and investment, consumption, production, and conditions--are also discussed. employment in the "real" economy. There may also be The potential impacts that risk and disasters, in turn, severe impacts in terms of financial flows, such as rev- can have on the pace and nature of globalization are enue and expenditure of public and private bodies then examined in a section on the implications of nat- (Benson and Clay 1998). The losses in stocks of capital ural hazards for globalization, focusing in particular and inventory and reductions in short-term economic on the issue of whether natural hazards can present a flows are sometimes confounded in reporting the costs fundamental obstacle to integration. of disaster impacts.3 These stock losses and short-term The next section presents evidence from three flow effects may be so extreme as to result in a modifi- countries--Dominica, Bangladesh, and Malawi-- cation in the medium- to longer-term trajectory or devel- illustrating a range of experience in terms of trends in opment path of an enterprise, region, or national economy vulnerability, forms that vulnerability can take, and the as well. role of varying external linkages and relations. The Vulnerability is the potential to suffer harm or loss in case studies also demonstrate that globalization is not terms of sensitivity, reliance, and reliability. Economic a new phenomenon; that it is possible for a country's behavior is sensitive to a disaster shock. This impact is level of integration into the global economy to decrease, reflected at a macro or sectoral level in the deviation of as well as increase, over time; and that the nature of economic aggregates from trends that were expected integration can change. The latter two factors, in turn, without taking into account the effects of this event. can have implications for an economy's sensitivity to Because economic activity is sensitive to many influ- natural hazards. ences, including other sources of shock, in practice it The paper concludes with some reflections on the can be difficult to isolate precisely the impacts of a spe- policy and research implications of the complex and cific disaster or disasters. The primary objective of our changing influences that determine an economy's sen- studies has been to seek to isolate and understand sitivity to natural hazards. these short- and long-term consequences of natural dis- The literature relating to both natural disasters and asters. Resilience is the speed of recovery in economic to globalization indicates some diversity in the use of activity, which may involve repair and replacement of basic terms. At the outset, therefore, it is useful to define lost and damaged capital. how key language is used in this paper: Disaster management literature commonly distin- A natural hazard is a geophysical, atmospheric, or guishes rapid-onset disasters, such as storm surges or hydrological event that has a potential to cause harm earthquakes, which cause immediate loss and disrup- or loss. Usually these are both uncommon and extreme tion, and slow-onset events, notably drought. In our events in terms of the range of natural phenomena empirical investigations of economic consequences, we such as rainfall, tropical storms, flooding, and so forth. have found it useful to distinguish climatic hazards Hence the need to determine risk, which is understood and related riverine and coastal hydrological hazards to be "a combination of the probability, or frequency, from geophysical hazards. of occurrence of a defined hazard and the magnitude Climate-related hazards present threats of varying of the consequences of the occurrence" (Royal Society intensity that are usually recognized at a local or national 1992: 4). level, and there is consequently some form of adapta- A natural disaster is the occurrence of an abnormal tion in terms of economic behavior and the technology or infrequent hazard that impacts vulnerable commu- in which capital--productive, housing and habitat, or nities or geographical areas, causing substantial damage, infrastructure--is embodied. The economic, and of disruption, and possible casualties, and leaving the affected course wider social, consequences of both individual communities unable to function normally. From an eco- events appear to be susceptible to investigation for most nomic perspective, a disaster implies some combina- lower- and middle-income developing countries. In con- tion of losses in terms of human, physical, and financial trast, potentially catastrophic geophysical hazards may Disasters, Vulnerability, and the Global Economy 5 be very rare in occurrence. Even in potentially high-risk under successive GATT negotiations, although negotia- geographical regions there may have been no extreme tions are currently underway in the WTO on a new agree- event in living memory or even within the historical ment on agriculture. Quotas have also remained on exports record. Consequently, such hazards pose quite different of textiles and clothing, discriminating by country. problems of risk perception and economic behavior. Indeed, both the World Bank (2002) and others are call- However, a global phenomenon--satellite television and ing for a "development round" of trade negotiations. linked media information--may be changing that as well. As a result of this broad process of liberalization as Globalization is the process through which there is well as increased FDI (see the following "External Trade" an increase in cross-border economic activities, in the section) and a relatively high rate of accumulation of form of international trade of goods and services, for- human and physical capital, many globalizing devel- eign direct investment (in turn comprising the financ- oping countries have shifted exports from agricultural ing of new investments, retained earnings of affiliates, to manufacturing products. In 1965, agricultural com- and cross-border mergers and acquisitions), capital modities accounted for about half of developing coun- market flows, and labor migration. It should be noted try exports and manufactures for only around 15 percent. that greater globalization is not necessarily synonymous By the late 1990s, around 80 percent of developing with a higher level of GDP, with increasing domestic or country exports were in the form of manufactured items, regional economic integration, or with market liberal- with agricultural products falling to around 10 percent ization, although these phenomena are commonly related. by 1998 (Martin 2001). Although there is considerable variation in the composition of exports between differ- ent developing countries, with some remaining as pri- Broader Implications of Globalization marily agricultural exporters, even many of these latter for Vulnerability countries have experienced some growth in manufac- turing exports. Exports of services from developing In this section major aspects of the globalization process countries have also increased significantly. are considered in terms of their implications for vul- Different productive activities are potentially differ- nerability to natural hazards. entially sensitive to natural hazards; thus, any change in the composition of production could be significant in terms of the level and nature of risk. Natural hazard International Trade events may reduce the availability of particular goods Reductions in trade barriers and transport and com- and services for export (either directly or via disruptions munications costs have resulted in a rapid growth in to transport and communications networks) while simul- openness since the mid-1950s, with increasing trade taneously increasing imports, to meet both disaster- in manufactures (involving more two-way trade) and a related domestic shortages and relief and rehabilitation fragmentation of the production process (Martin 2001). requirements. Ramifications throughout the economy Initially, developing countries typically liberalized trade can be significant. Depending on levels of foreign- more slowly, with a number favoring import substitu- exchange reserves and on government external borrow- tion policies instead, but since the mid-1980s devel- ing policy, a deterioration in the balance of trade could oping countries have also increasingly reduced barriers result in an increase in external borrowing, with impli- to trade, often unilaterally rather than under the aus- cations for future levels of debt servicing and, ultimately, pices of the World Trade Organization (WTO). Aver- economic growth. Any worsening of the balance-of- age tariff rates in developed countries are now low, payments position could also exert pressure on the although barriers remain in the two areas where devel- exchange rate and, thus, on international competitive- oping countries have a comparative advantage: agri- ness. There are also potential budgetary implications culture and labor-intensive manufactures (World Bank in so far as government revenue is derived from export 2002). In the case of agriculture, various exceptions and import duties and tariffs. Thus, it is important that have been made for domestic support price schemes a government be aware of the potential sensitivity of its 6 Building Safer Cities: The Future of Disaster Risk various exports to natural hazards and the possible con- and export decisions. In contrast, stockpiling agricultural sequences of any changes in both relative and absolute produce was often undertaken by public or quasi-public composition. As liberalization encourages trade, it also agencies, in part specifically to stabilize export earn- encourages shifts in the composition of an economy, ings. Governments need to recognize this change and with implications for livelihoods, their relative secu- consider whether new ways of managing balance-of- rity, and ultimately household vulnerability to natural payments risks--for example, encouraging international hazards, a theme explored in further detail below. financial risk transfer mechanisms or maintaining At first sight, diversification and the shift toward man- increased foreign exchange reserves--are required. ufacturing exports would seem a positive development The shift into manufacturing products also means from a natural hazards and balance-of-payments per- that many developing countries are now competing spective. Renewable natural resource commodities (agri- against developed countries for markets. Thus, when culture, forestry, fisheries) are often among the most disruptions to production occur--particularly where directly affected by natural hazards. The sector is par- just-in-time production practices are employed-- ticularly susceptible to climatic hazards such as droughts, contracts may be lost and future market shares lost. For excessive rainfall causing floods, and cyclones, although example, the shift from agricultural to manufacturing the extent and nature of impact depends in part on the exports and thus, at first sight, to an apparently less sen- timing of a hazard event relative to cropping cycles, and sitive form of economic activity, may not in fact have on the severity of the hazard itself. Moreover, it is often reduced the potential vulnerability of Bangladesh's export difficult to obtain insurance against crop losses. Natural earnings to natural hazards. Bangladesh faces severe hazards can also have indirect effects via their impact global competition in the export of ready-made garments. on agricultural equipment and infrastructure, such as In contrast, it was the world's primary jute producer drainage and irrigation systems, post-harvest and stor- and, as such, was a price-setter on the international market. age facilities, and boats, as well as generally on trans- Disruption to the production of ready-made garments port and marketing infrastructure. could result not only in the direct loss of export revenue Primary commodity exports, including metals, but also in the longer-term loss of markets overseas. minerals, and oil, as well as renewable natural resources, The concept of vulnerability also entails potential to are also vulnerable to commodity price shocks. Few recover. Again, in some instances agriculture can offer countries are price-setters in such markets and thus may certain advantages, as illustrated by banana cultivation experience coincidental contemporaneous fluctuations in Dominica (see below), but generally, manufacturing in international commodity prices, either offsetting or activitiescanoftenberestoredfaster.Intheeventofhazard- exacerbating balance-of-payments and inflationary related damage, however, there is a possibility that a impacts of disasters. particular productive activity will not be re-established That said, there is evidence that efforts have some- at all. Although there has been no research undertaken times been taken to dampen the impact of hazard-related in this area, it is plausible that manufacturing activi- falls in agricultural production. In Fiji, for instance, ties, which are less geographically tied than agricultural sugar reserves have been used to maintain export earn- ones, could simply be relocated elsewhere, with implied ings and prevent loss of export markets in the aftermath losses to the local economy and, where FDI is involved, of natural disasters (Benson 1997a). There is probably to the national economy. less scope for using stockpiles of manufactured items Despite these reservations, the broad shift in com- to manage risk in this way. Shifts in technology and fash- position of exports experienced by many developing ions make many manufactured items rapidly obsolete, countries in recent years is, on balance, almost certainly while modern management techniques often emphasize a positive development from the perspective of sensi- just-in-time production processes. Moreover, most man- tivity of exports to natural hazards. However, again from ufacturing production is in privately owned enterprises, a natural hazards perspective, the fiscal implications of with, by implication, little regard given to the stability of trade liberalization may be less beneficial, to the extent the broader external sector in undertaking production that liberalization reduces earnings from import duties. Disasters, Vulnerability, and the Global Economy 7 Revenue emanating from import duties is typically less horizontal but, relative to developed-country investment, sensitive to natural hazards. Import duties are also rel- much of the inflows to developing countries are verti- atively easy to collect--an important point where a cal (Shatz and Venables 2000). Both forms of FDI disaster results in administrative chaos and disruption. bring potential benefits in terms of increased supply of The precise implication of any reduction in import duties capital and access to technology, management expert- will depend on the precise structure of taxes in a coun- ise, and markets. Each can also alter the nature of sen- try, including not only the significance of import sitivity of an economy to natural hazards. duties but also the relative rates charged on different Horizontal integration, under which a firm supplies categories of imports (food, oil, inputs to industry, luxury a foreign market with its product by producing locally items, and so forth). rather than importing, implies that domestic availabil- Finally, over the over the past two to three decades, ity of a product may be reduced due to direct damage growth in various service industries linked into the to the operating plant, potentially placing additional international economy has offered another form of risk pressure on the balance of payments post-disaster. diversification as illustrated by the case of Dominica. Domestic production, rather than import, of a partic- International financial services and tourism are prob- ular item also changes the nature of demands on a coun- ably the most significant in this regard. International try's transport network; whether or not this is to the financial services can be structured in such a way that firm's advantage post-disaster remains unclear. Poten- performance is determined almost entirely by nondo- tial post-disaster slumps in an economy could also reduce mestic factors. The growth of tourism also offers some demand for a particular item, perhaps with implications opportunity to reduce an economy's overall sensitivity for demand for labor in the affected industry. to natural hazards. However, efforts are required to Vertical FDI involves shifting a stage of the produc- ensure that the transport, communications, and tourism tion process to low-cost locations, on the basis that "dif- infrastructure are hazard-proofed. Tourists them- ferent parts of the production process have different selves also need to be adequately protected in the event input requirements and, since input prices vary across of a disaster. It should also be borne in mind that demand countries, it may be profitable to split production" (Shatz is potentially highly sensitive to bad publicity. These and Venables 2000: 7). Vertical FDI offers the advantage are regionally and globally relatively footloose sectors that demand does not depend on domestic economic and so investment may cluster in perceived low-risk circumstances and thus is immune to the consequences locations. of any disaster-related slump, instead continuing to offer employment. However, it can be affected by tem- porary disruption to transportation and communica- Foreign Direct Investment tions networks. The globalization process has also involved increasing From a natural hazards perspective, both forms of flows of FDI, as already noted, in part stimulated by a FDI are also potentially significant in spreading risk, reduction in developing country restrictions on for- both from the perspective of individual producers, who eign investment (World Bank 2002). The majority of can hold assets in more than one country, and from FDI flows go from advanced industrial to advanced that of an economy, reducing relative levels of risk borne industrial countries. Advanced countries accounted domestically. Such benefits of foreign ownership were for 85.3 percent of total FDI outflows between 1993 apparent in the case of lime production in Dominica in and 1997; and for 71.5 percent of FDI inflows over the the past. Large multinational producers involved in period 1985 to 1997. However, the share of inflows to the production of primary commodities may be better developing and transition economies is increasing, jump- placed to transfer risk by taking advantage of commodity ing from 21.8 percent in 1988­92 to 39.8 percent in futures (offering the opportunity to buy and sell for- 1993­97 (Shatz and Venables 2000). ward or reserve the right to do so at a pre-agreed price) There are two basic forms of FDI: horizontal and ver- and reinsurance markets, by virtue of their greater knowl- tical. Much of the intra-industrial country investment is edge and experience.4 8 Building Safer Cities: The Future of Disaster Risk Foreign investors may also build factories and other technology, reducing the importance of geography, as buildings to companywide building standards which, well as by liberalization and privatization of public finan- where they exist, are often very high, reducing poten- cial institutions in developing countries. tial physical damage as a consequence of natural haz- From a natural hazard perspective, such instruments ards. This is not always the case, however. In Bangladesh, offer certain advantages. First, firms and households for instance, inward investment in garment manufac- may be able to smooth consumption and investment ture seeking low-cost sourcing that exploits potentially while meeting rehabilitation costs as they arise. Inter- temporary tariff loopholes may be associated with low- national banking also enables individuals to hold specification, poor safety designs in high-risk locations funds with institutions better able to diversify risks. An (see the following section on Bangladesh). extreme example of the need to diversify is the case of In summary, globalization in the form of increased the Montserrat Building Society during a volcanic emer- FDI flows will alter the nature of risk. The nature of gency (box 1.1). this change will depend on individual circumstances Increasing international financial integration could but, on balance, in many cases will probably play a also offer a future mechanism for the spread of risk by role in reducing broader economic sensitivity to natu- microfinance institutions (MFIs). MFIs provide finan- ral hazards. cial services to the poor, extending credit and provid- ing savings facilities. The loans they provide are typically very small, are mainly intended for productive purposes, International Financial Markets do not require conventional forms of collateral, and Financial globalization entails the integration of a coun- are extended on a nonprofit-making basis. MFIs are try's local financial system with international financial highly vulnerable to natural hazards because of tem- markets and institutions. It involves an increase in cross- porary liquidity difficulties as they try to support clients country capital movement, including the participation through difficult periods while also experiencing a tem- of local borrowers and lenders in international mar- porary drop in flows of debt repayments. Some MFIs kets and in widespread use of international financial are therefore beginning to explore options for disaster intermediaries (in part via their presence, largely in the insurance to protect themselves and enable themselves form of foreign banks, in local markets as well as in the to respond to the additional disaster-related needs of use of those located overseas) (Schmukler and Zoido- their clients. To date, the MFIs that have established Lobatón 2001). The process of financial globalization such schemes have basically opted for self-insurance, has been significantly aided by gains in information setting some resources aside into a calamity fund for Box 1.1 Financial fallout from the Montserrat volcanic eruption The volcanic eruption in Montserrat, which began in mid-1995, resulted in the displacement of 90 percent of the res- idents from their homes, with more than half eventually leaving the island. One of the financial casualties was the Montserrat Building Society (MBS), the country's only building society, which effectively collapsed. The MBS is largely dedicated to using savings to finance housing. The MBS estimated that, prior to 1995, it had accounted for approxi- mately 90 percent of mortgages on the island as well as for a high proportion of savings by residents and some non- resident migrants. However, following an escalation of the crisis in August 1997, most insurance policies were suddenly canceled by international companies that could easily give up business on an island that was a marginal part of their portfolio. The mortgaged assets held by the MBS immediately assumed a zero value, putting the Society into substantial deficit. Although the MBS remained open, following a temporary three-week closure, depositors were ini- tially only able to withdraw up to 35 percent of their savings while the Society remained in deficit. In early 1999, the MBS announced that savers could withdraw a further 35 percent of their savings. The contrasting behavior of inter- national insurers and a local financial institution illustrates the ambiguities of globalization that can alter but not nec- essarily reduce disaster risks (Clay and others 1999). Disasters, Vulnerability, and the Global Economy 9 use in the event of an emergency. In the event of a dis- developing countries. Although there is thus scope for aster seriously affecting a significant proportion of clients, benefits of greater financial integration to be reaped, however, such funds would be grossly inadequate. The there are also a number of practical obstacles that need alternative--placing the risk externally--would create to be overcome before coverage can be increased signif- additional overheads, making the cost of credit itself icantly. There is a need to reform the structure and legal more expensive. Instead, the solution could lie in and regulatory framework of the insurance industry in some sort of international syndicate of MFIs. Good prac- a number of countries, including removal of barriers to tice dictates that MFIs should not encourage a culture entry. The cost of insurance also needs to be affordable of default and that, instead, borrowers should ultimately and stable. At the same time, insurers need to remain repay any loans. Assuming this occurs and that default-- sufficiently capitalized to bear any losses, in turn requir- as opposed to deferment--rates are low (as evidenced ing detailed scientific information on current and future in, for instance, Bangladesh and Dominica), MFIs could risks. benefit significantly from temporary access to additional Despite the various potential benefits of financial inte- resources to smooth fluctuations in demand relative to gration from a natural hazard perspective, as discussed the availability of funds. Such resources could be pro- above, it should also be remembered that such inte- vided by other, unaffected, syndicate members. gration carries other, more general, risks. Although the Globalization has also brought with it increasing pos- World Bank generally favors greater openness to trade sibilities for the use of traditional and newer forms of and FDI because of its net beneficial implications for financial risk transfer. More traditional tools comprise economic development and poverty reduction, it is "more insurance and reinsurance. Newer instruments, devel- cautious about liberalization of other financial or capital oped over the past five years in response to dramatic market flows" (World Bank 2000: 2). As Schmukler and increases in more traditional ones, entail some form of Zoido-Lobatón (2001: 3) observe, "international market hedging transaction in capital markets. Weather deriv- imperfections, such as herding, panics and boom-bust atives involve automatic and immediate payouts (typ- cycles, and the fluctuating nature of capital flows can ically available within 72 hours) upon the occurrence lead to crises and contagion, even in countries with good of a predetermined trigger event, irrespective of the scale economic fundamentals." Banks and financial institu- or nature of damage. Catastrophe bonds provide attrac- tions can spread a crisis across countries, as demon- tive payments to investors unless the specified cata- strated by the emerging-market crises in East Asia and strophic event involves a reduction, and in some cases elsewhere in 1997­98. Natural hazards themselves could cancellation, of the principal and/or interest on a bond. even trigger such crises. The city of Tokyo, for instance, The potential advantages of these various mechanisms lies in a seismically active area. It experienced a major include the alleviation of post-disaster pressure on fiscal earthquake in 1923 and volcanologists warn that another and external balances; increased government control over major event is "long overdue." As early as 1995, finan- the financing of disasters, possibly including the imme- cial analysts were already forecasting that the next major diate and timely availability of funds; increased capacity Tokyo earthquake could result in bond and stock market for the relevant government to set its own priorities in crashes in the United States and a world recession, as the management of relief and rehabilitation; increased well as severe domestic economic difficulties (Hadfield transparency in the delivery of relief and reconstruction; 1995). There is clearly a need to balance risks from dif- and provision of a tool for promoting mitigation. ferent sources and, where possible, to seek to reduce In developed countries there are already well- them. The World Bank (2002), for instance, calls for established markets for insurance against a wide range building up supportive domestic institutions and poli- of natural hazards, including earthquakes, volcanic erup- cies to reduce the risks of a financial crisis before becom- tions, floods, droughts, and cyclones. Newer hedging ing involved. instruments are also gaining some popularity. However, Finally, as with FDI, private capital does not flow to insurance and capital market instruments have played all countries equally. Indeed, the share of flows to low- a relatively small role to date in the transfer of risk in and middle-income countries (excluding the top 12) 10 Building Safer Cities: The Future of Disaster Risk has increased over time (Schmukler and Zoido-Lobatón employment in certain southern African countries), 2001), implying that many hazard-prone developing which is relatively insensitive to water shortages. countries have yet to benefit from potential risk- spreading tools available via financial integration. Economic Growth Many of the countries that have grown fastest in recent Labor Mobility decades have also increased their participation in world Increased labor mobility, the third aspect of globaliza- trade most rapidly (e.g., Dollar and Kraay 2000; Martin tion, allows affected people a radical and socially ambigu- 2001). Although the direction of causality has yet to be ous way of coping with disasters. Mobility provides established, developing countries included in the latest a potential mechanism for spreading risk geographi- round of globalization, begun in the early 1980s, are expe- cally via the transfer of remittances across borders. As riencing rapid rates of growth and catching up with more the World Bank (2002:11) states, "geographic factors developed countries; this mirrors patterns of convergence make it unlikely that capital flows and trade will elim- between OECD countries during earlier waves of global- inate the economic rationale for migration. Too many ization (World Bank 2002). This pattern basically reflects parts of the developing world have poor institutions and improved resource allocation, in part driven by increased infrastructure that will not attract production; at the competition as well as the removal of distortive tariffs and same time, some of the existing production networks other barriers to trade that protect domestic production, in the North are too deeply rooted to move." Thus, labor and improved access to markets, with markets in turn mobility looks set to remain as a potentially significant expanding further as per capita incomes rise. way of reducing sensitivity to natural hazards. How- Economic growth is not necessarily synonymous with ever, there are potential costs in terms of loss of skills broader socioeconomic development, but higher per to the economy. capita countries also tend to be among those countries In the case of Bangladesh, for instance, flows of exter- classified as more developed. Certain broad general- nal remittances provide a significant source of foreign izations can, in turn, be made about the sensitivity of exchange and have played an important role post-dis- economies at different stages of development--as defined aster. A relaxation of restrictions on out-migration, in terms of complexity of intersectoral linkages, levels including professionals such as doctors in government of physical and human capital, the scale of secondary hospitals and medical colleges, was one of the meas- and tertiary sectors, and so forth--to natural hazards. ures adopted in Bangladesh in response to the economic In its earlier stages, development tends to alter, rather crisis associated with the 1974 floods and famine. Evi- than reduce, vulnerability. Socioeconomic change asso- dence from the 1998 flood again suggests that remit- ciated with development can lead to the breakdown of tances can increase sharply during times of crisis, traditional familial support, declines in traditional ways rising by 11.9 percent (in U.S. dollar terms) year-on- of life and associated coping measures, and the increased year in 1998­99 to US$1.7 billion. Most migration is occupation of more hazardous land, a process in part asso- temporary, with migrants eventually expecting to return ciated with urbanization. The increased provision of infra- to Bangladesh (Ahmed and Chowdhury 1998), imply- structure and services can also alter, even increase, ing that family ties are strong. vulnerability. The attempt to foster rapid growth may be The implications of migration for broad sensitivity to reflected in standards of construction unable to withstand natural hazards are extremely complex in Sub-Saharan extreme conditions. This appears to have happened in Africa, however, to the extent that migration is often to Dominica in the 20 years prior to independence. Simi- neighboring countries that may be simultaneously affected larly, private sector investment in conditions of rapid tech- by drought, a problem of co-variant risk. In such circum- nicalandmarketchangeoftensacrificessafetyanddurability stances, the impact depends on the nature of employment to short-term profitability. These are conditions in which of migrants--for instance, agriculture, which is highly there may be increased vulnerability to hazards, especially weather-sensitive, or mining (a major source of migrant those regarded as extremely unlikely to occur. Disasters, Vulnerability, and the Global Economy 11 At a macroeconomic level, greater domestic integra- Nevertheless, the typology serves as a reminder that tion increases the multiplier effects of adverse perform- economic development and growth are not necessarily ance in a particular sector or regional economy. For beneficial from a natural hazards perspective. Instead, example, droughts, floods, or hurricanes may impact natural hazards need to be taken into account in the the (larger) manufacturing as well as the agricultural and determination of priorities, policies, and strategies, includ- livestock sectors, particularly where initial growth of the ing those relating to integration into the global economy. manufacturing sector is based primarily around agro- processing. A notable exception is found in dual economies with largely self-contained extractive sectors that may Information Gathering and Exchange be relatively insensitive at a macroeconomic level to cli- matic shocks. Examples are Botswana and Namibia. Provision of various regional and global public goods-- As a country begins to develop, the structure of the that is, goods and services that are nonrival in con- financial sector is also likely to be more important in sumption (users do not reduce the supply available to shaping the impact of a natural disaster. Intermediate others) and nonexcludable--can clearly benefit from economies typically have more developed economywide improved transnational cooperation and integration. financial systems for the flow of funds, including small- From the perspective of natural hazards, the greatest scale private savings and transfers, which also diffuse benefit has almost certainly been felt in the area of sci- impacts more widely. For example, in Zimbabwe fol- entific monitoring and forecasting. This is most evi- lowing the 1991­92 drought, the transfer of remittances dent within meteorology and climatology (e.g. Lee and from urban to rural regions was facilitated by the well- Davis 1998; IRI 2001) articulated system for small savings. These transfers not There is a growth in regional and international coop- only mitigated the impact of the drought in rural areas eration in climatic forecasting for the three major cli- but also spread the effects more widely (Hicks 1993). matic regions in Sub-Saharan Africa, for instance. This In the later stages of development, evidence suggests cooperation links into and has been considerably that the relative scale of the economic impacts of disas- strengthened by research and monitoring of global cli- ters is likely to decline again. In part, this reflects the matic processes such as the El Niño Southern Oscilla- smaller role of the particularly hazard-vulnerable agri- tion phenomenon by international and industrialized cultural sector in GDP, as a source of employment, a country institutions such as WMO and NOAA, which source of inputs to other sectors, and an end user. have global monitoring networks and can draw on all Other factors also contribute to reduced sensitivity, includ- the power of remote sensing technologies. ing typically higher investment in structural mitigation Regional cooperation on water resources, which relies and proofing measures, generally higher building stan- more directly on the political cooperation of upper and dards and maintenance practices, greater use of finan- lower riperian states without global partners, is less cial risk transfer mechanisms (see the FDI section below), advanced. Recent disasters such as the devastating extreme fewer foreign exchange constraints, improved environ- 2000 floods in Mozambique and in southwestern mental management, and lower levels of poverty. Bangladesh have highlighted the considerable scope for This framework for relating vulnerability to natural progress on system modeling and flood forecasting.5 hazards to the growing complexity of the economy is a Another example of international cooperation is the very broad brush. As the three country cases presented global volcanology community. This is very close-knit, later suggest, a wide variety of other factors also deter- with a small team of international experts providing serv- mines sensitivity. For example, prevailing domestic ices around the world. The creation of this informal macroeconomic and sectoral policies, deliberate changes grouping has been greatly facilitated by improved com- in policy as a consequence of a disaster, the external munications and transportation. There are also major policy environment, contemporaneous fluctuations in research benefits in the sense that the close cooperation primary export and import prices, and the timing and has helped facilitate the building of a consolidated nature of other adverse shocks can all be significant. body of evidence from volcanoes around the world. 12 Building Safer Cities: The Future of Disaster Risk There are other areas where, from a natural hazard The Government of Bangladesh, for instance, identi- perspective, information gathering and exchange are fies natural hazards as one of the factors eroding the also advantageous; these include crop research, the income of the poor via crisis-related expenditure and development of building codes, the development of reductions in income-earning capabilities. Furthermore, strategies to control pollution, and the development of it recognizes that poverty alleviation cannot be achieved mechanisms for protecting the environment. For instance, simply by increasing income, but instead requires a range in the case of the latter, as the World Bank (2002:17) of other measures, including the strengthening of local notes, "some environmental issues, such as global warm- capacity to protect the poor against shocks (GoB 2002). ing, are intrinsically global. They require international Obviously, to the extent that globalization and related cooperation, and the habit of such cooperation is economic growth reduce poverty, they may help reduce easier in an integrated world." vulnerability. Globalization tends to encourage growth However, a constraint that is emerging as more infor- and creates new job opportunities, potentially allow- mation that could have important disaster reduction ing people to move to better jobs. According to the value is generated, is the capacity at country and regional World Bank (2002), in the long run workers gain from levels to interpret and utilize these data. National integration, with wages growing twice as fast in the more meteorological systems provided as a public good, for globalized developing countries than in the less glob- example, have to compete for recurrent expenditure alized ones and faster than in rich countries. However, with all other areas of public spending. In all of the a reduction in either poverty or vulnerability is not case studies undertaken by the authors, there was evi- inevitable. Indeed, the World Bank (2002:1) states that dence of insufficient spending. This was reflected, for although "global integration is already a powerful force instance, in inadequate operation and maintenance of for poverty reduction... it could be even more effective." monitoring systems.6 For example, skilled wages rise faster, implying that the education system needs to serve all levels of society in order to avoid increasing inequality. Poverty and Vulnerability In terms of vulnerability, economic growth and devel- Poor and socially disadvantaged groups are usually the opment may not solve problems of risk and vulnera- most vulnerable to and affected by natural hazards, bility, as already noted. The declining importance of reflecting their social, cultural, economic, and political agriculture--potentially one of the most hazard-sensitive environment. Disasters, in turn, are a source of tran- sectors--typically associated with globalization may sient hardship and distress and a factor contributing to reduce vulnerability, both directly and as those previ- persistent poverty. At the household level, poverty is ously dependent on agriculture take advantage of increas- the single most important factor determining vulnera- ing alternatives. Some 70 percent of the world's poor and bility, in part reflecting location of housing (e.g., on food-insecure people currently depend on agriculture floodplains, riverbanks, steep slopes, or contaminated for their incomes and food entitlements (FAO 2001). land previously occupied by industrial facilities), pri- Enhanced opportunities for diversification of household mary types of occupation, and level of access to finan- income can also help spread risk. However, traditional cial and other resources. The poverty-exacerbating nature coping mechanisms may be simultaneously disrupted. of vulnerability is attributable not only to post-disaster- Within the domestic economy, increased competition related damage, temporary loss of income-generating emanating from globalization can also imply increased opportunities, and increased indebtedness, but also to entry and exit of firms, at least in the shorter term, imply- deliberate risk-averting livelihood choices that poorer ing greater labor market turnover. This can increase sen- households may make. For example, poorer households sitivity to natural hazards and other shocks, requiring may choose to forgo the potential benefits of higher- efforts to ensure that adequate social protection pro- yielding crops in favor of more hazard-tolerant ones, grams are in place. As the World Bank (2002) notes, implying more stable and secure but, in most years, social protection may also be crucial in encouraging poor lower earnings. people to take the risks involved in entrepreneurship. Disasters, Vulnerability, and the Global Economy 13 Those facing higher levels of risk, such as those ema- private sector was allowed to cover a large part of the nating from natural hazards, may require particular deficit after the floods in 1998. Small open economies encouragement and support in recasting behavior from like Dominica, marginal to world and regional or global that of risk minimization to profit maximization. markets, face logistical but not access difficulties to Globalization and associated growth in the manu- food imports after a disaster. There are still important facturing sector as well as cuts in agricultural tariffs7 exceptions, however: countries like Malawi, which is cur- also fuel urbanization. This process is often rapid and rently experiencing a food crisis, that have difficulty in unplanned, by implication forcing poorer groups to live financing and organizing national food security. in more marginal and hazardous areas such as flood- It is well recognized within the considerable body plains, riverbanks, steep slopes, and reclaimed land of literature on food security that natural hazard events, (IFRC 2002). Sensitive and carefully designed meas- in particular droughts, are one of the principal triggers ures are required to help redress associated risks. of potential transitory food insecurity for particular seg- Meanwhile, the World Bank (2002) points out that ments of a population. In that light, it is relevant to con- in those countries left out of the globalization process-- sider the implications of globalization, particularly which contain some 2 billion people--many are facing agricultural trade liberalization, for sensitivity to chronic declining incomes and rising poverty. Whether this is and transitory food security. a direct consequence of the fact that they are not glob- Historically, agriculture has represented a special case, alizers is not clear. However, the fact remains that a sig- with various exceptions made for domestic support price nificant segment of the world's population, located in schemes under successive GATT negotiations, as already these countries, may remain poor and thus particu- noted. However, under the present WTO Agreement larly vulnerable to natural hazards, despite global trends on Agriculture, it was agreed that WTO member coun- toward increasing integration and growth. tries, other than LDCs, should reduce barriers to market access and market-distorting forms of domestic sup- port to agriculture. Developed countries have now imple- Food Security mented this agreement, while the implementation period Food security is "a situation that exists when all people, for developing countries will conclude in 2004. How- at all times, have physical, social and economic access ever, there are concerns that liberalization may not result to sufficient, safe and nutritious food that meets their in enhanced food security, as reflected, for instance, in dietary needs and food preferences for an active and "a common thread through many proposals by devel- healthy life" (FAO 2002a). oping countries that staple food crops should be exempted This emphasis on people's access as the key to food from limits on, or reductions in, support under WTO security is a measure of the considerable progress arrangements" (Roberts and others 2002: 40). made toward assuring food security at national and In theory, trade liberalization and associated move- international levels that is partly a consequence of the ments in relative prices of different crops should trig- liberalization of external trade and currency markets. ger a supply response, with more rational allocation of Most, but not all, developing countries are now able to resources. This may lead to an increase in aggregate acquire additional food imports to respond to tempo- agricultural production levels and net incomes. Such rary deficits. This is in stark contrast to the situation that responses would be more likely to reduce chronic, prevailed in the early 1970s. For example, Bangladesh, poverty-related food insecurity. Furthermore, the supply a low-income country with sometimes large, temporary response could be modified by various constraints relat- additional import requirements, was unable to finance ing to access to markets, agroclimatic factors, and the food imports in the famine crisis of 1974 and was fur- level and availability of assets (including land), skills, ther hampered by a U.S. embargo. Subsequently, its gov- and credit. As the 2001 IFAD Rural Poverty Report (IFAD ernment responded to major disasters with a combination 2001) states, "under globalization, market access becomes of making massive commercial purchases and seeking-- increasingly important as only those who have it can usually successfully--large-scale food aid. Finally, the exploit the new opportunities. Without market access, 14 Building Safer Cities: The Future of Disaster Risk the potential benefits of higher product prices and lower is mainly associated with poverty, and cheaper imports input prices are not transmitted to poor households. can be beneficial (Thompson 1999). This is most unam- Remoteness also restricts access to information about biguously so for the rapidly growing numbers of poor new technologies and changing prices, leaving the poor urban consumers dependent almost entirely on market unable to respond to changes in incentives." supply. If trade liberalization promotes economic growth Moreover, even if agricultural production does and this, in turn, reduces levels of poverty, then this, increase, this does not necessarily imply an improve- too, can improve food security, again by increasing access ment in food security. Any shifts between food and of the poor to food. nonfood cash crops and between tradable and non- In summary, the impact of trade liberalization on taxable can have implications for food security (FAO food security has been broadly positive at a global 2002b). Some people may lose their livelihoods as part level. But the short-term consequences of liberaliza- of the restructuring process associated with both tion are less clear. Food security continues to be a highly agricultural and broader liberalization, again with country-specific issue, in part depending on the nature potentially negative chronic food security and related and scale of agriculture and the significance of the sector poverty implications. Increased exposure to competi- as a form of employment. There will be both winners tion and world price fluctuations in countries where and losers, and impacts on food security are likely to agricultural industries were previously protected from vary between groups--for instance, between small-scale import competition could also expose some farmers and commercial farmers and between farmers, rural to transitory food insecurity. Oxfam (2000), for exam- nonfarm producers, and urban consumers. ple, asks whether small-scale farmers can compete in In terms of implications for sensitivity to natural haz- a liberalized environment and whether there is a ards, the impacts are, again, likely to vary between coun- need to retain some level of protection. Farmers in tries. From a consumer's perspective, increased access developing countries also typically have even more to world markets could dampen disaster-related food limited access to futures markets and other risk man- deficits resulting from reduced domestic production. agement tools (although globalization could help To the extent that globalization more generally facili- improve access--see section on FDI). In addition, tates the spread of risk associated with a decline in many have few financial reserves. The two factors com- production, it is also positive. bined leave farmers more exposed to sudden price fluctuations under more liberalized conditions, poten- tially restricting their productive capacity the follow- Environment ing season (Roberts and others 2002). Liberalization could cause increased short-run volatil- Finally, concerns have also been expressed about the ity in international grain markets, posing difficulties for impact of globalization on the environment. Environ- importing low-income countries. This possibility was mental degradation, both via greenhouse gas emis- highlighted by the severe price spike in international sions and physical destruction, has implications for wheat and coarse grain markets during 1995­96, the scale, frequency, and extent of the impact of natu- when there was a rapid reduction in U.S. and other stocks ral hazards. There is clear evidence that a number of to low levels. Food aid levels plummeted as well. They countries are becoming increasingly vulnerable to nat- also coincided and were thought to be associated with ural phenomena as a consequence of environmental the more liberal trade provisions of the 1995 U.S. Farm degradation, particularly deforestation, and increased Bill (Konandreas 2000). These developments signifi- cultivation and occupation of marginal lands. Defor- cantly increased, for example, the import costs for south- estation has disrupted watersheds, leading to more severe ern African countries of coping with the 1994­95 droughts and floods. It has also resulted in the siltation drought. of riverbeds, deltas, bays, and gulfs, again increasing From the perspective of consumers, food security is to the incidence of flooding. Meanwhile, impacts of changes a large degree an issue of affordability, food insecurity in the composition of the atmosphere on the frequency Disasters, Vulnerability, and the Global Economy 15 and intensity of climatic hazards are predicted to vary Implications of Natural Hazards for Globalization significantly between regions and subregions but there are expectations of more extreme weather variability, Risk in various forms can have potential implications with associated increases in the incidence of droughts for the pace and nature of globalization, whether related and floods, as well as sea level rises, in many parts of to such factors as exchange rate instability or natural the world. hazards. This section considers the role of the latter in Globalization is widely considered to be a cause of determining the extent to which countries are integrated environmental degradation, as illustrated in the quote into the global economy and, in addition, are able to from the Sienna Declaration cited earlier. In discussing reap the potential benefits of that integration. the impact of FDI more specifically, a recent WWF-UK It is beyond the scope of this paper to undertake an report states that "the past decade has ... seen all major empirical examination of factors determining differ- trends of environmental degradation accelerate--for ences in levels of global integration across countries or, example, greenhouse gas emissions, deforestation, loss in particular, to explore the linkages between disasters, of biodiversity. Such patterns of environmental damage growth, and patterns of globalization.8 Nevertheless, have been driven by increased economic activity, to natural hazards could be another factor preventing the which FDI is an increasingly significant contributor" growth benefits of globalization from being achieved (Mabey and McNally 1998:3). However, there is also a and, as discussed in further detailed below, in some cases counter argument that globalization does not neces- even inhibiting the pace of integration itself. As the sarily directly exacerbate this process. Regarding World Bank (2002: 5) states, "while the new globalizers deforestation, for instance, growth is often associated are beginning to catch up, much of the rest of the devel- with reductions in forest area, most obviously where oping world--with about 2 billion people--is becom- there is a timber export sector and land is being ing marginalized. Their aggregate growth rate was actually cleared for export-oriented production. However, the negative in the 1990s" (World Bank 2002: 5). World Bank (2002) argues that particularly high rates of deforestation in some countries may not be the direct Disasters, Growth, and Globalization result of globalization so much as they are domestic fac- tors. In discussing the more general argument that inten- The direction of causality between high growth and sification of competition creates a potential for a "race increasing participation in world trade has yet to be to the bottom" and "pollution havens," with govern- established. Nevertheless, it is widely observed that ments perhaps trying to attain a competitive advantage these two phenomena are correlated. More open, export- by lowering their environmental standards, the World oriented economies are also more successful in attract- Bank (2002) also argues that available evidence sug- ing FDI (see later discussion). Again, each affects the gests that this is not happening. Evidently, the costs other, but empirical analysis by Singh and Jun (1995) imposed by environmental regulation are small rela- suggests that, on balance, openness encourages FDI tive to other considerations, and so their impact upon rather than vice versa. location decisions between rich and poor countries is Increasing integration can occur without raising minimal. The WWF-UK report refutes this, however, growth, but this surely implies that some of the major arguing that studies on which such statements are based potential benefits of globalization--specifically, growth "have had serious flaws, and an excessive focus on site- and related rising per capita income and, hopefully, a specific environmental impacts and emissions of a few reduction in the level of poverty--will be lost. As Roberts industrial pollutants" (Mabey and McNally 1998:3). and others (2002: 36) comment, "whether such ready The report continues on to present "ample empirical movement does in fact occur depends on whether there evidence that resource and pollution-intensive indus- is sufficient growth in the economy and alternative activ- tries do have a locational preference for, and an influ- ities available to absorb resources displaced through ence in creating, areas of low environmental standards" trade liberalization," in turn requiring flexible economic (Mabey and McNally 1998:3). structures and sufficient demand for labor and other 16 Building Safer Cities: The Future of Disaster Risk resources to enable relatively rapid and substantial adjust- growth objectives, of various policy options in dealing ment between activities. with disasters. The study concluded that potential catas- Natural hazards could be another factor preventing trophes should be incorporated into economic projec- the growth benefits of globalization from being achieved tions for three reasons: high opportunity costs associated or, depending on the direction of causality, preventing with the diversion of scarce financial resources into post- increasing integration into the global economy by restrict- disaster relief and reconstruction efforts; the havoc ing growth. imposed by natural disasters on the already-complicated Theories of development place considerable empha- budgetary planning process; and the high demands that sis on the roles of capital and labor growth and pro- natural disasters place on international aid resources, ductivity (e.g., Solow 1956; Denison 1967). Yet capital diverting resources away from development uses. assets and other resources can be severely affected by There has been little empirical analysis of historical natural disasters while productivity of undamaged evidence on the impact of disasters on long-term growth, capital and labor can be reduced by associated disrup- however. Benson (forthcoming) attempts to address this tions to infrastructure and markets. There could be gap, examining comparative cross-sectional data on significant direct capital losses (except in the case of real GDP performance for 115 countries over a 34-year drought). All major types of disaster, including drought, period from 1960­93. The study involved regression can also disrupt longer-term investment plans, both in analysis and an analysis of relative movements in GDP.9 physical and human capital. Governments may divert Rather than attempting a ranking of countries accord- resources from planned investments to fund the relief ing to natural hazard risk, countries were simply divided and rehabilitation process. Disaster-related external assis- into two categories--higher and lower risk--based on tance may be extended, but this may not be entirely evidence on the incidence of disasters over the period additional, instead in part replacing development aid of analysis.10 Analysis was undertaken both including flows due to some combination of limited donor resources and excluding Sub-Saharan African countries. and local counterpart funding constraints. Other damage The results suggest that, over the past three decades, may be covered by insurance policies, but there are more hazard-prone low-income countries may have still opportunity costs relating to the payment of pre- experienced a relatively slower rate of economic growth miums. And some destroyed assets may not be replaced than their less hazard-prone counterparts who had sim- at all. In the shorter term, disasters and hazard risk can ilar levels of per capita income at the beginning of the also contribute to economic instability and an atmos- period. However, there are fundamental problems in phere of uncertainty. It is widely observed, for instance, undertaking such analysis, in particular, that less hazard- that disasters typically cause a short-term decline in prone countries were already typically among the set GDP (see, e.g., Benson and Clay 2000; Charvériat 2000). of more developed countries by the latter half of the Yet, research indicates that "macroeconomic stability is twentieth century. Thus, the results may simply reflect essential for high and sustainable rates of growth" (Ames Quah's (1993) broader finding of polarization toward and others 2001: 2). Thus, hazard risks combined a bi-modal distribution, with countries beginning at the with post-disaster related economic instability could be higher end of the income distribution likely to experi- a significant disincentive to potential new investment. ence further increases in income. Moreover, a wide range A recent research study undertaken by the Interna- of other factors also could determine rates of growth tional Institute for Applied Systems Analysis (IIASA), Nevertheless, the basic findings, if tentative, are in conjunction with the World Bank, confirms the poten- supported by anecdotal evidence from individual coun- tially adverse long-term impact of natural disasters. The tries, with poorer regions of a country also often more study sought to model the potential implications of nat- hazard-prone. Charvériat (2000), for instance, notes ural disasters for future longer-term growth in three that communities in the northeast part of Brazil and countries (Freeman and others 2001). The analysis coastal areas of Ecuador and Peru are typically poorer focused on their potential impact on capital accumu- than less hazard-prone parts of the same countries. In lation and quantified the implications, in particular for part, such patterns reflect differences in opportunities Disasters, Vulnerability, and the Global Economy 17 for growth and development as determined by the rel- structure. Natural disasters could, in fact, be viewed as ative risks faced by different communities. For exam- an aspect of adverse geography, although the literature ple, farmers in more hazard-prone regions of Vietnam on globalization and the role of geographical factors in have been less well placed to take advantage of higher- determining growth (e.g., Acemoglu and others 2000; yielding but less hazard-tolerant strains of rice, while Diamond 1998; Gallup and Sach 1999) tends to more hazard-prone regions of the country have also ignore them. received disproportionately small shares in private and In terms of the role of infrastructure, Limão and public investment and external assistance (Benson Venables (2001) argue that, now that recent liberaliza- 1997b). tions have reduced artificial trade barriers, the effective Disaster-related budgetary pressures can also affect rate of protection provided by transport costs is con- a country's ability to participate in the global economy siderably higher than that provided by tariffs for many in other ways. In the aftermath of a disaster, a govern- countries. They estimate the elasticity of trade flows ment will be obliged to meet potential budgetary pres- with respect to transport costs at approximately 2.5-- sures by increasing the money supply, drawing down that is, halving transport costs would increase the volume foreign-exchange reserves, or increasing levels of domes- of trade by a factor of five, or improving infrastructure tic and/or external borrowing. Foreign borrowing can from the 75th percentile to the 50th would increase result in an appreciation of the exchange rate, reduc- the volume of trade by 50 percent. Transport costs ing the price of imports and increasing that of exports. depend on various factors including distance, admin- In addition, it can place future strains on the economy istrative barriers, and the structure of the shipping indus- via higher debt-servicing costs.11 Natural disasters can try. However, Limão and Venables (2001) find that also trigger an increase in interest rates charged on infrastructure is also quantitatively important. For exam- new external loans by increasing the risk premia asso- ple, their results suggest that improving one's own and ciated with a country's assets. Another option, the run- transit countries' (that is, countries through which mer- down of foreign-exchange reserves, is limited by the chandise travels before reaching its destination) infra- size of those reserves and entails an appreciation in the structure from the 25th percentile to the 75th percentile exchange rate, with possible associated risks of capital would overcome approximately two-thirds of the dis- flight and a balance-of-payments crisis (Fischer and advantage associated with being landlocked. Easterly 1990). Natural hazard events can destroy transport and other infrastructure. Disasters can also result in the diversi- fication of resources away from new investment and External Trade into reconstruction, ultimately constraining efforts to Many of the nonglobalizers are Sub-Saharan African upgrade transportation systems. Efforts to improve the countries and former Soviet republics, with exports efficiency and economy of the Philippines' transporta- focused on a narrow range of primary commodities, tion systems, for instance, are reported to have been only making them highly vulnerable to commodity price moderately successful because most available resources shocks (World Bank 2002). Their failure to diversify were redirected in response to calamities, with knock- exports has been attributed to various factors, including on implications for the pace of improvement of rural poor policies (e.g., product standards and regulations, transport linkages (Philippine NLUC 1992). Moreover, health and safety regulations, labor and environmental disaster-related repairs can disrupt general maintenance regulations), weak institutions, poor access to infor- operations. In Dominica, unanticipated expenditure on mation, corrupt governance, limited technology, poor the repair of roads following landslides and storm damage infrastructure, adverse geography (e.g., being land- crowds out routine maintenance virtually every year. locked, greater proneness to disease), and climate Obviously, difficult tradeoffs often have to be made (Brahmbhatt 1998; World Bank 2002).12 between the quality and quantity of infrastructure. Natural disasters may certainly have contributed to Construction of less hazard-resistant roads could facil- some of these constraints, in particular poor infra- itate more rapid progress in improving market access. 18 Building Safer Cities: The Future of Disaster Risk However, the vulnerability implications of alternative They suggest that this reflects the fact that direct invest- levels of overall quality and strength (e.g., adequate ment is likely to be capital-intensive and so requires a drainage of roads) should also be explored, as hazards relatively more substantive and long-term commitment. could damage and disrupt transport networks. Disasters are another form of instability, also potentially As already noted, natural hazards can also affect threatening the long-term viability of an investment. patterns and levels of trade in terms of securing mar- Singh and Jun's analysis (1995) additionally indicates kets. If frequent occurrences affect reliability of supply, that work days lost in production, in turn affecting then exporting countries could face difficulties in secur- production efficiency, is more significant for low-FDI ing and maintaining trading partners. counties, presumably reflecting the fact that production in these countries is likely to be more labor-intensive. Frequent or extended natural hazard events (e.g., flood- FDI ing) could affect days worked. There is some evidence that FDI is spatially more The literature suggests that location and related inter- clustered than other forms of production, possibly due national transport costs, the cost of market access through to certain incentives to locate close to other firms, includ- exports, the quality of infrastructure, possession of ing spillovers created by research and development; the raw materials, labor costs, government incentives, polit- development of local networks of suppliers of special- ical risk, per capita income, the degree of industrial- ized goods and services; the development of local ization, and the size of domestic markets are all important labor markets with appropriate specialized skills; and in attracting FDI (Shatz and Venables 2000; Singh and confidence, and the possibility that firms "herd," with Jun 1995; Wheeler and Mody 199213). Generally, firms uncertain whether a particular country is a good more open, export-orientated economies are more location for FDI but willing to take the success of one successful in attracting FDI, as discussed above. Indeed, firm as a signal of underlying national characteristics the relative size of the export sector is the strongest (Shatz and Venables 2000). Again, this herding ten- explanatory variable for FDI flows according to Singh dency could discriminate against more hazard-prone and Jun's (1995) analysis. countries if potential investors are aware of the possi- There is little hard evidence reported in the literature bility of natural hazards, if perceived risk--whether or that natural hazards and related risk have influenced not correctly so--is high, and if few others have been decisions on FDI directly, although there is some anec- seen to invest there. dotal evidence that this may occur (see section on Dominica below). Again, however, natural hazards and risk may have had some indirect impact on factors Country Experiences in an Era of Globalization determining flows. One of the more interesting lines of investigation from a natural hazards perspective con- The general discussion presented earlier indicates that cerns the importance of a hospitable business environ- the linkages between globalization and vulnerability to ment. Singh and Jun (1995) examined this using an natural hazards are complex and that no easily sus- operation risk index based on a range of factors includ- tainable generalizations about impacts and effects can ing balance of payments performance, economic growth, be made. Our three most recent case studies suggest and infrastructure--all factors that natural disasters can that there are important, distinct country type situa- affect. Their results suggest that the business climate is tions. There are therefore likely to be country-specific important for high-FDI countries but not for low-FDI strategies for disaster reduction. These themes are countries. illustrated by a more detailed account of changing vul- Singh and Jun's (1995) analysis also suggests that, using nerability for the small, relatively less complex open a broad-based qualitative political risk index, political economy of Dominica. These developments are shown stability may be important for high-FDI countries, where to be country-specific by contrasting developments in the stakes are higher, but not for low-FDI countries.14 Bangladesh and Malawi. Disasters, Vulnerability, and the Global Economy 19 Dominica impact of three tropical storms in 1995, and Hurricane Lenny in 1999. Hurricane David, a Category 4 hurri- Dominica is an important case with which to begin to cane, directly hit the island and was extremely devas- explore forms of vulnerability and the role of a coun- tating, with severe environmental and demographic try's relationship with the global economy, as it exem- consequences.17 plifies the type of experience faced by many small island There are significant geophysical hazards, as the island economies. Such economies face a number of special is geologically extremely young and almost completely disadvantages associated with their size, insularity, and volcanic in origin. There was a volcanic alert in 1998­99, remoteness (Briguglio 1995), making them highly sen- the first since 1880, and scientists indicate a continu- sitive to economic shocks of any form, including nat- ing, significant risk of an extreme event in the twenty- ural hazards. They are often perceived as some of the first century with a related possibility of earthquakes. countries most vulnerable to natural hazards in the High rainfall in the mountainous, noncoastal areas world.15 Small island economies are typically very open, of the island also results in frequent localized flooding with relatively limited internal forward and backward and landslides, which are recurrent annual problems. linkages, instead relying on international trade to market Other potential hazards include drought, storm surges, their outputs and as a source of capital goods, inputs to bush fires, and tsunamis. domestic production, and consumables. Such countries often strive to find niche export markets, concentrat- Agricultural Exports ing the focus of economic activity accordingly. In so doing, many have secured some form of preferential Over time there have been significant shifts in the nature trade agreement; however, the WTO process is currently of Dominica's vulnerability to natural hazards relating to eroding the protected status of many such exports. changing levels of development and capital investment In the case of Dominica, the level of imports stood in the island and changes in economic activity. Shifts in at equivalent to 65 percent and exports to 25 percent the structure and composition of economic activity, in of GDP in 1997, making the economy very open. turn, have been closely tied to international political Since the 1950s the economy has been reliant on a single and economic interests and export market opportuni- export crop, bananas, for which it had preferential access ties. In the past, as a colonial plantation cum subsistence to E.U. markets. Agriculture and agro-processing com- economy, the impact of natural hazards was heavily bined continue to be the major productive sector, dependent on the sensitivity of the prevailing export crop although agriculture's share in GDP declined from an and the associated structure of production and market- average of 37 percent in 1977­78 to 20 percent in ing. In the first half of the 20th century, limes were the 1997­98. Other private sector activity remains small, dominant crop. Limes are relatively insensitive to high although experiencing some growth since the mid 1970s. winds. They were also grown on plantations owned by Dominica consistently runs a deficit on its external vis- U.K.-based companies able to absorb intermittent losses ible trade account, in part met through tourism earn- and associated recovery costs occurring in just one of ings. Tourism's contribution to GDP remains relatively their countries of operation. This production and mar- low, but by the late 1990s accounted for an estimated keting structure effectively acted as a geographical risk- 35 percent of external earnings (GoCD 2000). spreading mechanism. Meanwhile, small-scale farmers Dominica is susceptible to a wide range of natural produced much of the island's staple foods, roots and hazards. The most common, probable, and historically tubers, known locally as "ground provisions." significant are tropical storms and hurricanes. The major- From the 1950s banana exports, largely grown by ity of the population and infrastructure are located along smallholders, progressively displaced plantation agri- the coast, making Dominica particularly vulnerable to culture. Bananas were exported to the United Kingdom strong winds and high seas.16 There has been a sequence under a preferential access agreement that continued of disasters since 1978: Hurricanes David and Frederick after the United Kingdom joined the E.U. in 1974. This in 1979, Allen in 1980, Hugo in 1989, the cumulative (structural) change increased the overall vulnerability of 20 Building Safer Cities: The Future of Disaster Risk the agricultural sector to natural hazards. Bananas are economy, and it may be extremely difficult for agricul- highly sensitive to damage from winds of 40 or more ture to adjust to globalization and less assured markets miles per hour, so that even the fringe impacts of less of uncertain profitability. From the mid-1990s, how- severe tropical storms can cause serious damage. Small- ever, external factors resulted in declining banana pro- holders are also less able to bear heavy losses, because duction, with falling real prices and the loss of guaranteed of their lack of assets and access to credit. These preferential access to the European market. The WTO changes in the type and structure of production implied ruling against the E.U. is expected to eliminate prefer- increased vulnerability. ences on bananas within the decade (Schiff 2002; Roberts Hurricane David in 1979, followed rapidly the next and others 2002). year by Hurricanes Frederick and Allen, demonstrated Dominica's future, more diversified, agricultural sector that sensitivity, causing severe damage to banana plant- will be more sensitive to natural hazards and other risks. ings. However, this sequence of disasters also led to an Other subsectors lack the risk-spreading arrangements increase in the dominance of bananas, which offered a associated with bananas, namely, WINCROP, STABEX, fast, low-investment means of restoring agricultural liveli- and a protected export market. Some tree crops also hoods in an assured export market. Recovery only lack bananas' capacity for rapid post-disaster recovery. takes 9 to 12 months, even where plantings are totally Thus, a future disaster could be associated with a higher devastated. In contrast, production of copra, the other rate of default on agricultural loans, increased demand major commercial crop, took three to four years to recover. for credit, and slower post-disaster recovery. This dif- The rapid recovery in export production after Hur- ference in risks has been an obstacle in encouraging ricane Hugo in 1989 again demonstrated the resilience agricultural diversification, despite it being official policy of the banana economy. In this case, the compulsory throughout the twentieth century and despite the intense WINCROP banana crop insurance scheme, jointly intro- efforts of government and NGOs to foster a broader duced in 1987­88 by the banana marketing boards of economy. the four Windward Islands (Dominica, Grenada, St. In parallel with shifts within the agricultural sector, Lucia, and St. Vincent), also effectively encouraged replant- the wider economy's sensitivity to natural hazards has also ing of bananas by offering partial financial protection in changed over the past two decades as a consequence of the event of a disaster. The E.U., through STABEX, had changesinitscompositionacceleratedbytheWTOprocess. also provided the government of Dominica and other Agriculture's share of GDP halved to only 19 percent associate countries with a partial compensation mech- between 1977 and 1997, while manufacturing, tourism, anism for fluctuations in agricultural export earnings. and international financial services--the latter two by def- So public finances partially dependent on export earn- inition closely linked to the global economy--grew and ings were also buffered from the effects of disaster shock. increased their share of GDP. These latter service sectors The dominance of bananas in Dominica and similar are less sensitive to all except a catastrophic event, such as monocrop agricultural sectors in other small island Hurricane David. Indeed, if the country's recent expan- economies perhaps exemplifies a progressive adapta- sion into international financial services proves success- tion to a specific external economic environment, a ful, then a further decline in broad economic vulnerability process often accompanied by institutional innova- can be anticipated in the future. The international finan- tion. The structural change from estates to smaller com- cial services sector has little reliance on physical infra- mercial holdings in Dominica resulted in production structure and is not linked in any way to the domestic with relatively low overheads and fixed capital at risk. economy (including domestic financial markets). WINCROP--an outcome of regional cooperation-- helps to manage the risks associated with an extremely Infrastructure hazard-sensitive crop. Extension, credit, and market- ing arrangements are also closely tied to the specifics Development of the island's key infrastructure, in par- of this crop and its production structure. These insti- ticular harbors and the road system, provides another tutional arrangements become embedded within the example of changing long-term sensitivity to natural Disasters, Vulnerability, and the Global Economy 21 hazards, in this case linked to Dominica's broad level or among the general public of volcano-seismic hazard of development rather than the structure and compo- until the alert in 1998­99. Awareness was heightened by sition of economic activity. Until the 1950s, sea trans- mediacoverageoftheongoingeruptioninnearbyMontser- port was the primary form of intra-island movement, rat and the arrival of some Montserratian evacuees. implying rapid recovery of the transport network in Hurricane David in 1979 caused severe damage to the the aftermath of a storm. The more recent emergence whole of the island's capital stock. The population loss of roads as the major form of transport, coupled with from out-migration was not up for 20 years. There was a the mountainous terrain, which forces much of the road relatedunquantifiedlossofhumancapital.Tourism,largely network along the coastline, has effectively exacerbated uninsured and dependent on local finance, did not recover the direct and indirect impacts of storms. The scale of for almost a decade. Following this catastrophe and sub- physical damage to the transport network has become sequent severe storms, there has been piecemeal public far more severe and the pace of recovery much slower, investment in more hazard-proof infrastructure and pri- with knock-on implications for the movement of vate sector investment in industrial and service sector goods and people. Increasing vulnerability of this nature construction. Nevertheless, Hurricane Lenny in 1999 can have extreme consequences in a country like caused considerable temporary disruption and damage Dominica, with limited capital resources relative to to the infrastructure of ports and roads. The pattern of demand and thus a tendency to select least-cost solu- aggregate macroeconomic impacts of disasters in terms tions in meeting infrastructure needs; this vulnerabil- of GDP and sectoral product (as shown in figure 1.1) ity was exposed by the catastrophic Hurricane David. suggests that vulnerability to climatic hazards had peaked Similar issues relating to limited capital investment around independence, just prior to Hurricane David. Sub- resources have been encountered in constructing port sequently the impact of storms has become relatively less facilities. The expansion of external trade, including severe due to disaster-proofing and structural changes in highly bulky, refrigerated bananas, and the growth of the economy. Public finances were also in disarray and cruise ship tourism required more extensive port facil- there were problems of governance in 1979. ities. Funding such investment at apparently acceptable What were the longer-term development conse- rates of return, however, resulted in compromises in the quences? Dominica probably lost ground to other islands storm proofing of new facilities in the 1970s and 1980s, such as Barbados and St. Lucia on the post-independ- with costly consequences. The 1979 and 1989 hurri- ence tourist and financial services front. It also became canes created severe disruption and high repair costs. a source of less skilled labor to neighboring French and Anglophone islands. It is among the poorest of the smaller Caribbean economies. Changing Risks There are two important qualifications to the con- Gradual changes in the character of sensitivity of an clusion that vulnerability to natural hazards is declin- economy to natural hazards, such as those described ing. First, there are the uncertain consequences of climatic in the case of Dominica, can go unrecognized. Infor- change. Second, the scale of the threat posed by volcano- mants for the Dominica study suggested that the impact seismic hazard is increasing. Economic and popula- of Hurricane David in 1979 was in part so severe because tion growth have been increasingly concentrated in the the island had not experienced a hurricane for 40 years. capital, Roseau, which is in a relatively high-risk zone Thus everyone was caught unaware. Though Dominica in the event of a severe eruption. Scientific monitoring had not experienced a Category 4 hurricane since 1930, has indicated a significant risk of an extreme event in however, meteorological records show that there had, the twenty-first century. This is a real dilemma. Land- in fact, been a number of less severe storms. Instead, it use planning and regulation could reduce volcanic hazard would appear that the changing nature of and apparent risk. However, in a highly competitive regional econ- rise in the island's vulnerability to storms had not fully omy, with many islands seeking FDI in tourism and impinged on perceptions of risk. Similarly, there was little trying to develop financial services--Dominica's own awareness in government, in the business community, potential growth sectors--investors could easily be 22 Building Safer Cities: The Future of Disaster Risk Figure 1.1 Dominica--Annual fluctuations in agricultural, nonagricultural, and total GDP, 1978­99 H David & H Frederick H Allen TS Klaus H Hugo TS Debbie 3 storms TS Hortense H Lenny 30 20 10 change 0 centage per -10 -on-year earY-20 -30 -40 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Total GDP Nonagricultural GDP Agricultural GDP Source: Benson and Clay 2001. discouraged if attention is drawn to Dominica's hazard directly affected by the 1998 floods. Rapid-onset flash risks. flooding, tornadoes, and landslides are frequent causes of more localized but intense human suffering and loss. Severe earthquakes have been rare but are a potentially Bangladesh catastrophic hazard. Around 45 percent of Bangladesh's population is classified as poor and some 23 percent Bangladesh is one of the most disaster-prone countries live in absolute poverty. These people are typically living in the world. Most of its large, densely settled popula- and working in areas most at risk from natural hazards. tion of 130 million people is at significant risk to more At the household level, poverty is still the single most than one form of natural hazard, making it a test case important factor determining vulnerability. for international efforts in disaster reduction River flooding: there have been 4 extreme events in 30 years--1974, 1987, 1988, and 1998. Other very high floods in 1976 and 1984, though less severe when Hazards and Major Disasters measured in terms of height, maximum flow, and pro- In terms of area and number of people directly affected, portion of area inundated, caused widespread suffer- impact on economic activity, and damage and destruc- ing and losses and elicited an international emergency tion of assets, the types of hazard that have been most response. The implied annual risk of an extreme flood important since independence in 1971 are: exception- is a high 10­20 percent. ally widespread riverine flooding; severe tropical cyclones Over 100 years at least 14 very severe storms have and associated coastal storm surges; river bank erosion; impacted Bangladesh with an implied annual risk level anddrought.Accordingtoofficialestimates,139,000people of more than 10 percent. The worst storms accompanied were killed during the 1991 cyclone, and 31 million by storm surges have been catastrophic. The cyclone of Disasters, Vulnerability, and the Global Economy 23 November 1970 resulted in 300,000­500,000 fatali- of structural adjustment and trade liberalization along- ties; that of May 1991 caused 125,000 deaths. side more disciplined monetary management in the These events in particular have created a worldwide 1990s resulted in single-digit inflation and an annual perception of Bangladesh as one of the world's most dis- current account deficit below 2.5 percent of GDP. The aster-prone countries, described in the mid-1970s by reforms have also helped increase private sector devel- the U.S. Secretary of State as a nonviable "basket case." opment and foreign direct investment. Fiscal policy has not been so successful, however. There have been large fiscal deficits, a low tax-to-GDP-ratio, and relatively Economic Performance poor quality spending. Since independence, the Bangladesh economy has A simple assessment of the sensitivity of Bangladesh's achieved impressive rates of growth. It achieved rapid economic performance to major disasters in terms of recovery in the late 1970s following the devastating fluctuations in GDP and rates of growth in agricultural effects of natural hazards, war, and famine in 1970­75; and nonagricultural-sector product as shown in figure and an average real annual growth rate in GDP of 4.2 1.2 highlights some key issues: percent in the 1980s, rising to 5 percent during the · From 1965­75 there was extreme volatility in the 1990s. Average annual per capita GDP growth rose from largely agricultural economy, clearly linked to cata- an average of 1.7 percent in the 1980s to 3.3 percent strophic natural disasters. in the 1990s, reflecting higher GDP growth and declin- · With the notable exception of the most recent 1998 ing population growth. At the same time, there has been floods, major disasters have resulted in a downturn a change in the structural composition of the econ- in the agricultural sector's annual rate of growth. omy: agriculture's share of GDP has declined while the · The impact on the nonagricultural sector looks much industrial and service sectors have expanded, resulting less significant, but longer-term impacts of disasters in a sharp shift in the composition of the country's are not reflected in inter-yearly fluctuations: if resources exports. Exports also rose as a share of GDP from 4 are diverted from productive investment to disaster percent in 1980 to 14 percent in 2000, while imports response, the pace and nature of development will rose from 16 percent to 19 percent. A gradual process be adversely affected. Figure 1.2 Bangladesh--real annual fluctuations in GDP, agricultural, and nonagricultural sector product, 1965­2000 Floods '88 Floods '66 Cyclone '70 Floods '74 Drought '79 Floods '84 Floods '87 Cyclone '91 Drought '94 Floods '98 25 1 War of 20 Indep- endence 15 and change after- 10 math centage 5 per 0 -5 -on-year earY-10 -15 -20 0 1965/66 1966/67 1967/68 1968/69 1969/70 1970/71 1971/72 1972/73 1973/74 1974/75 1975/76 1976/77 1977/78 1978/79 1979/80 1980/81 1981/82 1982/83 1983/84 1984/85 1985/86 1986/87 1987/88 1988/89 1989/90 1990/91 1991/92 1992/93 1993/94 1994/95 1995/96 1996/97 1997/98 1998/99 1999/00 GDP Agricultural sectoral product Nonagricultural sectoral product Source: Benson and Clay 2002. 24 Building Safer Cities: The Future of Disaster Risk · The sensitivity of agricultural and non-agricultural Chittagong is exposed to possible cyclones and storm components of GDP to natural hazards appears to surges, such as that of 1991. There are other risks such be declining over time, suggesting greater resilience. as fire, outside the scope of this study. Finally, building standards in facilities with a short life expectancy in this and most other new industrial developments largely Declining Vulnerability? ignore seismic hazard. Part of Bangladesh's greater resilience is attributable to The third major development has been in the finan- structural change in the agricultural sector. Following cial system, with some important innovations in financial the 1987 and 1988 floods, the relaxation of restrictions services. After the chaotic hyperinflation that contributed on private agricultural investment and imports of equip- to the famine of 1974, the government has managed to ment was associated with a rapid expansion of much- maintain relative financial stability through periods fol- lower-risk dry-winter-season rice and, to a much lesser lowing disasters. Labor migration has played an extent, wheat. Since then, highly flood-prone deep water important role in financing economic growth through rice and jute have gradually been displaced, and after the remittances of incomes. For example, remittances the 1998 floods, the main monsoon-season transplanted increased by 18 percent in the financial year that includes rice finally lost its primacy as the dominant crop. the 1998 floods. Bangladesh has been a leader in devel- As Bangladesh approached self-sufficiency in rice, oping microfinance for the rural and more recently urban the national staple, internal prices displayed reduced poor. Microfinance played a significant although lim- seasonal volatility and moved closer to import-parity ited role in enabling the poor to cope with the costs of price levels with liberalization of the grain import trade. the 1998 floods (del Ninno and others 2001). Impor- After the floods of 1998 large-scale private sector imports tantly, the (central) Bangladesh Bank was also able to covered the greater part of the temporary food gap, lim- protect this critically important financial sector through iting pressures on prices and the public finances (del massive refinancing. Ninno and others 2001). The economic impacts do not reflect or parallel the Urbanization is rapidly creating large urban and severity of disasters in terms of loss of life and human peri-urban zones, including the capital Dhaka, which tragedy. Large, unprotected rural and peri-urban pop- is quickly becoming a sprawling, minimally planned ulations, increasing rapidly due to unchecked popula- megacity with weak, overstretched infrastructure. Since tion growth, remain vulnerable to riverine flooding. The the severe floods of the late 1980s, there has been a de exploitation of ground water for irrigation and human facto shift in flood control investment and protections use has had its downside in the widespread problem of from rural and agricultural to urban and industrial. This arsenic poisoning. Urban flood protection on a flood seems to have been at least partially successful. The plain with high population densities poses severe drainage 1998 floods, of longer duration and with higher river and pollution problems that require unprecedented levels than those of 1987 or 1988, did not severely affect improvements in management of the urban environment, the greater Dhaka metropolitan area or the secondary requiring technical sophistication, investment and oper- towns that received enhanced protection. ational funds, and improvements in governance. Any Export-oriented garment manufacture has been the major failure in urban flood protection would have mas- primary motor of export growth as inward FDI, and sive costs in human and economic terms. some local industrialists exploited the trading niche Positively, the construction of a system of cyclone offered by the MFA. In 1998, there was some disrup- shelters and improvement in storm warnings appear to tion to supply and export chains, but the industry, largely have reduced the considerable risks to human life posed based in less-flood-affected urban zones, proved resilient. by tropical cyclones and accompanying storm surges However, for the future it appears that risks have altered from the Bay of Bengal. But this threat to large popula- rather than been reduced. The industries' markets are tions settled in high-risk coastal areas is by no means far from assured and could be lost if there were a major eliminated. There are still considerable institutional disaster-related disruption. Manufacturing in coastal problems concerning control and access to shelters Disasters, Vulnerability, and the Global Economy 25 and the maintenance of coastal embankments that could Malawi mitigate the impacts of storm surges (IFRC 2002). There are also two major sources of increased hazard Since 1990 Malawi and other countries in Southern vulnerability. First, the scale of the threat posed by Africa seem to have experienced increased economic seismic hazard is increasing. Rapid economic and pop- volatility that is linked with climatic variability (figure ulation growth has been increasingly concentrated in 1.3). This apparent increase in vulnerability has occurred the capital, Dhaka, and other urban centers that would during a period of many complex interacting devel- be devastated by a major earthquake. Bangladesh is part opments in the region--some positive, such as the of a high-risk region. Minor tremors are common and political reintegration of South Africa and the end to one of the most extreme events, the 1897 earthquake conflict in Mozambique, and others negative, such as (8.8 on the Richter scale), had its epicenter in the nearby the increasing problems of governance in Malawi, Zambia, Shillong Plateau of the Indian State of Megalaya. Local and Zimbabwe and the HIV/AIDS epidemic, which are assessments provide only highly tentative risk zoning undermining the capacity to cope with shocks. These within the country in map form because of the inade- developments are highlighted by what has happened in quacies of available data (Ali and Choudhury 2001). Malawi. Second, there are the uncertain consequences of envi- Malawi, small and landlocked, recorded a population ronmental change, some recorded and others only so of 10.8 million in 2000. It is one of the poorest countries far identified as possible consequences of global climatic in Africa, with per capita GDP of US$170 in 2000. Health change. Human activity in Bangladesh and the imme- and social indicators are also among the lowest and diate region may also be altering the likelihood of spe- declining: average life expectancy fell from 43 in 1996 cific events as well as the associated effects. to 37 in 2000 and Malawi is one of the countries most Figure 1.3 Malawi--real annual fluctuations in GDP and agricultural, industrial, and services sector product, 1980­98 Drought Drought 60 50 40 30 change 20 centage 10 per 0 -on-year earY-10 -20 -30 -40 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 GDP Agricultural sector product Industrial sector product Services sector product Source: Clay and others 2003. 26 Building Safer Cities: The Future of Disaster Risk severely affected by HIV/AIDS. The loss of human cap- These influences have included some relating to changes ital and ill health among the economically active pop- in the external economic environment. ulation are likely making the country more disaster-prone. Agricultural development has stalled. Demographic Malawi still has a largely rural economy, with 89 growth averaging 2.6 percent in the 1990s has placed percent of the economically active population classified increasing pressure on agricultural systems that are an as rural. Agriculture accounted for some 40 percent of adaptation of shifting cultivation. Declines in soil fer- GDP in 2000, compared with 44 percent in 1980. Its tility on holdings of shrinking size are barely compen- share in GDP was declining but rose again in the sated for by increased fertilizer use and other technical 1990s, with industrial stagnation and contraction in the improvements that could increase productivity. Liber- public service sectors. Export earnings are dominated alization of internal agricultural markets has been rel- by agricultural commodities, largely rainfed tobacco, atively unsuccessful. The private sector has been unable making the economy sensitive to climatic variability and to take on and efficiently handle functions that were commodity price shocks. previously the responsibility of parastatals, especially Although there has been internal liberalization and the agricultural marketing agency, ADMARC. a reduction in tariffs, the Malawi economy has become Conflict in neighboring Mozambique, and more relatively less open over time. Exports have declined recently, the process of reintegration of South Africa into as a proportion of GDP from 28 percent in 1980 to the regional polity and economy have contributed to 24 percent in 2000. Imports fell from 43 percent to the failure of industrialization or service subsectors such 40 percent. as tourism to provide alternative sources of economic The main source of natural hazard vulnerability in growth and employment. Malawi is climatic variability. The major food staple, The relative deindustrialization of Malawi shows rainfed maize, accounting for more than 70 percent of the need for caution in assuming that regional devel- energy intake, is extremely sensitive not only to drought opment will be consistent with broader global trends. or low rainfall, but to erratic rainfall within the grow- The disruption to external communication because of ing season and, as the 2001 season showed, to abnor- the war in Mozambique from the late 1970s increased mally high rainfall. There were only two clearly transport costs, reducing export parity and raising import defined droughts in the twentieth century: the drought parity prices. This favored low-input, self-provisioning that caused a famine in 1949 and another that reduced rather than export-oriented agriculture, encouraging maize production by 60 percent in 1991­92. How- the development of small-scale manufacturing enter- ever, relatively unfavorable conditions such as the widely prises, although growth was checked by limited domes- reduced and erratic rainfall of 1993­94, extremely high tic demand. However, the more recent progressive rainfall as in 2001, or locally erratic rainfall as in 2002 reintegration of South Africa into the regional econ- pose increased food security and wider economic threats omy has exposed small-scale manufacture and pro- to a more vulnerable, less resilient economy. cessing of tradables in Malawi and other "front line Riverine flooding is an annual, relatively predictable states" to a larger-scale, absolutely more efficient com- hazard in the lower-population-density southern dis- petitor. This adjustment effect amounts to de-industri- tricts. Even in 2001, flooding did not have a widespread, alization, making the economy more exposed to catastrophic impact. There are apparently no other sig- agricultural sector volatility. nificant forms of natural hazard. Malawi and some neighboring countries have been beset with problems of conflict, governance, and weak public financial management. These have amplified the Sources of Increasing Vulnerability difficulties caused by economic sensitivity to climatic A variety of influences has interacted to make the econ- variability. In 1991­92, the economic effects of drought omy and society increasingly sensitive to climatic vari- were intensified by the effects of an influx of displaced ability,notjusttheextreme"drought"eventsthatarewidely people from Mozambique and the halt of bilateral assis- but simplistically perceived to impact Southern Africa. tance other than emergency relief. In 1994, the effects Disasters, Vulnerability, and the Global Economy 27 of an agricultural-sector shock were compounded by the reaction of both domestic and international mar- weak fiscal and monetary management in a hyperin- kets, which can trigger fundamental-based or self-ful- flationary situation. In 2000­01, there was donor filling crises." However, they also note that, although pressure to reduce parastatal debt by reducing grain the evidence on the impact of globalization is still very stocks. Then, as the food security situation deteriorated scarce, any observed increase in volatility seems to occur after the 2001 harvest, there was donor reluctance to in the short run only, and that volatility decreases in respond to aid requests from the government, which the long term. Indeed, they conclude that there is could not account for revenues from its grain market- scope for much deeper globalization, given its poten- ing operations, including local currencies generated by tial benefits, but that efforts are also required to seek to the sale of aid commodities. It is debatable whether the minimize associated risks. food security crisis that emerged in Malawi during This paper draws upon a limited number of in-depth 2001­02 should be categorized as the consequence of country studies. As such, its findings should be con- a natural hazard. Rather, climatic variability over two sidered as hypotheses for wider testing. Nevertheless, years within a range that had not previously been regarded it is striking to note that most of the findings confirm as disastrously destabilizing contributed to a crisis in an and elaborate conclusions and policy presumptions in economy made more vulnerable by structural changes the wider globalization literature, which focuses on and other developments that had reduced resilience at market-related and financial risks rather than natural all levels. Unfortunately, the onset of an El Niño event hazards. in 2002, with its prospect of low and erratic rainfall, From a natural hazards perspective, an important objec- increases the risk of a third, disastrous year. tive is to seek ways of using global markets to improve risk management. There may be opportunities in the area of smaller enterprises and consumers, as well as in larger Conclusions corporations and government. The Montserrat case (box 1.1) and potentially similar risks to narrowly based, The sensitivity of an economy to natural hazards is deter- locally important, and highly successful financial institu- mined by a complex, dynamic set of developmental, tions in Dominica and other smaller, hazard-sensitive economic, and societal influences, including powerful economies imply an urgent need to spread risks. Increas- external factors. The evidence presented in this paper ing global integration may create opportunities for spread- suggests that increasing integration of economies around ing risks borne by micro-finance institutions as well. the world has significant implications for the nature of Exploiting such opportunities may require international sensitivity to natural hazards. In particular, globalization encouragement and support. has expanded opportunities for risk diversification and, In the context of the December 2002 ProVention con- for nations as a whole, it seems to be a positive trend. ference highlighting urban disaster reduction, the coun- However, the question of whether globalization ultimately try studies suggest that different types of natural hazard exacerbates or reduces sensitivity, both of particular risk have distinctive economic dynamics. Developing economies and individual households, is complicated countries responding more successfully to the oppor- and depends on specific country circumstances, includ- tunities and challenges of globalization are showing some ing public action to reduce vulnerability. reduction in relative sensitivity (measured as a propor- On the downside, globalization exposes countries to tion of GDP or sector product) to more predictable, rel- new forms of risk, possibly exacerbating the impact of atively frequent, climatic hazards such as tropical cyclones natural hazards when different risk events coincide. Writ- in Dominica and extreme riverine flooding in Bangladesh. ing about financial globalization specifically, Schmukler An important qualification to such trends is the highly and Zoido-Lobatón (2001:18) ask: "Is the link between uncertain implication of global climatic change for the globalization, crises, and contagion important enough frequency and severity of natural hazards. to outweigh the benefits of globalization?" They cau- In contrast, the exposure to geophysical hazards tion that "in open economies, countries are subject to appears to be rising. Rapid urbanization--a process 28 Building Safer Cities: The Future of Disaster Risk often associated with globalization--creates large con- Africa have been reinvestigated in a study focusing more specif- centrations of people and physical capital, mostly built ically on Malawi (Clay and others 2003). There have been three with little regard for natural hazards either in choice of studies of small island economies: Dominica (Benson and Clay location or design. These geophysical hazards typically 2001), Fiji (Benson 1997a) and Montserrat (Clay and others have relatively low but difficult-to-determine risks, 1999) as well as three studies for larger Asian economies, less than 1 percent annually for an extreme earthquake Bangladesh (Benson and Clay 2002), Philippines (Benson 1997c) in Bangladesh or a disastrous volcanic event in Dominica. and Vietnam (Benson 1997b). Globally, such increasing hazard exposure implies rising 3. For example, an official assessment of the costs of the 1998 disaster-related losses. Bangladesh floods aggregated capital losses, such as damage to The most worrying position is that of countries and infrastructure with rice crop losses. An assessment of Hurricane even regions that are apparently being marginalized in Lenny in 1999 in Dominica included costs of physical damage the process of globalization. In re-examining the con- and reductions in income from small-scale fisheries. sequences of climatic variability in Southern Africa after 4. International companies operating in the sugar sector are almost a decade, there is substantial evidence of attempting to take climatic forecasts into account in this way. greater vulnerability to natural hazards. Natural haz- Private communication from Dr. M. Evans. See also Bohn, ards, in turn, may well be at least indirect compound- forthcoming. ing factors limiting opportunities and potential for 5. In both these cases flows from outside the country contributed globalization for certain economies, although the pre- to the disaster, and these flows were influenced by the actions cise nature of their role is complicated and, again, of public agencies responsible for water management. There were often highly country-specific. inadequate warnings to those responsible for flood response in For those countries that are becoming more closely the affected areas. A contributory factor was insufficiently pre- integrated into the global economy, risks emanating cise understanding of system dynamics and links to exception- from all types of natural hazards should be considered ally high rainfall (Akteer Hossain 2001; Christie and Hanlon in assessing the impacts of reductions in trade barriers 2001). and related changes in the composition of economic 6. For example, there was underfunding of volcano-seismic mon- activity, security of livelihoods, and measures taken to itoring in Dominica in 1998 at the outset of a volcanic emer- help protect vulnerable groups. More broadly, risks ema- gency and no proper wave level monitoring even during Hurricane nating from natural hazards should be taken into account Lenny in 1999. Bangladesh has effectively been without a seis- in the determination of priorities, policies, and strate- mic monitoring system since the separation of Pakistan and India gies, with enhancement of resilience to natural hazards at Partition in 1947. The meteorological system in Malawi lost as one of the basic objectives of government in hazard- access in 1991 to its historical database of climatic information, prone countries. It should also be recognized that suc- impeding investigations for over a decade. cessful risk management requires not only technical, 7. However, Roberts and others (2002) note that the rural non- structural solutions, but also a broader awareness of farm sector has also been expanding and thus some labor released underlying socioeconomic causes and appropriate action. from farming may remain in rural areas. 8. Even had this analysis been undertaken, the literature on glob- alization indicates a number of analytical difficulties that arise Notes in comparing relative integration across countries, implying that 1. The Siena Declaration on the Crisis of Economic Globalization. any findings would have been very approximate at best. Statement prepared by the Board of Directors of the Interna- 9. After Quah (1993), ratios of per capita income relative to the tional Forum on Globalization Siena, Italy, September, 1998. global average were discretized into intervals at 1/4, 1/2, 1 and 2. www.twnside.org.sg/title/siena-cn.htm Annual one-step transition matrices were then estimated by aver- 2. Previous studies have included Benson and Clay (1998) on the aging the observed one-year transitions over every year from economic consequences of drought in Sub-Saharan Africa with 1960-61 to 1992-93. The one 33-step transition between 1960 a more detailed country study of Zimbabwe (Benson 1998). More and 1993 was also analyzed. Analysis was undertaken on the full recently the economic effects of climatic variability in southern data set and three subsets (more hazard-prone, more hazard-prone Disasters, Vulnerability, and the Global Economy 29 excluding Sub-Saharan African (SSA) and less hazard-prone Bibliography countries). Acemoglu, D., S. Johnson, and J. A. Robinson. 2000. "The 10. A more sensitive ranking according to disaster impact is Colonial Origins of Comparative Development: An Empiri- fraught with difficulties, relating in part to incomplete data. cal Investigation." Processed. Berkeley, California, Department 11. In the Philippines, for instance, the 1990 earthquake and of Political Science; University of California, Department of 1989­90 drought were reported to have contributed to a 6.7 Economics; Massachusetts Institute of Technology, Depart- percent increase in total external debt, and a 22.4 percent increase ment of Political Science, and Sloan School of Management, in debt from official creditors alone, in 1990 (Ernst and Young Massachusetts Institute of Technology. 1991). An examination of the impact of the mid-1980s drought Ahmed, K. U., and H. U. Chowdhury. 1998. "The Impact of on external borrowing in six countries in sub-Saharan Africa Migrant Workers" Remittances on Bangladesh Economy." Indian revealed that the growth rate in total debt stocks accelerated Journal of Economics 78 (311). during the year of most severe drought in five of the countries Akhter Hossain, A. N. H. 2001. "Late Monsoon Flood in the South- (Benson and Clay 1998). The one exception, Zimbabwe, had west Region of Bangladesh 2000." Engineering News 42. Dhaka. been pursuing a deliberate long-term policy of debt reduction. Ali Md. H., and J. R. Choudhury. 2001. "Assessment of Seismic Disasters can also create additional external debt pressures to Hazard in Bangladesh." In K. Nizamuddin, ed., Disaster in the extent that they also destroy infrastructure and other assets Bangladesh: selected readings. Dhaka: University of Dhaka, Dis- funded with still-outstanding external loans. aster Research Training and Management Centre. 12. Brahmbhatt (1998) also discusses the role of various struc- Ames, B., W. Brown, S. Devarajan, and A. Izquierdo. 2001. tural factors in determining levels of international trade, includ- "Macroeconomic Policy and Poverty Reduction." Poverty Reduc- ingcountrysize,factorendowmentstructure,geographicalisolation, tion Strategy Program. (Draft. April 2001). World Bank. Avail- and the stage of development. Geographically large countries tend able from . diversity of domestic resources and a large home market, the latter Atkins, J. P., S. Mazzi, and C. D. Easter. 2000. "A Commonwealth implying some reaping of economies of scale even by producing Vulnerability Index for Developing Countries: The Position for domestic consumption alone. Countries with a highly spe- of Small States." Commonwealth Economic Paper Series 40. cialized structure of factor endowments, for example a relative London: Commonwealth Secretariat. abundance of natural resources such as oil, will tend to special- Benson, C. 1997a. "The Economic Impacts of Natural Disasters ize in its production and export, while importing more of their in Fiji." ODI Working Paper No. 97. London: Overseas Devel- other needs. Meanwhile, richer countries tend to devote a higher opment Institute. share of their output and consumption to services. Benson, C. 1997b. "The Economic Impacts of Natural Disasters 13. Cited in Shatz and Venables (2000). in Vietnam." ODI Working Paper No. 98. London: Overseas 14. Singh and Jun (1995) report that empirical evidence on the Development Institute. importance of political stability reported by others is inconclu- Benson, C. 1997c. "The Economic Impacts of Natural Disasters sive, in part depending on how political stability is defined. It in the Philippines." ODI Working Paper No. 99. London: has been variously defined as the number of changes in gov- Overseas Development Institute. ernment, internal armed attacks, riots and so forth. Benson, C. 1998. "Drought and the Zimbabwe Economy, 15. See, for example, Atkins and others 2000; UNDRO 1990; 1980­93." In H. O'Neill and J. Toye, eds., A World Without and the authors' case studies for Dominica (Benson and Clay Famine? London: MacMillan. 2001), Fiji (Benson 1997a) and Montserrat (Clay and others Benson, C., Forthcoming. "The Economy-Wide Impact of 1999). Natural Disasters in Developing Countries." Draft doctoral 16. Some 24 percent of the total population resided in the Roseau thesis. city area at the 1991 census (GoCD 1999). Benson, C., and E. J. Clay. 1998. "The Impact of Drought on 17. The historical incidence since 1886 of tropical storms impact- Sub-Saharan African Economies: A Preliminary Examination." ing directly on or passing very close to Dominica implies an World Bank Technical Paper 401. Washington, D.C.: World Bank. annual risk of 17 percent of one or more Hurricane Category 1 Benson, C., and E. J. Clay. 2000. "Developing Countries and or above and of 4 percent for Hurricane 4 or above. the Economic Impacts of Natural Disasters." In. Alcira Kreimer 30 Building Safer Cities: The Future of Disaster Risk and M. Arnold, eds., Managing Disaster Risk in Emerging Diamond, J. 1998. Guns, Germs and Steel: A Short History of Every- Economies. Washington, D.C.: World Bank: 11­21. body for the Last 13,000 Years. London: Vintage. Benson, C., and E. J. Clay. 2001. "Dominica: Natural Disasters Dollar, D., and A. Kraay. 2000. "Growth is Good for the Poor." and Economic Development in a Small Island State." Disas- Washington, D.C.: World Bank. Ernst and Young. 1991. "Boom- ter Risk Management Working Paper Series No. 2. Washington, ing and Busting." Economic Forecast. Australia: Asia-Pacific D.C.: World Bank Available from . FAO (Food and Agriculture Organization of the United Nations). Benson, C., and Edward J. Clay. 2002. "Bangladesh: Disasters and 2001. "Some issues relating to food security in the context of Public Finance." Disaster Risk Management Working Paper Series the WTO negotiations on Agriculture," in Food Security in 5. Washington, D.C.: World Bank. Available from . Palais des Nations. Bohn, L.E. Forthcoming. "Climate Forecasts in Swaziland: Per- FAO (Food and Agriculture Organization of the United Nations). spectives from Agribusiness." In Karen L. O'Brien and C.H. 2002a. The State of Food Insecurity in the World 2001. Rome Vogel, eds., Coping with Climate Variability: The Use of Sea- FAO (Food and Agriculture Organization of the United Nations). sonal Climate Forecasts in Southern Africa. Aldershot: Ashgate. 2002b. "Trade and Food Security: Lessons of the Past 20 years. Brahmbhatt, M. 1998. "Measuring Global Economic Integration: Analytical Issues and Framework." Rome. ESCP, Commodi- A Review of the Literature and Recent Evidence." Washington, ties and Trade Division. (Draft). D.C.: The World Bank. Available from . 5(2):127­42. Briguglio, L. 1995. "Small Island Developing States and their Freeman, P. K., L. A. Martin, R. Mechler, and K. Warner. 2001. Economic Vulnerabilities." WorldDevelopment 23(9): 1615­1632. "Catastrophes and Development: Integrating Natural Cata- Charvériat, C. 2000. "Natural Disasters in Latin America and the strophes into Development Planning." (Draft). Laxenburg: Caribbean: An Overview of Risk." Research Department Working International Institute for Applied Systems Analysis. Paper:434.Washington,D.C.:Inter-AmericanDevelopmentBank. Gallup, J. L., and J. D. Sachs. 1999. "Geography and Economic Clay, E. J., and others. 1999. "An Evaluation of HMG's Response Development." Consulting Assistance on Economic Reform II to the Montserrat Volcanic Emergency." 2 Vols. Evaluation Report Discussion Papers 39. Cambridge, M.A.: Harvard Institute for EV635. London: Department for International Development. International Development. Clay, Edward J., Louise Bohn, Enrique Blanco de Armas, Singand Government of Bangladesh. 2002. Bangladesh: a National Strat- Kabambe, and Hardwick Tchale. 2003. "Malawi and Southern egy for Economic Growth and Poverty Reduction. Dhaka: Eco- Africa: Climatic Variability, Economic Performance."DisasterRisk nomic Relations Division, Ministry of Finance. Management Working Paper Series 7. Washington, D.C.: World Government of the Commonwealth of Dominica. 1999. Statis- Bank. Available from . Central Statistical Office, Ministry of Finance, Industry and Christie, F., and J. Hanlon. 2001. "Mozambique and the Great Planning, Government of the Commonwealth of Dominica. Flood of 2000." African Issues. Oxford: International African Government of the Commonwealth of Dominica. 2000. Dominica: Institute in association with James Currey and Bloomington: Medium Term Strategy, 2000­2002. Roseau: Government of the Indiana University Press. Commonwealth of Dominica. del Ninno, C., P. A. Dorosh, L. C. Smith, and D. K. Roy. 2001. "The Hadfield, P. 1995. Sixty Seconds that Will Change the World: How 1998FloodsinBangladesh:DisasterImpacts,HouseholdCoping the Coming Tokyo Earthquake Will Wreak Worldwide Economic Strategies, and Response." IFPRI Research Report 123. Wash- Devastation. Second edition. London: Pan Books. ington, D.C.: International Food Policy Research Institute. Hicks, D. 1993. "An Evaluation of the Zimbabwe Drought Denison, E. F. 1967. Why Economic Growth Rates Differ: Postwar Relief Programme 1992/1993: The Roles of Household Level Experience in Nine Western Countries. Washington, D.C.: Brook- Response and Decentralized Decision Making." Harare: World ings Institution. Food Program. Disasters, Vulnerability, and the Global Economy 31 IFAD (International Fund for Agricultural Development). Royal Society. 1992. Risk: Analysis, Perception and Management. 2001. Rural Poverty Report 2001: The Challenge of Ending London: Royal Society. Rural Poverty. Oxford: Oxford University Press. Schiff, M. 2002. "Regional Integration and Development in Small IFRC (International Federation of the Red Cross and Red Cres- States." Policy Research Working Paper 2797. Washington, cent Societies). 2002. World Disasters Report 2002: Focus on D.C.: The World Bank, (March). Available from . IRI (International Research Institute for Climate Prediction). Schmukler, S. L., and P. Zoido-Lobatón. 2001. "Financial 2001. "Coping with the Climate: A Way Forward." Prepara- Globalization: Opportunities and Challenges for Developing tory report and full workshop report. A multi-stakeholder Countries." Washington, D.C.: The World Bank. Available from review of Regional Climate Outlook Forums, concluded at . South Africa. IRI-CW/01/1. Palisades, NY. Shatz, H. J., and A. J. Venables. 2000. "The Geography of Inter- Konandreas, P., R. Sharma, and J. Greenfield. 2000. "The Uruguay national Investment." In G.L. Clark, M. Feldman and M.S. Round, the Marrakesh Decision and the Role of Food Aid." Gertler, eds., The Oxford Handbook of Economic Geography. In E. Clay and O. Stokke, eds., Food and Human Security. Oxford: Oxford University Press. London: F Cass. Singh, H., and K. W. Jun. 1995. "Some New Evidence on Lee, B., and I. Davis. 1998. Forecasts and Warnings: Programme Determinants of Foreign Direct Investment in Developing Overview. U.K. National Coordination Committee for the Countries." Policy Research Working Paper 1531. Washington, IDNDR. London: Thomas Telford. D.C.: International Economics Department, The World Limão, N., and Venables, A.J. 2001. "Infrastructure, Geograph- Bank. ical Disadvantage, Transport Costs and Trade." World Bank Solow, R. M. 1956. "A Contribution to the Theory of Economic Economic Review. Growth." Quarterly Journal of Economics, 70: 65­94. Martin, W. 2001. "Trade Policies, Developing Countries, and Thompson, R. L. 1999. "Rural development: challenges in the Globalization." Washington, D.C.: Development Research next century." Presentation prepared for the Latin American Group, World Bank. and Caribbean Association of Agricultural Economics, Port- Mabey, N., and R. McNally. 1998. Foreign Direct Investment and of-Spain, Trinidad and Tobago, June 30 July 2. Washington, the Environment: From pollution havens to sustainable develop- D.C.: Rural Development Department, World Bank. ment. Godalming, Surrey: WWF-UK. UNDRO (United Nations Disaster Relief Office). 1990. Prelim- Munich Re. 1999. "A Year, a Century, and a Millennium of Nat- inary Study on the Identification of Disaster-Prone Countries Based ural Catastrophes are all nearing their End." Press release of on Economic Impact. New York/Geneva. December 20, 1999. Munich: Munich Re. Wheeler, D., and A. Mody. 1992. "International Investment Loca- Oxfam. 2000. "Agricultural Trade and the Livelihoods of Small tion Decisions: The Case of U.S. Firms." Journal of Interna- Farmers." Oxford: Policy Department. tional Economics 33: 1­2 and 57­76. Philippine NLUC. 1992. National Physical Framework Plan, World Bank. 2000. "What Is Globalization?." World Bank Brief- 1993­2022. Manila: National Land Use Committee, National ing Paper. Washington, D.C.: PREM Economic Policy Group Economic and Development Authority. and Development Economics Group, The World Bank. Quah, D. 1993. "Empirical Cross-Section Dynamics in Economic Available from . Roberts, I., B. Buetre, and F. Jotzo. 2002. "Agricultural Trade World Bank. 2002. Globalization, Growth, and Poverty: Building Reform in the WTO: Special Treatment for Developing an Inclusive World Economy. Washington, D.C.: Oxford. Countries." Abareconomics. Chapter 2 Natural Hazard Risk and Privatization Paul K. Freeman Increases in natural hazard losses intensify the need for allocation of risk are central themes of manuals designed financing dedicated to reconstruction. Multilateral insti- to assist in the privatization process.1 tutions are addressing this issue and establishing new This paper explores the role that privatization can play programs. For example, the World Bank recently imple- in shifting the risk of financing post-natural-disaster mented a Private Sector Development Strategy (PSDS) reconstruction from the government to the private sector. with the objective of increasing private participation in This topic has not been explored in detail in the exist- infrastructure. As noted by the Bank: ing literature. Current practice allocates risk of infra- structure loss from natural hazard events to governments. Most poor people in developing countries have little or Existing practice is predicated on the long-standing no access to efficient infrastructure services. Typically, gov- principle that governments are best able to cope with ernment policies aim at expanding access to infrastruc- large, uncertain risks--the types of risk that character- ture services and at rendering it affordable. Yet, progress has been slow in a number of the poorest countries. (PSDS ize natural hazard catastrophes. Through the power of 2002:10) taxation, governments can efficiently transfer these types of risk to taxpayers. The tool to expanding and accelerating access to infra- For a number of developing countries, the risk of loss structure services is the harnessing of the private market from natural hazards may be handled more efficiently to improve those services, particularly in telecommu- by the private market. The relative cost of transferring nications, electricity distribution, and water pipeline risk to taxpayers may be more expensive than that of systems. The PSDS focuses on activities that increase transferring risk to the private sector. For these coun- the use of the private market to provide essential serv- tries, considering natural hazard risk as a part of the ices, including infrastructure. bundle of risks transferred and a component of the pri- The process of privatization is complicated. For each vatization process may be warranted. project, establishing the macro conditions necessary for This paper will address natural hazard risk and pri- privatization and conducting a detailed cost/benefit vatization through discussion of the increasing costs of analysis are required at a national level. Privatization disasters to infrastructure, detailing the existing justifi- can be described as a process that transfers responsi- cation for allocating natural hazard risk to governments bility for the provision of goods and services from the in the privatization process; exploring circumstances government to the private sector. The process also in which existing practice may be inappropriate; and allocates risk for the provision of these goods and serv- discussing areas where additional research is needed. ices from the government to a private party. In exchange, the private party is provided revenue sources. A difficult task in the privatization process is allo- Natural Hazard Losses to Infrastructure cating risk to the participating parties. A key principle of risk allocation is the assignation of risk to the party The losses to infrastructure from natural hazards are best able to cope, though the risks associated with pri- significant and continue to escalate at an increasing rate. vatization are often complex. The identification and Research indicates two main factors that contribute to 33 34 Building Safer Cities: The Future of Disaster Risk these losses: increasing concentrations of people and Relationship between Increased Damage assets in hazard-prone regions of the world and increases and Climate Change in the intensity and frequency of severe weather-related Researchers have isolated several factors that contribute events. This section will briefly review these trends. to the rising trend in direct damage from catastrophes. One significant factor is the acceleration in weather- Rising Total Direct Damages from Extreme Events related natural hazard events such as hurricanes, cyclones, and flooding. They account for nearly two-thirds of all The ever-increasing losses from natural hazard events are losses from natural hazards, while earthquakes account an important issue for economic development and poverty for most of the remaining third. Figure 2.2 divides losses reduction. Over the last 10 years, economic losses from into specific types of events and shows that, while earth- natural disasters have averaged nearly $580 billion a year quake occurrences have remained relatively stable (figure 2.1). This is a 7.7-fold increase in losses from the over time, the incidence of weather-related events has decade of the 1960s (Munich Re 2002). Due to differ- accelerated. The economic costs of rainstorms, floods, ences in size of the economies in industrialized and devel- droughts, and other extreme weather events have oping countries, however, the economic losses per capita increased 14 times from the decade of the 1950s to the were 20 times greater in developing countries (Bendimerad decade of the 1990s (Munich Re 2002). 2000). From 1991 through 2000, 2.1 billion people were affected by natural disasters, an average of 211 million people annually. Of that number, 98 percent lived in Socioeconomic Factors and Increased Vulnerability medium- and low-development countries as classified by to Natural Hazards theUnitedNations(IFRC2002). Between 1990 and 1998, 94 percent of the world's major natural disasters and 97 While the increasing frequency and severity of extreme percent of all natural-disaster-related deaths occurred in weather events affect the cost of natural hazard risk, developing countries (World Bank 2001). the most important variable increasing damage is the Figure 2.1 Economic losses from natural catastrophes in the 20th century US$ 150 bn 80 70 60 50 40 30 20 10 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 Economic losses (2000 values) of which insured losses (2000 values) Trend of economic losses Trend of insured losses (Amounts in US$ bn) Source: Munich Re 2002. Natural Hazard Risk and Privatization 35 Figure 2.2 Natural catastrophe trends in the 20th century 16 14 12 10 8 6 4 2 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 Others Flood Windstorm Earthquake, volcanic eruption Source: Munich Re 2002. concentration of human populations and their assets though increases in population do not necessarily trans- in hazard-prone regions.2 late into increased vulnerability to natural hazards. It is estimated that natural disaster losses will increase Populations are concentrating in urban areas. The dramatically over the next 50 years. The global cost of movement of people toward cities has accelerated in the natural disasters is anticipated to top $300 billion annu- last 40 years, with 47 percent of the world's population ally by 2050 (UNISDR 2001). Two broad demographic now living in cities, compared to one-third in 1960. trends directly impact the increasing losses from natu- The growth of cities results from births and migration ral hazards in the developing world: population growth to the cities from rural areas. In developing countries, the and the concentration of populations in megacities. In proportion of people living in cities has doubled since 1999, the world's population surpassed 6 billion. This 1960, with more than 40 percent now living in urban represents a tripling of population since the beginning areas. This trend is expected to continue, and by 2030, of the twentieth century. According to U.S. Census nearly 57 percent of the population in less developed Bureau projections, the world's population will increase regions will live in urban areas. In Latin America and the to nearly 8 billion by the end of 2025 and reach 9.3 Caribbean, it is projected that more than 75 percent of billion by 2050--a 50 percent increase above current the population will reside in urban areas by 2030 (UNFPA levels (U.S. Bureau of the Census 1998). Ninety-nine 1999). Urban concentrations in Latin America are the percent of the global increase will occur in developing highest in the world (Charveriat 2000). countries. In 1960, 70 percent of the global popula- Increasing population concentrations in urban regions tion lived in less-developed regions. By 1999, that per- are primarily located in "megacities" with populations centage had increased to 80 percent (UNFPA 1999), of more than 10 million people. In 1960, only New York 36 Building Safer Cities: The Future of Disaster Risk and Tokyo had populations greater than 10 million. By upon the frequency and severity of weather-related events 1999, there were 17 cities of that size, 13 of which were in a given year. Based on historical data, infrastructure in less developed countries. It is projected that by 2015, losses in 1995 alone were $32.6 billion (Munich Re 2002). there will be 26 mega cities, 22 of which will be in less- Research suggests that different types of infrastruc- developed regions of the world. Nearly 10 percent of the ture are vulnerable to different types of natural hazard world's population will live in these cities, up from just events. Housing and roads are particularly vulnerable 1.7percentin1950(UNFPA1999).Urbanizationincreases to earthquake damage (Albala-Bertrand 1993). Droughts risk by concentrating people and investments in limited may have minor impacts on infrastructure and produc- geographic zones. As a result, natural hazards can inflict tive capacity, but can result in heavy crop and livestock substantial damage in a short period of time. Hurricane losses. Floods can cause extensive damage to infrastructure Andrew inflicted $20 billion in damage in a few hours and other production capacities, for example, wiping when it struck Miami in 1992 (Blaikie and others 1994). out agricultural yields (Benson and Clay 2000). Megacities are highly vulnerable to natural disas- In a manual for estimating the socioeconomic effects ters. Nearly half of the world's largest cities are situated of natural disasters, the United Nations Economic Com- along major earthquake zones or tropical cyclone tracks mission for Latin America and the Caribbean (ECLAC) (Bendimerad 2000). Floods, earthquakes, and tropical provides broad outlines of the most probable types of cyclones often strike the same geographic zones more infrastructure damage by type of disaster. For example, than once, and some of the highest risk areas are also the manual explains how floods can impact clean water the most populous. India, China, and Southeast Asia supplies; damage buried and semi-buried tanks and dam are at high risk of seismic activity and floods, hurri- structures; and harm pump equipment. Floods were con- canes, and cyclones. 3 Increasing population concen- sidered to cause damage in all infrastructure categories, trations in urban areas far outstrip the capacity of cities deteriorating or destroying integral structural compo- to absorb this growth. In the 1990s, 60 to 70 percent nents, deforming the land on which infrastructure rests, of urban growth was unplanned (UN/ISDR 2001). Since orrenderingituselesswhenwindandwaterhavedeposited governments in many developing countries already extraneousmaterialsuchasmud,ash,anddebris.Droughts struggle to provide basic services to burgeoning popu- tend to impact infrastructure more mildly, though high- lations, however, it is already difficult to find resources ways can buckle and railroad tracks can misalign when to decrease the vulnerability of poor residents of megac- soldered rails become distorted. Windstorms put addi- ities to natural disaster risk. The spiraling costs of nat- tional stress on buildings, affecting both structural and ural hazard events in developing countries are linked nonstructural elements, though only minimally affecting to the increasing number and intensity of events as foundations and underground elements (ECLAC 1999). well as to concentrations of urban poor in hazard-prone Because different types of infrastructure are differently regions. Losses from catastrophes will be reflected in affected by disasters, the impacts of climate change are increasing infrastructure losses. likely to affect critical infrastructure. Increases in flood- ing and windstorms will have implications for buildings, bridges, roads, and water systems, whereas droughts will Current Infrastructure Losses as a Component affect agriculture and some transportation systems. of Worldwide Direct Losses Since total direct damage increased dramatically world- wide over the past decades, it can be assumed that infra- Allocation of Natural Hazard Risk structure damage as a portion of overall losses increased in the Privatization Process as well. Data from World Development Indicators (World Bank 1999) show that 24 percent of invested capital stock The loss of infrastructure due to natural hazard events is public infrastructure. As of 2001, total direct losses is well known, though at issue--and the topic of this for infrastructure had reached $9.6 billion, though this paper--is finding the most appropriate entity to bear annual loss figure can vary significantly, depending the cost of post-disaster reconstruction. As a general Natural Hazard Risk and Privatization 37 proposition, the risk of loss to government-owned assets The risk-neutral status of governments has influenced is best borne by the government. The justification lies the behavior of industrialized countries in dealing in the ability of governments to use their power of tax- with natural catastrophe risk in the privatization process. ation to spread the cost of loss to their taxpayers. Gen- In examining opportunities to privatize energy activi- erally, the relative size of the loss to the capacity of the ties in the United States, the government explicitly retains government to spread the loss across taxpayers has made the risk of natural hazard losses in the privatization governments the most efficient instrument to manage process. Guidelines from the U.S. Department of Energy risk. "It is profitable for all concerned that risks should (referenced in footnote 1) state that the risk of loss be shifted to the agency best able to bear them through from natural hazards is less expensive in the hands of its wealth and its ability to pool risks. The government, the government than any other party. As a result, the above all other economic agencies, fits this description" government is presumed to be better off by retaining (Arrow 1992). The characteristic assumption that a gov- the risk of loss from natural hazards than by attempt- ernment is the most efficient entity to bear risk under- ing to transfer the risk as a component of the privati- pins decisionmaking about government investments. zation process. The same justification has been applied Most governments maintain a risk-neutral position: to privatization of government-owned facilities in devel- they ignore the risk in making public investment deci- oping countries (Lewis and Mody 1998). sions. Justification for this approach is found in the work While the theory of government risk-neutrality may of Kenneth Arrow (Arrow and Lind 1970). The critical be applicable to most industrialized countries, the use question for Arrow is: "What is the per capita cost of of this theory to justify the assumption of natural hazard public risk-bearing?" The answer lies in computing indi- risk by governments in all cases is limited. In regard to vidual costs and weighting them against the benefits of the privatization of infrastructure projects, the theoret- public risk bearing. As the cost of risk approaches zero ical limits of the policy have a particular application. in the hands of each individual, the risk also approaches Arrow recognized limitations on the general theory zero for the government. As the cost of risk per citizen of the risk-neutral status of governments (Arrow and or per person that can be taxed approaches zero, the Lind 1970). He was particularly interested in projects risk for a country's total wealth decreases. As Arrow where the loss would be felt by one specific group, as states, "when the government undertakes an investment in the case, for example, of a bridge that serviced a lim- it, in effect, spreads the risk among all the taxpayers." ited geographic region. In this case, it is appropriate that A second basis for government risk-neutrality is based the risk profile of the geographic region dictate how risk on portfolio theory. Since risk may be reduced by port- for that project should be managed. Another example folio diversification as well as by spreading risk over cited by Arrow concerns government projects directed large populations, the government as owner of a wide at a particular population group that receives the ben- variety of investments has the potential for the widest efits of a project and bears its risk of loss. An irrigation portfolio diversification. This theory underpins work project in a defined watershed region would be such by Paul Samuelson on capital investment decisions made an example. In this case, if the risk of loss were to be by governments, which arrives at conclusions similar borne by those dependent upon the project, Arrow to those in Arrow's analysis (Samuelson and Vickrey maintained that the risk profile of the directly impacted 1964; Arrow and Lind 1970). group, not the entire population, would be most rele- Because of the risk-neutral status of governments, they vant. Both of these examples highlight the same prin- arethenaturalentitiestowhichriskisoftenshiftedbysoci- ciple: if the risk of an investment is borne by a limited eties. In industrialized countries, the government sup- group, the risk-adversity of that group should domi- plies insurance for retirement. In many countries, medical nate the investment decision (Arrow and Lind 1970). care, particularly for the aged and indigent, is financed by As to the comparative advantage of the government the government (Priest 1996). In regard to natural hazard in diversifying risk through its portfolio strategy, a com- losses,thegovernmentisoftentheagenttoassumetherisk pelling argument can be made that this is a small advan- of loss for its citizens (Lewis and Murdock 1999). tage when dealing with correlated risk. The risk of loss 38 Building Safer Cities: The Future of Disaster Risk from catastrophes in smaller countries is correlated, since are weak (Ferranti and others 2000). The imperfections the catastrophe is likely to impact the entire country at of the financial markets severely limit their ability to the same time. The core of the portfolio strategy lies in diversify risk and reallocate financial resources during aggregating independent risks, whose error terms cancel times of distress. Another factor of economic recovery out increasing predictive ability. This does not occur is the political will of the country to reallocate costs of when aggregating highly variant risks (Priest 1996). adjustment programs within the domestic economy Several arguments emerge as to why developing coun- (Rodrik 1998). In a study examining differences in recov- tries should be risk-averse. For many, the large size of ery from the recent debt crises in East Asia and Latin the potential loss in absolute terms compared to their America, Rodrik found that countries with the ability to internal risk-spreading ability means that risk does not distribute the cost of risk with few political repercus- approach zero in the hands of the population. Honduras sions were able to recover relatively quickly. These coun- presents such an example: in 1998, Hurricane Mitch tries tended to be located in East Asia. However, countries caused total losses approaching $5 billion in a country that lacked the ability to allocate the cost of risk with- with a total population of 4 million. Seventy percent of out considerable political turmoil took long periods of the population earns less than $2 per day (World Bank time to recover from external shocks to their economies. 1999). The average per capita cost of $1,250 per person, These countries tended to be in Latin America. In the therefore, is not insignificant, and the cost of risk does meantime, economic growth within these slowly recov- not begin to approach zero for Honduras or countries ering countries was significantly curtailed (Ferranti and in similar circumstances. others 2000). The importance of this observation is The mere calculation of per capita cost does not reflect that risk aversion at a government level should be influ- the difficulty countries have in increasing internal tax enced as much by the capacity of the country to allo- revenues. Many have only a limited ability to spread the cate risk within society as by the relative size of the risk. cost of risk internally through taxation (Rodrik 1998). Since many countries have difficulty reallocating inter- Many countries rely on a system of indirect taxes, the nal resources, reliance on external debt is often the pre- increase of which has political implications for specific ferred tool to address the need for additional resources segments of society. For these countries, the budgetary (Cornelius 2000). Additional external debt may have adjustment process is difficult and politically costly the least internal political cost. (Lewis and Murdock 1999). As a result, the process of As described earlier, it is not clear that governments spreading even small dollar amounts of risk has high in smaller countries can efficiently spread highly corre- political costs (Meier 1995). In addition, shallow lated risk. Since natural disasters tend to recur within financial markets and weak financial systems limit the geographically defined regions and can impact signifi- capacity to access internal and commercial external sav- cant portions of smaller countries, no benefit arises from ings in times of catastrophes (Ferranti and others 2000). aggregation of risk at the country level (Priest 1996). Therefore, even if the cost of risk approached zero in the This analysis begins to define investment decisions where hands of each taxpayer, the barriers to transferring the the traditional assumption of government risk neutral- risk to each taxpayer could be high in many countries. ity should be challenged. The first group of decisions This raises a critical issue when examining risk spread- includes those for which the risk of loss relative to the ing. The issue for some governments is not only the capacity of the population to absorb the risk is high. absolute size of the risk; it is also the relative ability of Honduras is a good example of this problem: the risk a country to dedicate resources to dealing with the risk. of loss on a per capita basis is very high. In examining external shocks to economies, a growing Risk analysis is different for another group of coun- body of research is emerging to describe why some coun- tries that includes those where the cost of catastrophe tries recover from shocks better than others. These stud- losses per capita is small, but the ability to shift the ies relate primarily to credit and commodity price shocks cost of risk to the population is limited. These are coun- (Cornelius 2000). A core factor is that financial markets tries with fragile taxation systems and those with weak in many countries remain shallow, and financial systems democratic institutions that lack the power to impose Natural Hazard Risk and Privatization 39 costs on entrenched power elites within the country. for the Inter-American Development Bank, a technique As described earlier, Rodrik has identified a group of that measures the capacity of a government to finance countries that lacked the ability to institute required probabilistic losses from natural hazard events was devel- political change to adjust to noncatastrophe external oped. For the Dominican Republic, El Salvador, Bolivia, stocks, despite the considerable costs borne by these and Colombia a "resource gap" was developed. A resource countries due to a lack of economic growth. The coun- gap is a measurement of a country's ability to finance tries he identified, primarily in Latin America and Africa, its reconstruction obligations following a disaster. would likely lack the political will to allocate the risk Calculation of the resource gap requires the following of loss internally through taxation. The World Bank computations: has also developed an index that describes countries · Country risk from natural hazard losses. The risk is with imperfect financial markets that tend to magnify a function of the probability of hazards of different rather than absorb the cost of external shocks (Ferranti magnitudes impacting a country and the vulnera- and others 2000). A third situation arises when invest- bility to loss of the potentially exposed population ments target the poor. If a specific group benefits from and assets. government investment, the risk profile of that group · The financial responsibility of the government to should dictate how risk is handled. finance country losses. Primary losses from disas- Determining when the government should assume ters may be the responsibility of various parties in risk associated with past investment decisions has a addition to the government: industry, businesses, direct bearing on which risks should be assumed or homeowners, and individuals. For this analysis, we transferred by governments in the privatization process. are concerned about the portion of loss borne by As noted earlier, the risk-neutral status of governments the government. in industrialized countries leads them to retain natural · The capacity of the government to meet its financial hazard risk in the privatization process. The issue of obligations. To the extent that a government lacks concern for developing countries, particularly for those the resources to fund its obligations, there is a nat- with high catastrophe exposure, is whether the same ural-hazard-resource gap. The required resources theory holds true. may come from international aid, government rev- enues (taxes), reserves, insurance proceeds, bor- rowing, and the diversion of resources from other Capacity of Countries to Absorb Natural programs. Hazard Risk A natural-hazard-resource gap articulates the ability of a government to meet the needs of financing post-dis- Key principles in the privatization process are the aster reconstruction. For countries with a resource identification and allocation of risks to the parties best gap, it means that significant costs to meet the risk of able to cover them.4 With respect to infrastructure loss to natural hazard risk will be incurred. projects, a wide variety of risks, including risks during construction, projected use of new infrastructure, the Resource Gap willingness of people to pay to use the infrastructure, and the maintenance and ongoing operations required, A natural-hazard-resource gap is developed for each of must be addressed. As discussed earlier, another risk is the case study countries. The resource gap is a meas- the partial or complete destruction of a project by a urement of the inability of a country to finance its recon- natural hazard event. Since this risk may be large, and struction obligations after a disaster. The measurement such losses are often difficult to predict, it is a risk gen- of the resource gap requires the calculation of a catas- erally retained by governments. trophe exposure for each country. Catastrophe expo- Increasingly, research has measured the risk of loss sure is determined by combining hazard and vulnerability from natural hazards in developing countries and the estimates for each country. The calculated catastrophe capacity of countries to manage that risk. In a recent study exposure estimates are presented in table 2.1: 40 Building Safer Cities: The Future of Disaster Risk Table 2.1 Catastrophe exposure in case study countries These estimates represent government responsibility Country 20-year event 50-year event 100-year event for reconstruction of government-owned property, as Bolivia 200 600 1,000 well as the assumption of risk for private housing, agri- Colombia 2,000 5,000 8,000 culture, and programs targeting the poor during post- Dominican 1,250 3,000 6,000 disaster periods. Republic Once an estimate of future financing needs has been El Salvador 900 3,000 4,500 determined, the next question addressed is the ability All values in millions of U.S. dollars. of the government to meet those needs. Is there a gap between the probabilistically determined resources and For example, Bolivia can expect direct losses to cap- the ability of the government to fund the required ital stock of at least $200 million approximately every resources? The potential difference is a resource gap. A 20 years; more specifically, there is a 1-in-20 chance resource gap is calculated by comparing a government's every year that there will be a catastrophic event equal- probabilistic or contingent need for reconstruction funds ing or exceeding $200 million in losses. Likewise, in the current year with its anticipated access to inter- there is a 1-in-50 chance, or 2 percent probability, every nal and external funds. Table 2.3 shows the estimated year of at least $600 million in direct losses. The mag- resource gap for the four countries. nitude of that figure could double, however, if indirect The resource gap for each country depends on crit- losses from lost production, tourism, and other serv- ical assumptions regarding the ability to access inter- ices were included. nal and external resources. For example, the resource Because the capacity of governments to finance gap in Colombia is primarily affected by the ability to obligations after a disaster is often limited, it is essen- raise funds through taxation. Historically, Colombia has tial to know the responsibility of the government for a raised internal taxes as a major tool in financing natural country's catastrophe exposure. Generally, two broad disaster losses. It is assumed that Colombia could raise categories of governmental responsibility can be defined: taxes by an additional $1.5 billion, if necessary. If it has risk to government-owned property and the risk a the ability to raise taxes by $2 billion, the resource gap government assumes from others. In the former, the for the 100-year event would disappear. The resource risk of loss is to government buildings, schools, and gap also depends on a series of assumptions regarding hospitals, and infrastructure such as roads, bridges, and future financing sources. The report details all assump- airports. The second category is the risk that the gov- tions used and the source of the data. The chart raises ernment assumes from others. This generally includes this question: How are countries able to finance their the risk to homeowners, agriculture, local and provin- probabilistic losses from natural hazards? The resource cial governments, and the poor. gap provides the basis for evaluating whether a coun- It is estimated that the government will finance 50 try can efficiently absorb losses to infrastructure. In some percent of the losses in the four countries. Table 2.2 cases, transferring the risk of loss from natural hazards (which is 50 percent of table 2.1) shows the share of to the private market as a component of the privatiza- losses borne by the government: tion process might be the best option for countries with a high resource gap. Before addressing this option, however, the variations Table 2.2 Government financing needs in case study in the resource gap among countries must be understood. countries According to this analysis, Bolivia can anticipate no Country 20-year event 50-year event 100-year event resource gap over the range of 20-, 50-, and 100-year Bolivia 100 300 500 events, although Bolivia is the poorest country in South Colombia 1,000 2,500 4,000 America and would be expected to be the most vul- Dominican 625 1,500 3,000 Republic nerable. Its level of hazard risk, however, is so low that El Salvador 450 1,500 2,250 it should have sufficient resources (assuming substan- All values in millions of U.S. dollars. tially increased borrowing) to respond. Colombia, by Natural Hazard Risk and Privatization 41 Table 2.3 Resource gap in case study countries contrast, has a very high natural hazard risk, but per 20-year recurrence capita incomes are high and risks are geographically Bolivia Colombia Dom Rep El Salv diverse. The government should be able to absorb the cost of disasters until it reaches a 100-year event. Even Direct damages 200 2000 1250 900 Gov responsibility 1000 1000 625 450 then, its resource gap can be covered if the govern- Aid 2 17 11 8 ment is able to raise taxes. Alternatively, El Salvador Insurancce 5 50 31 23 and the Dominican Republic can anticipate resource Payments gaps, given their catastrophic risk exposure. Both coun- Budget realloc 250 1500 500 250 New taxes 0 500 0 90 tries are small and have limited geographic diversity Domestic credit 100 0 150 0 with respect to risk, a high exposure to large-scale nat- External credit 100 100 100 100 ural disasters, and limited financial resources. For these IDB/WB External credit 0 0 800 800 two countries, there is at least a 1-in-100 chance of being market struck by an event that outstrips their ability to raise Resource gap none none none none post-disaster reconstruction funds. Resource gap none none none none The calculation of a resource gap for countries is the w/o IDB/WB Additional debt 0 0 83 80 beginning of a process. The resource gap identifies possible sources of financing for losses from natural haz- 50-year recurrence ards, but the analysis does not quantify the cost of access- Bolivia Colombia Dom Rep El Salv ing those resources. As discussed earlier, accessing Direct damages 600 5000 3000 3000 available resources has a cost. There are political costs Gov responsibility 300 2500 1500 1500 Aid 5 43 26 26 to raising taxes and diverting budgetary allocations. The Insurancce 15 125 75 75 use of increased debt absorbs borrowing capacity that Payments may be better used for other purposes. The borrowing Budget realloc 250 1500 500 250 gap calculation frames the issue so that a determina- New taxes 0 1000 0 180 Domestic credit 100 0 150 0 tion of whether it is more efficient for a government to External credit 200 200 200 200 retain or transfer risk of natural hazard losses as a com- IDB/WB ponent of the privatization process can be made. External credit 0 0 800 800 market Resource gap none none none none Resource gap none none none 169 w/o IDB/WB Conclusions and Future Research Additional debt 30 0 899 969 100-year recurrence The need to expand the provision of infrastructure in Bolivia Colombia Dom Rep El Salv developing countries is clear. The use of privatization Direct damages 1000 8000 6000 4500 as a tool to assist in the extension of infrastructure to Gov responsibility 500 4000 3000 2250 the poor is a priority of the international financial Aid 9 69 52 39 community. While myriad issues are associated with Insurancce 25 200 150 113 Payments privatization, defining and allocating risk as a compo- Budget realloc 250 1500 500 250 nent of the privatization process to determine the most New taxes 0 1500 0 270 cost-effective allocation is especially difficult. If risks are Domestic credit 100 0 150 0 large and difficult to control, government retention of External credit 200 200 200 200 IDB/WB the risk might prove the best option. These risks External credit 0 0 800 800 would be extremely expensive to shift to the private market sector and, in the worst case, could prove to be a "deal Resource gap none 531 1148 579 killer." Justification for government assumption of risk Resource gap 116 731 1348 779 w/o IDB/WB is that governments are best able to handle unknown Additional debt 216 200 1150 1000 risk through their power of taxation. The ability to spread 42 Building Safer Cities: The Future of Disaster Risk risk to taxpayers is an enormous efficiency advantage Research on the proper allocation and financing of of governments. natural and man-made hazard risk for developing coun- Circumstances exist where it may be more efficient tries is in the early stages. For many countries, infor- for the risk of natural hazard losses be shifted to the mation about the level of risk to infrastructure has not market as a component of privatization. The main cir- been systematically developed, although techniques to cumstances are: make the necessary calculations are well understood. · The project benefits a limited geographic area. The concept of "risk aversion" for governments is not · The project assists a limited population group, such well understood, either. As globalization continues, it as the poor. may be that the market (including large international · The size of the risk is larger than the capacity of the corporations) is much more efficient in coping with nat- government to shift the risk efficiently to the popu- ural hazard risk. If so, it makes the most sense for that lation. For those countries with a resource gap, it risk to be assumed by the market as a component of may not be possible for them to finance the natural the privatization process. hazard risk. Specifically, the following issues need to be addressed: · Countries without a resource gap may have institu- · The exposure of an infrastructure project to natural tional barriers that prevent them from shifting risk hazard risk. to their populations. For countries lacking the abil- · The capacity of a country to absorb the risk. ity to develop institutional compromise, the political · An evaluation of infrastructure projects to determine costs of financing post-disaster reconstruction may their primary beneficiaries and an analysis of the risk make it more desirable to include risk as a compo- tolerance of that particular group. nent of the privatization process. This may be the · The cost of the assumption of the natural hazard case for a number of Latin American countries. risk by the market on a project-by-project basis. In these instances, the cost of transferring risk to the A number of these questions are novel in the con- market should be considered as a component of priva- text of developing countries. With increased losses to tization. It may be that the cost of assigning the risk to infrastructure from natural hazard events, the efficient the market is too high and the risk must be absorbed assumption of risk will be increasingly important. by the government, but natural hazard risk should not be placed automatically, as is currently done, in those categories of risk best left with the government. As the Notes losses to infrastructure continue to escalate, the ability 1. In 1998, the United States Department of Energy issued to shift natural hazard risk to the market will become two sets of guidelines for privatization projects that remain increasingly important. benchmarks for work in privatization. The first is a Program- The analysis presented in this paper also applies to Project Manager's Privatization Guide and the second is a Pri- other types of risk that may arise from the privatization vatization Cost Estimating Guide. These documents can be process. In Eastern Europe, for example, the privatiza- found on the website for the Department of Energy (http:// tion of state-owned manufacturing enterprises is ongo- www.em.doe.gov/ private/projmangu.html). The Organization ing. Manyofthesefacilitiesinvolvechemicalandhazardous of Economic Cooperation and Development (OECD) has a waste risk from prior operations. Mechanisms such as Working Group on Privatization and Governance of State- environmental insurance exist for market assumption owned Enterprises with a comprehensive list of documents of private enterprises and the risk associated with con- related to privatization (http://www.oecd.org/EN/document/0, tamination created from prior business operations, EN-document-80-3-no-20-19549-80,00.html). As relates to including chemical facilities (Freeman and Kunreuther work in developing and emerging economies, the Asian Devel- 1997). It may be that risks from prior hazardous opment Bank has a Public-Private Infrastructure Advisory Facil- chemical operations may also be more efficiently han- ity(http://www.adb.org/Documents/ADBBO/RETA/35078012.ASP) dled by the transfer of the risk as a component of the that has examined best practices for privatization in a number privatization process. of industries. Natural Hazard Risk and Privatization 43 2. Paul K. Freeman, "Natural Disasters in Developing Countries: Cornelius, Peter. K. 2000. "Reforming the Public Sector's Risk Vulnerability from Increasing Population Concentrations," in Management Emerging Markets." Development Discussion Encyclopedia of Population (forthcoming). Paper No. 751. Boston: Harvard Institute for International 3. Earthquake risk lies along well-defined seismic zones that incor- Development. porate a large number of developing countries. High-risk areas ECLAC (Economic Commission for Latin America and the include Turkey, Pakistan, Afghanistan, India, China, Indonesia, and Caribbean). 1999. Manual for Estimating Socio-economic Effects the west coasts of North, Central, and South America. The pattern of Natural Disasters. United Nations International Decade for of hurricanes in the Caribbean and typhoons in South Asia, Natural Disaster Reduction. New York: United Nations. Southeast Asia, and the South Pacific is well established. Floods Ferranti, D., and others. 2000. Securing Our Future in a Global occur on 1 percent of the worldwide landmass. (Swiss Re 1997). Economy. Washington, D.C.: World Bank. 4. The discussion in this section is based on work done for the Freeman, Paul K., and Howard Kunreuther. 1997. Managing Regional Policy Dialogue of the Inter-American Development Environmental Risk with Insurance. Boston: Kluwer Academic Bank on Natural Hazard Risk. The resource gap described in this Publishers. section is based on methodology prepared by Leslie Martin and Freeman, Paul K., Leslie A. Martin, Reinhard Mechler, Koko described in greater detail in Freeman and Martin, "National Sys- Warner, and Peter Hausmann. 2001. Catastrophes and Devel- tems for Comprehensive Disaster Management: Financing Recon- opment: Integrating Natural Catastrophes into Development struction," May 1, 2002. The paper and methodology can be Planning. Disaster Risk Management Working Papers Series found at the Regional Policy Dialogue website: (http://www.iadb. No. 4. The Disaster Management Facility, World Bank, org/int/drp/). Washington, D.C. IFRC (International Federation of Red Cross and Red Crescent Societies). 2002. World Disaster Report 2001: Focus on Recov- Bibliography ery. Bloomfield, CT: Kumarian Press Inc. accessed June 2002. asters with Special Reference to Developing Countries. Oxford, Lewis, C.M., and A. Mody. 1998. Contingent Liability for Infra- U.K.: Clarendon Press. structure Projects. Washington, D.C.: World Bank. Arrow, K. J. 1992. "Insurance, Risk and Resource Allocation." Lewis, C.M., and K.C. Murdock. 1999. "Alternative Means of In G. Dionne and S.E. Harrington, eds., Foundations of Insur- Redistributing Catastrophic Risk in a National Risk Man- ance Economics: Readings in Economics and Finance. Boston: agement System." In K.A. Froot, ed., The Financing of Cat- Kluwer. astrophic Risk. Chicago: National Bureau of Economic Arrow, K. J., and R.C. Lind. 1970. "Uncertainty and the Evalu- Research. ation of Public Investment Decisions." The American Eco- Meier, G.M. 1995. Leading Issues in Economic Development. New nomic Review Vol. 60: 364­378. York: Oxford. Bendimerad, Fouad. 2000. "Megacities, Megarisk." The Disaster Mitchell, James K., ed. 1999. Crucibles of Hazard: Mega-Cities and Management Facility, World Bank, Washington, D.C. Available DisastersinTransition.NewYork:UnitedNationsUniversityPress. from accessed June 2002. Review: Natural Catastrophes 2001. Munich: Munich Re. Benson, C. and E. Clay. 2000. "Developing Countries and the Priest, G.L. 1996. "The Government, the Market, and the Prob- Economic Impact of Natural Disasters." In Alcira Kreimer and lem of Catastrophic Loss." Journal of Risk and Uncertainty No. M. Arnold, eds., Managing Disaster Risk in Emerging Economies. 12: 219­237. Washington, D.C.: World Bank. PSDS (Private Sector Development Strategy). 2002. "Private Sector Blakie, Piers, Terry Cannon, Ian Davis, and Ben Wisner. 1994. Development." World Bank, Washington, D.C. Available from At Risk: Natural Hazards, People's Vulnerability, and Disasters. accessed June 2002. New York: Routledge. Rodrik, D. 1998. "Where Did all the Growth Go? External Shocks, Charveriat, Celine. 2000. Natural Disasters in Latin America and Social Conflict and Economic Growth." Available from the Caribbean: An Overview of Risk. Inter-American Develop- http://ksghome.harvard.edu/~.drodrik.academic.ksg/ ment Bank: Washington, D.C. publications.html. 44 Building Safer Cities: The Future of Disaster Risk Samuelson, Paul, and W. Vickrey. 1964. "Discussion." American from accessed June Economic Review Proceedings May(59):88­96. 2002. UNISDR (United Nations International Strategy for Disaster U.S. Bureau of the Census. 1998. "World Population Profile." Reduction). 2001. Natural Disasters and Sustainable Develop- U.S.CensusBureauOnline.Availablefrom accessed June 2002. ment and Natural Disasters. Working Paper No. 5. New York, World Bank.1999. World Development Indicators. Washington, N.Y.: Department of Economic and Social Affairs. D.C.: World Bank. UNFPA (United Nations Population Fund). 1999. "The State World Bank. 2001. World Development Report 2000/2001: Attack- of World Population 1999." UNFPA Online. Available ing Poverty. New York: Oxford. Chapter 3 Natural Disaster Risk and Cost-Benefit Analysis Reinhard Mechler Natural Disaster Risk and Development CBA and Natural Disaster Risk Natural disasters constitute a serious challenge, par- Natural Disaster Risk and Impacts of Disasters ticularly for a number of developing countries where Natural disaster risk can be defined as the following: the impacts of these disasters are substantially larger than they are in more developed countries, due to a The exposure or the chance of loss (of lives, persons injured, typically higher degree of vulnerability. Factors con- property damaged and economic activity disrupted) due tributing to increased vulnerability comprise wide- to a particular hazard for a given area and reference period. It may be expressed mathematically as the probability that spread poverty, high unemployment, distributional a hazard impact will occur multiplied by the consequences inequalities, high population growth, and lack of of that impact (ADPC 2000). strong national and local institutions for dealing with disasters (Smith 1996: 42­46; Anderson 1995: 45; In case of a disaster event, the following effects may ECLAC/IDB 2000: 1). However, natural disaster risk occur: humanitarian effects, including the loss of life is often insufficiently accounted for in decisionmak- and persons injured; ecological effects among other ing. A major decisionmaking tool commonly used in damage to ecosystems; and economic effects, the focus the economic and financial evaluation of public invest- of this paper, comprising different effects on the econ- ments is cost-benefit analysis (CBA). In the context of omy and grouped into three categories--direct, indi- natural disaster risk, CBA is not used sufficiently. rect, and macroeconomic costs (see figure 3.1). This risk is often neglected in CBA assessments of Direct losses describe physical impacts on capital investment projects, risk management measures to stock such as infrastructure, machinery, and buildings. reduce natural disaster risk are often not assessed by They can be caused by the disaster itself or by follow-on CBA, and risk is commonly represented by average values only. Figure 3.1 Impacts of natural disasters This paper will analyze these shortcomings and their consequences, focusing on the economic impacts of dis- asters. The next section discusses the impacts of natu- Direct Loss of capital stock ral disasters and the basic elements of CBA. Then the incorporation of natural disaster risk into CBA is Economic costs Indirect examined, followed by a review of the current short- e.g., business interruption comings in using CBA in the context of natural disas- ter risk and resulting consequences. A short case study Disaster Humanitarian effects Macroeconomic impacts illustrates some of the issues discussed. The final sec- e.g., loss of GDP tion summarizes the findings and provides recom- Ecological effects mendations for using cost-benefit analysis in assessing natural disaster risk. 45 46 Building Safer Cities: The Future of Disaster Risk physical destruction (e.g., through fires). Indirect losses generated by the project compared to the situation occur as a consequence of these direct stock losses and without the project. In relation to natural disaster risk, include production and wage losses due to business inter- additional benefits arise from the savings in terms of ruption. Macroeconomic or secondary impacts comprise avoided direct, indirect, and macroeconomic costs as the aggregate impacts on economic variables like gross well as the reduction in variability of project outcomes. domestic product (GDP), consumption, and inflation There are several limitations to CBA. One is the dif- due to the effects of disasters as well as to the realloca- ficulty of assessing nonmarket values. Although meth- tion of government resources to relief and reconstruc- ods exist, this involves making difficult ethical decisions, tion efforts. Because macroeconomic effects reflect indirect particularly regarding the value of human life: CBA damage as well as the relief and restoration effort, these should probably not be used for this purpose. Another effects cannot simply be added up without causing dupli- issue is the lack of accounting for the distribution of cation (Otero and Marti 1995: 16­18). A substantial benefits and costs in CBA.1 The general principle under- number of studies on the assessment of these costs exist. lying CBA is the Kaldor-Hicks-Criterion, which holds A useful source of information is a manual that assesses that those benefiting from a specific project should the various costs of natural disasters developed by the potentially be able to compensate those who are dis- Economic Commission for Latin America and the advantaged by it (Dasgupta and Pearce 1978: 57). Caribbean (ECLAC 1999). Whether compensation is done in practice, however, is often not of importance. Another issue is the ques- Elements of CBA tion of discounting. Applying high discount rates expresses a strong preference for the present while poten- Cost-benefit analysis (CBA) is an economic technique typ- tially shifting large burdens to future generations. ically used to organize data, present costs and benefits, However, when keeping these limitations in mind, and finally estimate the cost-efficiency of projects like CBA is a useful tool whose main strength is an explicit building new infrastructure, which are undertaken by gov- and rigorous accounting framework for systematic cost- ernments and public institutions to increase public wel- efficiency decisionmaking. CBA provides a common fare(Koppandothers1997:53).CBAmeasuresthechange yardstick with a money metric against which to meas- with and without the specific project. In essence, it com- ure projects for social improvement (Kopp and others pares the costs of a planned project with its benefits and 1997: 53). It is a fact that economic efficiency is recommends its adoption when benefits exceed costs. important to many (government) decisionmakers. For There are three steps to CBA. First, costs and bene- example the United States has "at times dominated the fits must be identified and estimated. If these are given policy debate on natural hazards" (Burby 1991: 154). in physical terms, monetary values need to be assigned However, CBA should not be the sole criterion for (there are several methods for doing this). Next, future evaluating policies, but should be complemented by costs and benefits need to be discounted to render cur- other, noneconomic considerations. rent and future effects comparable. Last, costs and benefits are compared under a decision criterion to assess whether benefits exceed costs. Several criteria exist: Incorporation of Natural Disaster Risk into CBA the main ones are the net present value, the cost-ben- efit ratio, and the internal rate of return (Zerbe and Cost-efficiency evaluations by means of CBA are under- Dively 1994: 177; Dasgupta and Pearce 1978: 165). taken in the context of uncertainty which, when it can The costs in a CBA are the specific costs of conducting be measured probabilistically, is called risk, or "meas- a project. First there are the financial costs, the mone- urable uncertainty," according to Knight (compare with tary amount that has to be spent for the project. How- Brent 1998: 206). ever, of greater interest are the opportunity costs of using A number of methods for CBA in the context of risk these funds--the benefits foregone from not being able have been suggested; these are well described in Kramer to dispose of these funds for other important objectives. (1995: 61­76). There are basically two ways to include Usually, the benefits in a CBA are the additional outcomes risk in project analysis: the limited-information approach Natural Disaster Risk and Cost-Benefit Analysis 47 and the probability-based approach. When no specific For a project evaluator, it may be important to exam- information or only partial information on natural haz- ine the probability of net benefits becoming negative, ards and their impacts is available, limited-information i.e., to determine how marginal a project is. Consequently, approaches are used. Among these, sensitivity analysis is if marginality is likely, a decision to abandon the project often used, where important variables are varied in an or site it elsewhere, where hazard exposure is lower, or ad-hoc fashion to study the sensitivity of outcomes to these include risk management components into the project variations. Although natural disasters are rare events may be necessary. and, thus, abundant information on them does not exist, These risk management measures, or "secondary proj- data and software tools that can conduct a fuller proba- ects," protect a primary project's outcomes (Brent 1998: bilistic analysis provide more insight than the limited- 220). Benefits of these projects are the savings in terms informationapproachesareincreasinglybecomingavailable. of damage avoided and the decrease in volatility in A probabilistic approach entails obtaining probabil- outcomes. Secondary projects may be mitigation proj- ity distributions on disaster events and linking them to ects that reduce risk, or risk transfer projects that cede major economic variables. With that information, a risk to other parties willing to accept it. probability distribution of the project outcomes (net benefits) can be generated (figure 3.2). Status of Application of CBA in the Context Assume that A is the projected net benefits (i.e., of Natural Disaster Risk benefits less costs) from a project before natural disas- ter risk has been included in the assessment. A is pos- Natural disaster risk is only one risk among several itive, so this project seems a worthwhile undertaking. (including, for example, exchange rate or commodity If a probabilistic project analysis is conducted and a price fluctuations) that must potentially be taken into probability distribution is estimated, the average out- account. When a risk is judged to be negligible, it may come--the expected value E--can be determined. In not have to be considered formally. In contrast, when the case of natural disaster risk, which is a purely down- a risk is found to be large, it needs to be accounted for side risk,2 the expected value will be lower than the properly, in order to allow an efficient allocation of originally projected deterministic value A that didn't resources to these projects. account for risk. However, the average outcome E rep- Natural disaster risk is often not considered suffi- resents only that value that over a certain time horizon ciently in CBA. When examining the relevant litera- will materialize on average. Actual outcomes may lie ture on CBA in the context of natural disaster risk, three along the whole range of A and B (here B is assumed to issues emerge: be the worst outcome); net benefits could be negative · Natural disaster risk is commonly not accounted for if a disaster destroys a significant part of the project in CBA for investments and primary projects. and only a few benefits materialize while project costs · Secondary (risk management) projects are rarely have accumulated already. assessed in a CBA framework. When this is done, the Figure 3.2 Project analysis under risk focus is on mitigation only. · Usually, risk is not included explicitly, but by averages. These issues are further analyzed in the following section. Probability Natural Disaster Risk Is Not Included in Evaluation of Investment Projects For a number of countries, natural disaster risk is a seri- B O E A ous risk. However, this source of risk usually is not - Net Benefits + sufficiently accounted for in developmental planning and appraisals of investment projects by governments 48 Building Safer Cities: The Future of Disaster Risk and multilateral finance institutions (MFIs) (Kramer causing large developmental impacts. Finally, if sufficient 1995: 62; OAS 1991: 8). Vermeiren and Stichter remark funds for reconstruction and relief are not available, in an assessment of the costs and benefits of mitigation serious negative long-term impacts on socioeconomic in the Caribbean: development may result (compare with Freeman and others 2002). Contributing to the precarious state of the infrastructure is the region's vulnerability to natural disasters--hurricanes Assistance by MFIs in their capacity as "reinsurers in particular --and the tendency of development decision- of last resort" is often sought post-disaster. These assis- makers, in the public as well as private sectors, to make tance needs are volatile and rising. For example, between decisions concerning major investment projects without 1980 and 1999, the World Bank, the world's largest due consideration of natural disaster risk (Vermeiren and multilateral lending institution, financed 102 post- Stichter 1998: 1). catastrophe reconstruction projects in 56 countries, As a consequence, the uncertainty in project bene- amounting to a total of about $7.5 billion (Gilbert and fits is thus not duly accounted for, which results in an Kreimer 1999: 1). In relation to World Bank lending, 3 inefficient allocation of resources. There is the possi- the total sum over this period amounted to 1.9 percent, bility that risky projects may be selected and that pro- with a range from 0.2 to 6.0 percent on an annual basis jected benefits and investment will be lost when a disaster (see figure 3.3). There was some recurrence: 22 countries occurs. Furthermore, investment funds often have to have had two or more lending operations over this period. be borrowed externally in developing countries if inter- These figures underestimate the World Bank's recon- nal resources are insufficient. In the case of a disaster, struction financing support, as they do not include funds this investment is lost and the capacity to service the reallocated from other operations. Over the course of recently accumulated debt in the future is not increased. the 1990s, the Inter-American Development Bank lent Also, funds for continuing these projects or rebuilding around $2 billion in post-disaster assistance, mainly lost assets often have to be diverted from other projects, to rebuild and rehabilitate damaged infrastructure Figure 3.3 World Bank post-disaster reconstruction loans in relation to total World Bank lending, 1980­99 7 6 5 lending Bank 4 orld W total 3 of centage 2 Per 1 0 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Year Note: Word Bank lending includes lending by the International Bank for Reconstruction and Development (IBRD) and the International Development Association (IDA). Sources: Gilbert and Kreimer 1999, World Bank 2001. Natural Disaster Risk and Cost-Benefit Analysis 49 (IDB 2000: 20). From 1996 to 2000 alone, it lent $1.5 Figure 3.4 Risk management of natural disasters billion to affected countries to help recover from disas- ters, which increased annual average disaster-related 1. Identification: lending by a factor of 10 compared to the average over catastrophe risk the previous 15 years (IDB 2000: 1). 2. Risk assessment: Also, the supply of donor assistance is becoming potential costs increasingly limited as the dwindling amounts of offi- 3. Risk control: cial development assistance (ODA) show.4 ODA assis- potential benefits Cycle tance (in constant 2000 U.S. dollar terms) decreased 4. Cost- Mitigation benefit from $69 billion in 1990 to $53 billion in 2000 analysis Residual risk (OECD 2000; 2001). Risk transfer Demand for the inclusion of natural disaster risk in project appraisal methodologies has increasingly been voiced (Kramer 1995: 62; OAS 1991: 53). Consider- ing natural disaster risk in project appraisal allows for measures should be considered. These may be mitiga- more careful selection and design of projects as well as tion measures that reduce risk (e.g., building a dam to the identification and development of secondary risk- prevent floods) or the transfer of risk to other parties management measures to protect the benefits of primary (most commonly by means of insurance). For the design projects. More careful project and development plan- and development of risk-control measures, cost-efficiency ning is called for when considering loss-increasing trends, considerations as conducted by means of CBA are (or such as increased urbanization and high population should be) an important element of such a risk man- growth in developing countries, which concentrate cru- agement process in light of the fact that resources for cial assets that may be at risk, and a possible increase risk management are generally in short supply. Thus, in the frequency and severity of natural disasters due it is crucial to optimally allocate available resources to to climate change. those measures where benefits are largest. Ideally, risk management measures are planned and assessed in con- junction with (main) investment projects and rou- CBA of Risk Management Projects Rarely Done tinely included in project appraisals, in the same manner Due to concerns of MFIs about increased spending for that environmental impact considerations are now reconstruction lending and disaster aid, ex-ante risk- included in such appraisals (Gilbert and Kreimer 1999: management measures--that is, measures planned and 44; Vermeiren and Stichter 1998: 8). undertaken before the occurrence of disaster events-- Since 1980, the World Bank has lent $6.5 billion for are increasingly being promoted: 96 projects that included at least one mitigation com- ponent (Gilbert and Kreimer 1999: 51­53). However, International aid and development funding agencies, as can be discerned from the fact that about half of the besides sharing consternation at delays, disruptions, and increased costs, have the strong view that wisely planned top clients for reconstruction borrowing from the World hazard and vulnerability reduction efforts and financing Bank do not appear among the main borrowers for measures taken before a catastrophe pay excellent dividends mitigation projects, ex-ante disaster risk management in reducing economic impacts. Mitigation expenditures measures cannot yet be said to have sufficiently per- are a tiny fraction of the funds spent on reconstruction in vaded development-planning activities. the aftermath of catastrophes (Pollner 2000: 44). In a similar vein, the necessity of estimating these Risk management can be broken down into four com- risk management dividends by means of CBA has only ponents (figure 3.4). been acknowledged in the past few years, and CBA is The first step is the identification of risks, followed by still not widely practiced for natural disaster risk man- an assessment of their potential impacts and magni- agement projects (Dedeurwaerdere 1998: 1ff.). When tudes. If a specific risk is considered large, risk control evaluating risk management measures by means of CBA, 50 Building Safer Cities: The Future of Disaster Risk in general the focus is only on mitigation (Kramer 1995; measured by average values in decisionmaking. The Dedeurwaerdere 1998). Also, indirect costs and bene- variability of outcomes is not considered of impor- fits are rarely included in a CBA (Tobin and Montz 1997: tance in this case, as governments are assumed to be 269), while secondary impacts are usually completely able to pool a large number of risks so that the aggre- neglected. As Gilbert and Kreimer (1999: 37) demand: gate cost of risk is negligible (diversification of risk) or "There is greater need for an explicit and transparent to spread risks over a large population base so that the estimate of all the costs and benefits of natural disaster cost of risk to the individual is negligible. management." Several assessments have demonstrated that risk man- [...] when the risks associated with a public investment agement measures can bring about significant benefits. are publicly borne, the total cost of risk-bearing is insignif- icant and, therefore, the government should ignore uncer- Dedeurwaerdere (1998) estimated the benefits of dif- tainty in evaluating public investments (Arrow and Lind ferent prevention measures undertaken against floods 1970: 366). and lahars in the Philippines and calculated benefits of 3.5 to 30 times the projects' costs.5 Vermeiren and Stichter Arrow and Lind do not argue in favor of completely (1998) calculated potential avoided losses of 2 to 4 times neglecting uncertainty; rather they argue for assessing mitigation costs had mitigation been undertaken when average/expected values only: "[...]the government should building infrastructure like ports and schools in Jamaica behave as an expected-value decision maker" (Arrow and Dominica. Both of these projects limited benefits and Lind 1970: 366) without accounting for volatility. to avoided direct losses in the appraisal; including However, there are a number of important qualifica- indirect benefits would have increased the efficiency of tions to this theorem, mostly applicable to developing these preventive investments. On a larger scale, it is esti- countries (table 3.1). mated that the $3.15 billion spent on flood control meas- The qualification of the narrow tax and financing ures in China over the last four decades of the twentieth resources base can be illustrated by looking at the impacts century have averted losses of about $12 billion. In addi- of the largest disasters in terms of direct economic losses tion, the World Bank and the U.S. Geological Survey that have occurred in the United States (Northridge have estimated that economic losses worldwide from Earthquake 1994) and Honduras (Hurricane Mitch natural disasters in the 1990s could have been reduced 1998) and comparing these losses to important eco- by $280 billion if $40 billion had been invested in pre- nomic indicators (table 3.2). ventive measures (Benson 1998: 12). For the United States, disaster risk on an aggregate level is not a significant risk. In absolute terms, the enor- mous loss of $45 billion from the Northridge earth- Risk Not Included Appropriately: Necessity quake amounted to only 0.6 percent of GDP and 2.9 of Accounting for Volatility of Natural percent of tax revenue. In Honduras, on the other hand, Disaster Impacts the losses due to Hurricane Mitch, the largest disaster In cases where probabilistic CBA assessments are con- so far, have had severe implications, and the resource ducted, risk is often not included appropriately since base for financing the losses was clearly overwhelmed. average values are used (cf. Kramer 1995: 61; Szekeres In terms of GDP, losses from this event amounted to 41 2000; Dedeurwaerdere 1998). The information about percent and, in terms of tax revenue, they amounted to costs and benefits contained in the probability distri- 292 percent (figure 3.5). Consequently, post-Mitch bution is compounded to the expected value measure. Honduras experienced significant aggregate economic Other information from probability distribution (as illus- impacts, with an economic recession in 1999 after years trated in figure 3.1) is not used. with a growing economy. Focusing only on averages can theoretically be jus- On the other hand, when examining only average tified by the Arrow and Lind theorem (Arrow and Lind annual losses in Honduras, the dimensions of this risk 1970), which postulates that governments are usually are concealed and the impression is provided that it risk-neutral, i.e., they can neglect risk other than that can be handled without major difficulties: the expected Natural Disaster Risk and Cost-Benefit Analysis 51 Table 3.1 Qualifications to applicability of risk neutrality­theorem Qualifications related to risk pooling Existence of few and large Usually, developing countries' governments undertake just a few large investment projects, a course of government projects action that will not result in a highly diversified portfolio of projects; thus risk pooling is not viable (Brent 1998: 217­218). Large local or regional Disaster risk is covariant risk: disasters usually will affect whole regions; thus there is loss correlation. consequences when assets are lost Qualifications related to risk spreading Narrow tax and financing In smaller developing countries, the tax base is often too narrow to spread risk sufficiently. Other resources base for financing potential government financing sources such as domestic credit or private sector lending used to losses of projects spread risk are generally very limited as well. Distributional impacts In developing countries, large distributional impacts may occur post-disaster when infrastructure projects whose prime goal is poverty reduction (e.g., through road or sanitation projects) are affected. The poor are the group most affected by a loss of infrastructure. Irreversibility If additional funds are not available to continue crucial projects or rebuild assets, there can be irreversible effects, such as on health service provision (Little and Mirrlees 1974: 320). shallow financial markets; and high indebtedness, Table 3.2 Disaster losses and availability of resources with little access to external finance. for spreading risk for the United States and Honduras · Small countries with few large infrastructural assets Honduras: and high geographical correlation between those U.S.: Honduras: Average Northridge Mitch annual loss assets. · Countries with concentrated economic activity cen- GDP (million $) 7,834,000 4,725 4,725 GDP/per capita $ 29,267 790 790 ters (e.g., large urban agglomerations) exposed to Aid/GDP -- 6.3% 6.3% natural hazards. Loss (million $) 45,181 1,946 45 When discussing the issue of risk-neutrality, the OAS Loss/per capita $ 168.8 325.3 7.5 (1991) argued more than a decade ago that: Loss/GDP 0.6% 41% 1.0% Loss/tax revenue 2.9% 292% 6.8% The reality of developing countries suggests otherwise. Loss/Gross Domestic 3.6% 189% 4.4% Savings Government decisions should be based on the opportu- Loss/Net domestic credit 0.7% 165% 3.8% nity costs to society of the resources invested in the proj- ect and on the loss of economic assets, functions and All values are in current 1997 U.S. dollars, and economic data refer to products. In view of the responsibility vested in the public 1997. Data sources: World Bank 2001; Freeman and others 2002; Münchener sector for the administration of scarce resources, and Rück 2000. considering issues such as fiscal debt, trade balances, income distribution, and a wide range of other economic and social and political concerns, governments should loss of $45 million constituted 1 percent of GDP and not act risk-neutral (OAS 1991: 40). 6.8 percent of tax revenue in 1997. Thus, it is neces- sary to look at the extremes rather than the averages, Adopting a risk-averse perspective and including the as disasters are by definition low-frequency, extreme- volatility of disaster risk in decisionmaking has important consequence events. Averages do not capture these char- implications for the evaluation of primary and risk man- acteristics well. agement projects. In the assessment of primary projects, The validity of the A&L theorem is generally restricted risk is more appropriately captured, and a more careful to more developed countries; a number of developing projectselectioncanbeconductedwhentheextreme-event countries should act in a risk-averse manner: characterofnaturaldisastersisproperlyaccountedfor.For · Countries subject to high natural-hazard exposure. secondary risk management measures there is increased · Countriessubjecttohigheconomicvulnerability--that benefitinconductingthoseevaluations,asbenefitsinterms is, those with low tax revenue; low domestic savings; of avoided impacts are higher. The latter point will be 52 Building Safer Cities: The Future of Disaster Risk Figure 3.5 Important indicators for ability to spread would be desirable to insure public assets against nat- disaster risk for Honduras and the United States ural disaster risk in Honduras (Mechler 2002). Hon- Natural disaster losses compared duras is one of the developing countries where insurance to economic indicators against natural disaster risk for public assets (infra- structure, schools, hospitals) has recently been proposed 300% (e.g., Pollner and Camara 2001) and is currently under 250% investigation. The author examined the cost-efficiency 200% of the government's purchasing insurance for the 150% entire portfolio of public assets from a foreign insurer. Costs in this analysis were the opportunity costs due 100% to the premium payment (less government funds 50% available for other important objectives). The benefits 0% arose due to receiving insurance indemnity payments USA-- Honduras-- Honduras-- post-disaster to be used for reconstructing critical public Loss/GDP revenue assets (roads, bridges) necessary for quick economic ax Savings credit Northridge M recovery. The desirability of this project was evaluated Loss/T average itch Domestic domestic oss with the change in GDP as the main indicator. Loss/Net loss Several strategies for insuring certain risk layers Loss/Gr (so-called excess of loss (XL) insurance) and their cost- illustrated by the following case study on the evalua- efficiency were studied while stochastically simulating tion of a risk transfer project in Honduras. the occurrence and the resulting economic impacts of storm, flood, and earthquake events, which are the major natural hazards in Honduras. Case Study on CBA of Risk Transfer in Honduras For example, when purchasing insurance in the form As discussed, the costs and benefits of transferring catas- of a risk layer stretching from the 50- to the 100-year trophe risk are rarely assessed, and existing analyses catastrophe events (i.e., all disaster events with a focus on mitigation; in addition, the macroeconomic return period between 50 and 100 years were insured), costs and benefits are usually not accounted for in CBA. projected GDP paths in Honduras over a time horizon To shed light on these issues, a recent prospective from 2002 to 2008 differed from the projections with- study undertaken by the author analyzed whether it out insurance, as illustrated in figure 3.6 and table 3.3. Figure 3.6 Projection of GDP paths with and without insurance of public assets in Honduras Insurance for public assets No Insurance 120,000 120,000 100,000 100,000 80,000 80,000 60,000 Lempira] 60,000 GDP Lempira] GDP 40,000 40,000 [Million [Million 20,000 20,000 - - 2002 2003 2004 2005 2006 2007 2008 2002 2003 2004 2005 2006 2007 2008 Year Year Note: Lempira is the currency unit in Honduras. One dollar is approximately 16.50 lempira. Natural Disaster Risk and Cost-Benefit Analysis 53 Table 3.3 Assessment of costs and benefits of insuring public assets in Honduras Difference GDP (million lempira) No insurance Insurance Difference: net benefit [% of no insurance case] Mean sum of GDP (2002­2008) 411,373 411,239 -134 -0.03% Standard deviation 18,378 13,999 -4,379 -23.83% Net benefits as measured by the change in the aver- If possible natural disaster risk should be incorpo- age sum of GDP over this period due to the insurance rated into CBA in a probabilistic manner (i.e., data on arrangement were slightly negative (134 million lem- probability distribution of disasters need to be obtained). pira, 0.03 percent of baseline GDP) compared to the Increasingly, the data and tools for such probabilistic no-insurance case (table 3.3).6 The fact that this aver- analyses are available. With these data at hand, proba- age outcome was negative (albeit only marginally) can bility distributions for important project outcomes can be explained by the fact that risk transfer--in contrast be assessed to study the impacts of the incorporation to mitigation--does not reduce risk but shifts risk to an of natural disaster risk on the viability of projects. insurer who in turn demands compensation in the form For new investment projects, accounting for risk in of an insurance premium payment every year.7 Due to CBA leads to more careful project selection and design, the large loss potentials characteristic of a disaster, decreasing potential losses when a disaster strikes. insurance premia for disaster risks generally exceed the Considering natural disaster risk also allows deter- annual average loss substantially (Pollner 2001: 21; Froot mination of the need for and efficiency of risk manage- 1999: 6­7); thus considerable opportunity costs accrue.8 ment (secondary) projects that secure the benefits of However, a major benefit of insurance in this analy- main investment projects. Benefits of these projects con- sis (and in general) was the reduction in volatility. The sist of savings in terms of disaster losses averted and a large volatility (vertical lines) around the expected values decrease in volatility of the primary project's outcomes. (horizontal lines) was reduced when undertaking insur- In principle, it would be desirable to integrate risk and ance. Volatility, as measured by the standard deviation, risk management measures into project evaluation, just decreased by about 24 percent. This is of importance as environmental impact analysis is nowadays routinely for risk-averse countries, where disaster impacts cannot conducted when appraising new investment projects. easily be absorbed and the stability of economic devel- Furthermore, a number of developing countries with opment is a major concern. When assessing such a risk high natural hazard exposure and a limited ability to transfer project, the benefits in terms of reduced volatil- cope with disaster impacts need to be risk-averse to nat- ity are important and need to be factored in, in addi- ural disaster risk. In these cases, natural disaster risk tion to the average outcomes. needs to be considered properly, and the volatility of projects' outcomes should be factored into the deci- Conclusions sion as to whether to conduct a proposed project. Par- ticularly for risk transfer measures, taking volatility of Although natural disaster risk is significant for a number project outcomes into account shows or increases the of developing countries, this risk is not sufficiently incor- desirability of such measures. porated into cost-efficiency evaluations as conducted by Cost-efficiency as measured by CBA should not be means of cost-benefit analysis. Keeping the limitations of the sole criterion when planning and assessing develop- CBA in mind (focusing on measurable, mostly economic ment and risk management projects, but it provides impor- project outcomes, distribution of costs and benefits not tant information for a more efficient, less risky allocation accounted for, discounting of future costs and benefits), of scarce funds and thus can aid in bringing about more undertaking CBA for investment and risk management robust development. CBA considerations should be an projects in the context of natural disaster risk improves integral element of decisionmaking in a "culture of pre- decisionmaking and the allocation of scarce resources to vention" and thus may contribute to creating more intan- the most profitable and least risky undertakings. gible, but probably even more important, benefits: 54 Building Safer Cities: The Future of Disaster Risk Prevention not only minimizes damage but promotes a Arrow, K. J., and R. C. Lind. 1970. "Uncertainty and the Eval- stable environment, incentives for investment and enter- uation of Public Investment Decisions." The American Eco- prises and the sense that people can control their own eco- nomic Review 60: 364­378. nomic destiny. These are crucial for sustainable long-term Benson, C. 1998. "The Cost of Disasters." In John Twigg, ed., development (Anderson 1991: 27). Development at Risk? Natural Disasters and the Third World. U.K. National Coordinated Committee for the International Decade Notes for Natural Disaster Reduction. Oxford: Oxford Centre for 1. Methods to account for the distribution of costs and benefits Disaster Studies: 8­13. have been proposed, but are not used in practice (Little and Brent, R. J. 1998. Cost-Benefit Analysis for Developing Countries. Mirrlees 1990: 358­62). Cheltenham: Edward Elgar. 2. As contrasted with speculative risk involving chances to gain Burby, R., ed. 1991. Sharing Environmental Risks. How to Control and lose at the same time. Governments' Losses in Natural Disasters. Boulder, CO: West- 3. All dollar amounts are U.S. dollars. The World Bank focuses view Press. on the financing of investment and productive assets for build- Dasgupta, A. K., and D. W. Pearce. 1978. Cost-Benefit Analysis: ing infrastructure and institutions fostering socioeconomic devel- Theory and Practice. London: Macmillan. opment rather than relief operations (Lester 1999: 179). Dedeurwaerdere, A. 1998. Cost-benefit Analysis for Natural Disas- 4. ODA encompasses all financial support given to developing ter Management - A Case-study in the Philippines. Brussels: CRED. countries in the form of loans and grants with a grant element Dinwiddy, C., and F. Teal. 1996. Principles of Cost-benefit Analysis of at least 25 percent in order to promote economic develop- forDevelopingCountries.Cambridge:CambridgeUniversityPress. ment and welfare. It is provided bilaterally by the more devel- ECLAC (Economic Commission for Latin America and the oped countries and by multilateral finance institutions. Caribbean). 1999. Manual for Estimating the Socio-economic 5. Lahars are volcanic mudflows. Effects of Natural Disasters. United Nations International Decade 6. GDP was discounted at a standard discount rate of 12 percent. for Natural Disaster Reduction. New York: United Nations. 7. Risk could also be transferred to the capital markets (e.g., by ECLAC/IDB. 2000. "A Matter of Development: How to Reduce means of catastrophe bonds). Vulnerability in the Face of Natural Disasters." Paper before 8. The insurance premium for catastrophe risk demanded by seminar: Confronting Natural Disasters: A Matter of Devel- an insurer is determined by the expected losses to the assets opment, New Orleans, March 25­26, 2000. insured (the average annual losses), a surcharge for expenses Freeman, P. K., L. Martin, R. Mechler, K. Warner, with P. Haus- and profit, and a risk-loading component accounting for the man. 2002. Catastrophes and Development, Integrating Natural high variability of catastrophe risk. Generally, the premium Catastrophes into Development Planning. Disaster Risk Man- charged is considerably higher than the expected losses. agement Working Paper Series No. 4. The Disaster Manage- ment Facility, World Bank, Washington, D.C. Froot, K., ed. 1999. The Financing of Catastrophe Risk. Chicago: Bibliography University of Chicago Press. ADPC (Asian Disaster Preparedness Centre). 2000. Workshop Gilbert, R. and A. Kreimer. 1999. Learning from the World Bank's on Urban Disaster Mitigation. University of Karlsruhe, Experience of Natural Disaster Related Assistance. Washington, 23­27 October 2000. D.C.: World Bank. Anderson, M. B. 1991. "Which Costs More: Prevention or Recov- Gramlich, E. M. 1981. Benefit-Cost Analysis of Government Pro- ery?" In Alcira Kreimer and M. Arnold, eds., Managing Nat- grams. Englewood Cliffs, New Jersey: Prentice-Hall. ural Disasters and the Environment. Washington, D.C.: World Hecker, G. 1995. "A Review of the Disaster-Related Activities of Bank: 17­27. the Asian Development Bank: An Economic Perspective." In Anderson, M. B. 1995. "Vulnerability to Disaster and Sustain- Mohan Munasinghe and C. Clarke, eds., Disaster Prevention able Development: A General Framework for Assessing Vul- for Sustainable Development: Economic and Policy Issues. Wash- nerability." In Mohan Munasinghe and C. Clarke, eds., Disaster ington, D.C.: World Bank: 77­85. Prevention for Sustainable Development: Economic and Policy IDB (Inter-American Development Bank). 2000. Facing the Chal- Issues. Washington, D.C.: World Bank: 41­59. lenge of Natural Disasters in Latin America and the Caribbean. Natural Disaster Risk and Cost-Benefit Analysis 55 An IDB Action Plan. Washington, D.C.: Inter-American OECD (Organization for Economic Cooperation and Develop- Development Bank. ment) and D. A. Committee 2000. Net ODA Flows in 2000. Kopp, R. J., A. Krupnick, and M. Toman. 1997. Cost-Benefit Paris: OECD. Available from http://webnet1.oecd.org/pdf/ Analysis and Regulatory Reform: An Assessment of the Science and M00005000/M00005137.pdf.Internet. the Art. Discussion Paper 97­19. Washington, D.C.: Resources OECD, and D. A. Committee 2001. Net ODA Flows from 1950 to for the Future. 1999. Paris: OECD. Available from http://webnet1.oecd.org/xls/ Kramer, R. A. 1995. "Advantages and Limitations of Benefit-Cost M00002000/M00002883.xls. Internet. Analysis for Evaluating Investments in Natural Disaster Mit- Otero, R. C., and R. Z. Marti 1995. "The Impacts of Natural igation." In Mohan Munasinghe and C. Clarke, eds., Disaster Disasters on Developing Economies: Implications for the Inter- Prevention for Sustainable Development: Economic and Policy national Development and Disaster Community." In Mohan Issues. Washington, D.C., World Bank: 61­76. Munasinghe and C. Clarke, eds., Disaster Prevention for Sus- Kreimer, A., and M. Arnold. 2000. "World Bank's Role in Reduc- tainable Development: Economic and Policy Issues. Washington, ing Impacts of Disasters." Natural Hazards Review 1 (February): D.C.: World Bank: 11­40. 37­42. Pollner, J. 2000. Managing Catastrophic Risks using Alternative Risk Lester, R. 1999. "The World Bank and Natural Catastrophe Financing and Insurance Pooling Mechanisms. Washington, D.C., Funding. The Changing Risk Landscape: Implications World Bank. for Insurance Risk Management." Proceedings of a confer- Pollner, J., and M. Camara. 2001. Honduras. Catastrophe Risk ence sponsored by Aon Group Australia Ltd., Sydney, Exposure of Public Assets. An Analysis of Financing Instruments Australia. for Smoothing Fiscal Volatility. Washington, D.C.: World Bank. Little, I. M. D., and J. A. Mirrlees. 1974. Project Appraisal and Reutlinger, S. 1970. Techniques for Project Appraisal under Planning for Developing Countries. London: Heinemann. Uncertainty. Washington, D.C.: World Bank. Little, I.M.D., and J. A. Mirrlees. 1990. "Project Appraisal and Smith, K. 1996. Environmental Hazards. Assessing Risk and Reduc- Planning Twenty Years On." In S. Fischer, ed. Proceedings of ing Disaster. London: Routledge. the World Bank Annual Conference on Development Economics. Szekeres, S. 2000. La consideracion de desastres naturales en la Washington, D.C.: 351­382. evaluacion de proyectos federales en Mexico. Analisis y recomen- Mechler, R. Forthcoming. Natural Disaster Risk Management daciones. Washington, D.C., World Bank. and Financing Disaster Losses in Developing Countries. Ph.D. Tobin, G., and B. Montz. 1997. Natural Hazards. Explanation and thesis. Karlsruhe: University of Karlsruhe. Integration. New York: The Guildford Press. Münchener Rück. 2000. Topics: Jahresrueckblick Naturkatas- Vermeiren, J., and S. Stichter. 1998. "Costs and Benefits of Hazard trophen 1999. Munich. Mitigation for Building and Infrastructure Development: A Munasinghe, M., and C. Clarke, eds. 1995. Disaster Prevention Case Study in Small Island Developing States." Paper before con- for Sustainable Development: Economic and Policy Issues. Yoko- ference of The International Emergency Management Society. hama: World Bank. World Bank. 2001. World Bank Development Indicators. Wash- OAS (Organization of American States). 1991. Primer on Nat- ington, D.C. ural Hazard Management in Integrated Regional Development Zerbe, R. O., and D. D. Dively. 1994. Benefit-Cost Analysis in Planning. Washington, D.C. Theory and Practice. New York: Harper Collins. Chapter 4 Globalization and Natural Disasters: An Integrative Risk Management Perspective Torben Juul Andersen The increased global exchange of merchandise, serv- economic impacts of natural disasters. While a com- ices, and capital is a key characteristic of the contem- prehensive analysis of all linkages is beyond the scope porary business environment. International firms position of this paper, three areas are explored: (1) the relation- themselves to gain competitive advantage from oppor- ship between natural catastrophes, economic develop- tunities offered in this larger and more open economic ment, and global competitiveness; (2) the relationship system. Similarly, governments reap rewards from firms between global trade and investment, economic growth, that establish operations within their borders. There are and sustainable competitive advantage; and (3) the strong arguments in favor of international trade, foreign role of proactive risk management and the potential direct investment, and globalization since they often benefits from global market access. promote economic growth. Though there can be sig- nificant variance in growth in developing countries, it is recognized that some economic trade and activities Natural Catastrophes, Economic Development, are more beneficial than others. Developing countries, and Global Competitiveness therefore, must position themselves to take advantage of opportunities in the global market. Natural catastrophes reflect the negative economic The frequency and severity of natural disasters have impacts on human settlements and productive assets increased markedly worldwide. Economic losses asso- from extreme natural phenomena such as windstorms, ciated with natural hazards are increasing exponentially flooding, and earthquakes. Direct economic losses from in developing countries, where local risk-transfer mar- natural catastrophes over the past decade exceeded $700 kets are generally weak. Hence, natural catastrophes billion1 (all amounts are in U.S. dollars). These losses have devastating socioeconomic consequences when are estimated to increase to a total amount of $6 to $10 they strike populated areas in less developed economies, trillion over the next 20 years, far beyond the growth where they are bound to have adverse impacts on the in aid and development programs (ICRC 2001). Over global competitiveness of exposed countries. Disasters the past ten years, natural catastrophes have caused have a negative impact on economic activity and the more than 800,000 deaths and affected the livelihoods associated economic uncertainties hamper investment of more than 2 billion people worldwide. Total reported in long-term commercial relationships. Conversely, par- losses from natural catastrophes, ranging from $30 to ticular types of economic activity and a truncated policy $190 billion annually, have averaged roughly $65 billion focus can increase a country's economic vulnerability annually.2 More than 60 percent of the reported eco- to natural disasters. These relationships need to be made nomic losses in recent years have related to events in more explicit and managed more effectively so devel- developing countries. Approximately half of the losses oping countries are not disadvantaged in the global in industrialized countries were covered by formal insur- market. ance contracts, while only some 5 percent of reported This paper incorporates perspectives from econom- damages in developing countries were covered.3 ics, finance, and strategic management and identifies Global catastrophic events seem to be occurring with several links between market globalization and the increased frequency. Over the past thirty years, the 57 58 Building Safer Cities: The Future of Disaster Risk number of reported catastrophes has quadrupled, and by a factor of 2,000-3,000 and total insured losses by several factors have resulted in increased economic expo- a factor of 1,000 (figure 4.1). The implied increase in sure to natural catastrophes. Climate patterns seem to economic losses associated with natural catastrophes be changing in ways that increase the frequency of cer- by far outweighs economic growth figures for the same tain natural events. For example, El Niño influences the period.4 The dramatic increase in direct economic losses intensity of storms, rainfall, floods, and landslides in much per hazard event points to the increasing significance of the world. At the same time, the population is grow- of catastrophe risks. If this trend continues unabated, ing and economic assets are being placed in areas more catastrophe risk exposure will seriously challenge the exposed to natural hazards (Kleindorfer and Kunreuther economic sustainability of developing countries that are 1999). This combination of higher hazard frequency exposed to natural catastrophes.5 and greater exposure of economic assets extends the The number of victims associated with natural catas- potential damage that can be inflicted by natural haz- trophes, as reflected in the numbers of dead and affected, ards. Though there is no indication that the frequency is heavily skewed toward developing countries. The of earthquakes is increasing, changing climatic condi- number of deaths, however, has fallen over the past 30 tions seem to be causing more frequent and severe wind- years, from 2 million during the 1970s to 800,000 during storm events. Hence, the combination of a burgeoning the 1990s. Hence, local risk mitigation and disaster relief world population, increasing urbanization, and an efforts may bear fruit, but the number of victims is still expanding economic asset base extend economic expo- large and most are related to events in developing coun- sure to natural catastrophes. tries. At the same time, the number of people affected Whereas event frequency has quadrupled over the by natural catastrophes has increased significantly from past thirty years, reported economic losses have increased 740 million in the 1970s to 2 billion in the 1990s. Figure 4.1 Development in reported and insured catastrophe losses, 1970­2001 (three-year moving averages) 3,500 3,000 2,500 2,000 1,500 1,000 500 1972 = 100 0 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Reported losses Insured losses Source: Centre for Research on the Epidemiology of Disasters (CRED) and natural catastrophe losses reported in various issues of Sigma (Swiss Re). Globalization and Natural Disasters: An Integrative Risk Management Perspective 59 Over the past decade, natural disasters have resulted affected by natural catastrophes and economic growth.9 in damage that has constituted 10 to15 percent of an This may appear counterintuitive, although in the absence exposed country's annual gross domestic product (GDP). of proactive risk management practices and effective These extreme situations usually apply to relatively small, risk-transfer markets, post-disaster financing is typi- vulnerable countries with less diversified economies. cally made available through emergency facilities Such effects can have a significant impact on economic extended by multilateral institutions and other foreign activity and the appropriation of public funds.6 Economic aid donors.10 International assistance is often prompted growth rates typically hover around 1 to 3 percent annu- when the number of victims is high. Since disaster- ally, so a direct-loss impact of 5 to10 percent of GDP related capital inflows have an economic impetus, they can have an abrupt effect on a country's economic devel- lead to a positive relationship between human devas- opment. Studies indicate that real GDP growth decreases tation and economic growth in exposed countries. in the year of the disaster and then increases the next This somewhat perverse relationship seems to indicate one to two years, as public and private reconstruction that the availability of international emergency sup- investment boost the growth rate (Charveriet 2000). In port and disaster financing shields the countries from many cases, post-disaster reconstruction efforts may actu- the adverse ex post economic impacts of natural disas- ally improve the quality of economic assets and lead to ters while too little is done to prevent the effects of the increased productivity (Albala-Bertrand 1993). There- catastrophes on an ex ante basis.11 Although unintended, fore, if exposed developing countries take precautionary the financing of catastrophe losses through international measures and establish disaster risk financing arrange- donations constitutes a powerful disincentive to imple- ments, they may be able to reinstate new, and hence more ment more proactive risk management practices that efficient, economic assets after major disasters. However, could help reduce the socioeconomic cost of natural dis- a sample of developing countries actively engaged in asters in exposed countries.12 Prevention is important global trade7 indicates that a high level of catastrophe in reducing human suffering from catastrophes, but losses is generally associated with lower economic growth8 political leadership often considers risk management (figure 4.2). efforts an unnecessary cost rather than an investment in Further analysis of the sample shows a positive rela- a better environment. Hence, there is a clear need to adopt tionship between the percentage of the population policy measures that effectively integrate risk assess- ment, risk mitigation, risk transfer, and emergency preparedness (Andersen and Masci 2001). Figure 4.2 The relationship between economic growth and catastrophe losses, 1990­2000 Since post-disaster economic recuperation in devel- oping countries is typically based on the availability of 9 China multilateral relief facilities and humanitarian aid, there 7 Argentina are few political incentives to adopt a more proactive 5 Malaysia risk management approach. This situation makes exposed Bangladesh 3 countries highly dependent on the international com- 1990­99 Zimbabwe munity as "lender-of-last-resort" to cope with the 1 worst economic impacts of natural disasters. This bailout Growth -1 Central Afr. Rep. hinders the development of disaster prevention and mit- Nicaragua Cameroon Jamaica GDP -3 igation measures, since leaders are not pressed to make Haiti -5 advance arrangements. By contrast, economic entities in Annual industrialized countries obtain insurance and alterna- -7 tive risk-transfer cover in financial markets to ease post- -9 disaster reconstruction efforts. This risk management 0 2 4 6 8 10 12 14 16 18 Losses/GDP 1990­99 approach would be beneficial to developing countries, Sources: Centre for Research on the Epidemiology of Disasters too. Risk management can reduce a country's vulnera- (CRED), World Factbook, and World Bank data. bility to catastrophe risks and secure reconstruction 60 Building Safer Cities: The Future of Disaster Risk funding that significantly lessens the economic sever- which constitutes an increasing share of economic activ- ity of natural catastrophes. ities, has been somewhat higher, at an annual rate of In the absence of an active risk management approach, 7.2 percent, while the volume of foreign direct invest- developing countries exposed to natural catastrophes ment has grown at the phenomenal rate of 17 percent are often forced to divert funds from existing develop- per year during the period.16 ment programs to fund temporary disaster relief efforts. The ability to exchange primary and manufactured This distorts commitments to longer-term economic goods as well as commercial services across borders investment. A country that has insufficient post-disaster has the potential to create economic net benefits because financing arrangements often faces delays in compen- it provides global market access for offerings that con- sating economic losses as governments await approval stitute comparative advantages.17 Increased global com- from multilateral credit facilities and other financing petition provides new opportunities for companies to sources. Furthermore, disaster relief in the form of bilat- improve customer service and increase economic effi- eral donations typically has conditions that limit the ciencies. The international mobility of capital can also uses of funds. Hence, a lack of risk management reduces funnel overseas financial resources to promising busi- the prospects for a more immediate economic recovery ness ventures and provide access to risk transfer arrange- after a disaster, particularly when a country's fiscal resources ments in the global financial markets. However, investors' are stretched and critical economic infrastructure has willingness to provide cross-border funding to economic been affected. Economic entities operating in economies activities in a country depends on the soundness and that are vulnerable to natural catastrophes have diffi- stability of a country's economic policies, since global culty establishing dependable, long-term business rela- investors are lured by promising returns with reason- tionships. If these essential stakeholder relationships able risk characteristics. are jeopardized by excessive catastrophe risk exposure, If the economic arguments for global trade have merit, a country may encounter difficulties in its attempts to the evidence should indicate that a higher level of support economic activities that have the potential to international trade is associated with economic devel- generate more sustainable competitive advantages.13 opment. For example, if a country is able to exploit Hence, effective management of catastrophe risk should comparative advantage in the global exchange of support competency-based economic activities, thereby goods and services, then a higher level of trade inter- increasing the potential economic benefits from inter- action should lead to higher economic growth in the national trade. The following section takes a closer look country. However, in a sample of developing countries at these relationships. with many international linkages, there seemed to be no clear relationship between the ratio of global trade and growth in GDP.18 Rather, the data seem to indicate Globalization, Economic Growth, and Sustainable a negative relationship between the level of trade and Competitive Advantage economic growth.19 These results do not fit with our simple international trade hypothesis. It is possible the As restrictions on cross-border transactions have eased discrepancy can be explained by the fact that large coun- in recent decades, the volume of global trade has tries with more diversified economic bases are better expanded faster than economic growth in the world hedged against the negative impacts, including natural economy.14 The annual compound growth rate in mer- disasters, of exogenous shocks to the economy. There chandise exports from industrialized countries has aver- seems to be a clear relationship between country size and aged 6 percent over the past 20 years.15 Certain countries, its concentration on specific export merchandise. Smaller including China, Thailand, Malaysia, Indonesia, and countries, for example, are generally more dependent the Philippines, have taken advantage of new global trade upon specific export products. This dependency may opportunities, while other developing economies, includ- make them more economically vulnerable to natural ing those in Sub-Saharan Africa, have displayed low catastrophes and other disasters. Hence, it does not seem growth rates. Overall growth in the export of services, to be trade volume in itself that matters, but rather the Globalization and Natural Disasters: An Integrative Risk Management Perspective 61 type and diversity of economic activities and global trade that have shown above-normal growth rates in global transactions a country pursues. demand.21 These product groups represent more skill It can also be argued that it is the trade policies pur- and technology-intensive merchandise with the poten- sued by developing countries that influence economic tial to achieve further productivity gains than relatively development. If a country has reduced its import tariffs, simple, labor-intensive manufactures and factor-based it reflects a general commitment to international trade primary goods. and global competition. When protective tariffs are Global capital flows can take place as investments in reduced, domestic economic entities are more exposed marketable financial assets or directly in productive to global competition, forcing these entities to improve assets through foreign direct investment (FDI). Invest- operational efficiencies to thrive and survive.20 An ment in domestic financial assets provides financial analysis of the country sample confirms that tariff poli- resources to local operators who maintain managerial cies in favor of global trade seem to be associated with control over the economic assets acquired. Since many higher economic growth. In other words, a reduction of these investments are placed in tradable securities in tariffs is associated with lower economic growth rates and syndicated facilities, they constitute relatively mobile (figure 4.3). capital that can change hands quickly if market senti- Hence, trade in and of itself provides little guaran- ments turn unfavorable. This may have repercussions tee for sustainable economic development, while eco- on foreign exchange rates when global market condi- nomic policy measures that favor a more global and tions change. By contrast, FDI gains managerial con- competitive business environment appear to induce eco- trol over business activities through direct corporate nomic growth. This suggests that it is the type of mer- acquisitions and investment in economic assets man- chandise a country exports that matters more than the aged by local affiliates. Since FDI constitutes investment actual trade volume. Developing countries as a whole in controlled economic assets, it is often considered a have increased their share of manufactured products from more stable source of trans-border financing. 25 percent of total exports to 70 percent over the past Whereas FDI commits financial resources in support two decades. However, the most successful developing of longer-term commercial activities, there are reserva- countries have had a higher emphasis on market-dynamic tions that this may not always be an advantage.22 Hence, product categories, including computer products, it is argued that developing countries should not actively electrical equipment, and manufactured garments, seek FDI at any price, but they should consider focus- ing policy efforts on improving economic conditions Figure 4.3 Economic growth and changes in tariff rates to attract capital to a country on the merits of under- lying business propositions and the expected returns 9 China from genuine economic activities. Developing countries 7 Argentina should attract investment that can build country-specific Malaysia 5 Uruguay skills and capabilities and create competitive advantages with the potential to drive more sustainable economic 3 1990­99 growth. FDI made primarily to exploit particular factor 1 endowments in a developing country does not repre- Growth Jordan -1 sent the most favorable type of investment. Instead, FDI GDP -3 in support of competency-based economic activities is Haiti much more attractive. Annual-5 The largest FDI amounts have accrued to economies -7 Rwanda focused on manufactured goods and with relatively low - 9-80 concentrations of particular merchandise exports. Brazil, -60 -40 -20 0 20 China, Mexico, and Thailand represent some of the Precentage change in tariffs, 1980s­1990s prime recipients of FDI.23 By contrast, developing coun- Source: World Bank data, Dollar and Kraay (2001). tries focused more narrowly on specific types of exports 62 Building Safer Cities: The Future of Disaster Risk have fared considerably worse. An analysis of the Figure 4.4 Economic losses and export concentration countries in the sample shows a significant negative 18 relationship between the degree of concentration on Nicaragua specific export products and economic growth. Over- 16 dependence on the export of specific product groups 14 Jamaica makes a country's foreign currency earnings vulnerable 12 to changes in global demand and the relative terms-of- trade. Prices for many primary commodities and labor- 1990­99 10 Zimbabwe intensive manufactures have decreased substantially 8 over the past decade, highlighting the risk of high export concentration.24 Losses/GDP 6 Bangladesh These revelations may guide developing countries Nepal 4 Costa Rica R. B. de in positioning themselves to take advantage of global Venezuela 2 market opportunities. Lower trade barriers and regu- China Mali latory restrictions make cross-border business transac- tions easier and provide greater flexibility in establishing 0 0.2 0.4 0.6 0.8 Export concentration index global corporate structures and networks. International Sources: Centre for Research on the Epidemiology of Disasters companies have taken advantage of the ability to inte- (CRED), United Nations Conference on Trade and grate national comparative advantages into their global Development (UNCTAD), and World Factbook. organization structures. Hence, various corporate func- tions may be located in countries that represent the highest potential value for the corporation and even- not only pursue stable and sound fiscal and monetary tually its customers. For example, labor-intensive oper- policies but also engage in risk management practices ations may be located where there are ample and qualified to stimulate economic activities that have the potential human resources, product development may be located to create more sustainable competitive advantages. around centers of research excellence, and global mar- Analysis of the countries sampled shows a positive keting may be coordinated from locations with high relationship between FDI and GDP growth.26 There concentrations of specialized sales agents. appears to be evidence of positive development effects Evidence from the sample countries supports the from FDI as a means to enhance national comparative contention that a relatively undiversified economic base advantage. However, that does not necessarily imply aggravates the adverse economic impact of natural dis- that FDI always supports production of competency- asters, since there is a positive relationship between based merchandise or enhances development of skills export concentration and the relative size of a coun- and capabilities. On the contrary, FDI is often made in try's catastrophe losses25 (figure 4.4). Countries with a pursuit of favorable factor cost conditions in host coun- high export concentration are typically more depend- tries. The challenge for developing countries is to min- ent upon factor endowments and developments in global imize dependence upon comparative advantage in one commodity prices than are countries that emphasize or a few fields and leverage it with the development of the export of competency-based merchandise. To reduce more competency-based manufacturing before the ini- economic vulnerability, these countries must find ways tial factor price advantages fade. Analysis of the coun- to diversify their economic bases. The ability to diver- tries sampled does not show that FDI has a positive sify the economic base and gear business activities in influence on the development of competency-based eco- support of more durable competitive advantage relies nomic activities, however. Rather, there is evidence of on an environment that is conducive to investment in a positive relationship between the level of FDI and export skills and capabilities-enhancing activities. Such an envi- concentration.27 This implies that FDI is positioned to ronment requires stable economic policies and man- take advantage of favorable factor costs in resource-rich aged exposure to catastrophe risk. Governments should developing countries. It could then be argued that FDI Globalization and Natural Disasters: An Integrative Risk Management Perspective 63 to a large extent supports commodity-based exports. From a strategic perspective, business entities are For example, exports from Venezuela, Mali, and Jamaica better off if they can establish advantages based on are highly focused on oil, cotton, and coffee. On the their organizations' inherent skills and capabilities.31 other hand, the largest FDI in absolute dollar terms Such advantages can provide value to customers through has been directed to countries like China, Brazil, Mexico, unique products, services, and delivery features and andThailand,whichhavemorediversifiedeconomicbases value to businesses through the development of eco- and a higher ratio of technology-intensive manufactures. nomic efficiencies in sourcing and internal processes. From a corporate perspective, the ability to place Both offer the company competitive advantages. To the functional entities at optimal locations around the globe extent that a competitive advantage is based on unique provides new opportunities for increasing efficiencies and firm-specific capabilities difficult for competitors and improving innovation by accessing specific skills, to imitate, a competitive advantage can become sus- capabilities, and knowledge and integrating them into tainable over time. Governments that establish economies organizational activities.28 From a country perspective, that support local companies and overseas investors and international corporations' investment dollars might be enable them to develop specialized skills and compe- attracted if special-factor endowments, skills, and capa- tencies provide countries with the ability to create sus- bilities can be used to reinforce economic activities.29 tainable competitive advantages. If local and multinational Companies headquartered in developing countries businesses are successful, a stronger economic base for may also invest overseas to exploit the same advan- more sustainable long-term economic development tages in the global markets, thereby attracting new skills will emerge. and resources to the economy. If local companies have Though government planning can provide support the ability to create their own global corporate struc- for increased economic activity, the development of tures, this may assist in attracting needed skilled work- essential skills and capabilities needed for companies ers to developing countries. Knowledge transfer and to succeed in the global economy often comes from capabilities-based commercial linkages arise not only innovations within a business. Policymakers can sup- from incoming FDI, but also from outgoing FDI, as local port increased commercial activity, however, by estab- companies learn through their global network rela- lishing a stable socioeconomic environment, improving tionships. However, many governments in emerging education, supporting research facilities, and main- markets are reluctant to ease restrictions on overseas taining a well-functioning public infrastructure.32 Fur- capital investment by domestic entities. This may not thermore, government investment programs can support be in a country's best interests, however, as this restric- the development of specific skills and capabilities that tion limits the ability of companies to expand and can shape future core competencies. To achieve this, learn from the global marketplace and puts them at a there is a need to reduce economic vulnerability to catas- competitive disadvantage. trophe risk and improve responsiveness to major exoge- Hence, governments should understand the compar- nous shocks to the economy. The next section takes a ative advantages that might drive an economy. It is less closer look at this issue. complicated to exploit existing factor endowments, includ- ing primary commodities such as metals, agricultural, raw materials, and labor. This approach is valid as long Risk Management and Global Market Access as the government also encourages the development of skills and capabilities that have the potential of creating The exponential growth in direct economic losses from longer-term comparative advantages to economic enti- natural catastrophes has an adverse impact on future ties operating in the economy. The difficulties with over- economic growth in an exposed developing country, reliance on cheap labor is that, once wage levels start to unless the impact is mitigated. Over-dependence on increase, simple comparative advantage erodes and com- international catastrophe funding and aid aggravates panies move manufacturing facilities to developing coun- the "moral hazard" (see endnote 16) problem reflected tries with even lower wage costs.30 in insufficient risk mitigation efforts and ineffective 64 Building Safer Cities: The Future of Disaster Risk post-disaster reconstruction. The uncertainty associ- factors can have at least three origins: catastrophe risk ated with uncontrollable catastrophe exposure and exposure, extreme price instability, and a deteriorating other exogenous economic shocks is detrimental to competitive position in global markets. Since these risk capabilities-based global linkages that could improve a factors are interrelated, risk management should inte- country's competitiveness. The development of more grate all factors into a country's aggregate risk exposure. knowledge-intensive competencies is hampered if eco- Countries with a high concentration of exports are nomic entities operating in developing countries are usually dependent upon specific commodities such as considered vulnerable counterparts. Hence, develop- food products, agricultural raw materials, and metals ing countries must become less vulnerable to natural (see figure 4.5). For example, countries such as Costa disasters and more responsive to changing economic Rica and Nicaragua are dependent upon world prices conditions. A proactive risk management approach for their coffee exports, so falling prices have a severe would help countries cope better with exogenous impact upon livelihoods.34 Similarly, Mali and to some environmental and economic shocks. extent Paraguay, depend upon the price of cotton, while Vulnerability to risk is a highly individualized phe- Zambia is highly dependent upon the price of copper. nomenon that depends upon the perceptions of a soci- When world prices drop, economic conditions in export- ety to risk. A society that fails to address the risks of ing countries are affected. Extreme dependency upon possible economic threats caused from such factors as primary commodities, therefore, provides little resilience price volatility, global competitive developments, and for an economy to withstand price declines and closely devastation from natural catastrophes assumes much links these commodity exports with poverty.35 higher risks than those that take steps toward preven- Whereas natural catastrophes have a direct economic tion and mitigation. It is in a country's self-interest to impact on exposed countries, business conditions can manage key risks in a proactive manner. International also be affected indirectly by climatic events in other businesses and financial institutions sensitive to high parts of the world that influence supply and demand levels of risk may avoid investing in such uncertain cir- conditions in commodities markets.36 When the prices cumstances. This avoidance further negatively impacts of primary agricultural products increase due to scarcity, a country's ability to create global linkages and attract this often provides a temporary economic bonanza for funds for new investment. producers. But when prices drop on the world market, A country that is able to maintain a relatively stable the decrease in export earnings results in an overall drop economic environment will attract significantly more in demand for all goods in export countries. Lower com- FDI than a country with a volatile economy. Economic modity prices should represent market opportunities entities operating in stable countries prove more reli- for countries to offer goods more cheaply; however, pri- able business partners for international firms and their mary producers are often unable to take advantage of better risk practices facilitate long-term, knowledge- such opportunities because they lack the skills to engage intensive partnerships that have a greater potential for in international product development and global sales receiving international funding for new ventures. Risk initiatives. Hence, successfully engaging national com- management facilitates long-term business activities and panies in global secondary market activities, including increases the potential of creating competitive advan- product development, packaging, sales, and distribu- tages and more sustainable growth. A number of con- tion, hedges the economy against the adverse effects of ditions must be satisfied in a stable socioeconomic deteriorating terms-of-trade. environment. There must be prudent fiscal and mone- Primary and secondary industries constitute a dimin- tary regimes as well as trade and foreign exchange ishing share of total factor income in the global econ- policies that assist foreign entities in establishing omy, though tertiary service-oriented economic activities skills-based linkages with local companies and sup- are increasing at a much faster rate than conventional porting domestic companies in their global expansion industries, particularly in industrialized countries. Thus, to overseas markets.33 There is also equal need to manage a developing country that focuses on the production of the major risk factors that expose a country. These risk primary commodities will likely see their terms-of-trade Globalization and Natural Disasters: An Integrative Risk Management Perspective 65 Figure 4.5 Commodity price developments, 1990­2000 Classes of Primary Goods Individual Primary Goods (food, agricultural, metals) (coffee, cotton, copper) 140 200 130 180 160 120 100) 140 100) 110 = = 120 100 (1990 (Jan-97 100 90 80 Value Index 80 60 Index Price 70 40 60 20 0 50 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Jan-97 Jul-97 Jul-98 Jul-99 Jul-00 Food Agricultural raw materials Metals Mar-97May-97 Sep-97Nov-97Jan-98Mar-98May-98 Sep-98Nov-98Jan-99Mar-99May-99 Sep-99Nov-99Jan-00Mar-00May-00 Sep-00Nov-00 Coffee Cotton Copper Sources: International Monetary Fund (IMF), Food and Agriculture Organization of the United Nations (FAO), and the London Metal Exchange (LME). deteriorate in the future. The only way to avoid this vicious reliable and consistent economic policies that main- commodity price trap is to encourage and support a tain fiscal and external trade balances within reason- focus on higher value-added business activities by going able boundaries. It also depends upon a country's beyond an emphasis on primary commodities and engag- integrative risk management capabilities that allow a ing in product development and new ways of creating country to cope with the economic effects of natural customer value to end users in global markets. Gov- disasters. Without the ability to manage and dampen ernments may support the development of specific skills the adverse impacts of external shocks, it is difficult to and capabilities and encourage local business entities to develop a sustainable base for economic value cre- become further engaged in competency-based economic ation. Firms and government entities insure themselves activities.37 This may induce FDI that links overseas against various kinds of risk that is beyond their con- distributors with domestic supplier affiliates and pro- trol and that otherwise could jeopardize firm survival vides local companies with opportunities to expand over- or severely damage public investment programs. If a seas and pursue linkages with sales affiliates in global firm or government assumes extreme risk exposure, markets. Effective linkages between local companies and insolvency risk increases and can reach levels where foreign affiliates require a mix of skills and technologi- credit becomes scarce and considerably more expen- cal know-how. This is even more critical when local com- sive.39 Potential restrictions on funding have adverse panies want to expand into overseas markets. There is impacts on economic activity levels that may cause an equal need to develop basic management skills and irreparable harm to important stakeholder relationships, international business capabilities to support the over- e.g., shareholders, employees, customers, suppliers, part- seas expansion of local companies.38 ners, etc., and strain profitability and future business The development of competency-based economic initiatives.40 This causes investment activities to drop as activities requires a relatively stable socioeconomic envi- viable funding sources dry up or become excessively ronment. This in turn depends upon the pursuit of costly. Hence, a highly disruptive business environment 66 Building Safer Cities: The Future of Disaster Risk without effective risk-transfer and hedging markets and factor income development may be sensitive to restrains economic growth. Indeed, the ability to iden- developments in primary commodity prices, foreign tify and manage risk in a proactive manner is heeded exchange rates, and global demand conditions. Income as a key characteristic behind the economic success of generation and economic growth are dependent upon modern society.41 the relative competitiveness of economic entities oper- A formal risk management process comprises a ating in a country and economic indicators in the global number of sequential tasks: risk identification, risk meas- economy. The potential direct losses deriving from urement, and risk monitoring. The contemporary risk various natural catastrophes can be determined on the management paradigm suggests that all relevant risk basis of advanced model simulations that are informed factors should be considered and integrated into the by data describing historical meteorological and seis- process and monitored on a continuous basis.42 Hence, mological event patterns and data describing the char- the risk management process in developing countries acteristics of the exposed economic infrastructure. It is should address exposure to a number of risk factors that also possible to develop econometric models that stip- affect economic performance, including market volatil- ulate the associated secondary effects on economic ity, natural catastrophes, and competitive risks. Differ- demand, investment activities, and government finances. ent risk exposures require different responses, but a The exposure position of different risk factors can diversified economy focused on competency-based man- be incorporated into a formal reporting system that ufacturing is generally more resilient to exogenous allows policymakers to monitor the manner in which economic shocks. A risk management process would a country's overall risk exposure is evolving. As the typically follow a series of sequential steps performed economic infrastructure, global market conditions, and in a continuous process (figure 4.6). As an initial step catastrophe frequencies change, risk exposure reports in the process, all risk factors that could affect an econ- reflect the consequences of the changed environmen- omy should be identified. Potential sources of risk must tal reality. As environmental conditions continue to be determined up front to devise alternative responses change, risk management frameworks should reflect an that could counteract the potential adverse effects of the ongoing and dynamic process. Registration and quan- risk exposure. Once key risk factors are identified, tification of important risk factors provide the basis for economic exposure associated with each factor should an informative mapping of the risk exposures that influ- be analyzed. ence a country's economic development path. This Exposure to each of the risk factors identified can be overview of a country's risk landscape allows decision- quantified and measured to assess relative importance. makers to evaluate the effects of alternative actions to For example, a country's balance-of-payments flows modify or limit the overall risk profile. The analytical framework also provides a basis for reducing risk exposure by making risk-transfer arrangements. Figure 4.6 Elements of the dynamic risk management Depending upon the nature of the risk, residual risk process exposure can be hedged through financial derivatives,43 3. reinsurance, and alternative risk transfer (ART) instru- ments. Economic exposure related to changing com- Risk monitoring petitive conditions in the global marketplace cannot be insured in the financial markets. Competitive advan- Risk Risk 4. 2. reassessment measurement tage typically relates to firm-specific, nontradable, intan- gible factors, so no market-based instruments exist to Risk hedge these exposures. It may be possible, however, to identification adopt a real options perspective to managing these long- 1. term exposures.44 The real options concept is the van- Source: Culp 2002. guard of strategic risk management, and it provides Globalization and Natural Disasters: An Integrative Risk Management Perspective 67 interesting new ways to respond to idiosyncratic non- rate in developing countries. The current approach to marketable economic exposures.45 New business oppor- multilateral catastrophe funding causes moral hazard tunities planned by economic entities, but not yet problems that leave too few incentives to engage in implemented, can be conceived as an options portfo- proactive risk management that could promote more lio that gives a country economic flexibility and enhances effective risk mitigation and post-disaster reconstruc- its development path.46 tion. There is an urgent need to support developing Options are not always obvious, however, and some- countries in managing the results of the current trend times must be innovated. The creation of options depends and assist them in pioneering new integrative risk upon the existence of economic entities in a country management practices. that are innovative and able to take on new business Open international trade relations can benefit all trad- initiatives. Governments can support the establishment ing partners, but a high export concentration of pri- of an economic environment conducive to serious options mary commodities among the world's poorest nations creation, such as managing excessive risk exposure, has entrapped them with terms-of-trade that continue maintaining public security and health standards, build- to deteriorate. Other industrialized developing coun- ing economic infrastructure, and investing in education tries are trapped by over-reliance on favorable labor and intellectual capital.47 The ability to create options costs, which constitute an unsustainable advantage. The and manage the associated flexibility can add signifi- poorest developing countries are also hit hardest by cant value to an economy. The more options available the economic devastation of natural disasters that in an economic portfolio, the more responsive and often aggravate an already-strained economic situation. resilient it can be to external shocks, whether from High dependence upon particular commodities pro- natural disasters or changing conditions in the global vides little room for responsiveness to adverse economic market. shocks. The obvious response is to create a more diverse A government that manages all risk factors on an industry structure and advance capabilities-based eco- integrative basis can cover excessive economic risk expo- nomic activities that have a better potential for creat- sure by combining a number of risk-transfer techniques ing competitive advantages. This approach requires a to ensure that sufficient funds will be available to stable economic environment founded in sociopoliti- retain economic responsiveness and quickly refurbish cal stability and active management of exposure to nat- essential infrastructure in the event of a disaster. Gov- ural catastrophes and other exogenous economic shocks. ernments should take steps to identify and continuously Prudent economic policies and proactive risk manage- survey risks that could impact the economy. They should ment practices can help developing countries establish also determine a level of prudent risk exposure and a business environment that is more conducive to a sus- manage the country's risk profile, within limits, through tainable development path. a combination of self-insurance, risk-transfer opportu- The paper has sought to derive conclusions from nities, a diversified industrial structure, and an eco- empirical evidence, but the underlying sample of devel- nomic options portfolio that builds flexibility into an oping countries has its limitations. As the conclusions economy. appear fairly generic, robust further analyses of more comprehensive data sets may be warranted. These stud- ies could consider some of the constructs introduced in this paper and define classes of competency-based Conclusions business activities, types of competitive advantage, and economic option portfolios. Risk management Globalization has significantly spurred trade and invest- approaches could also be tested in country-specific pilot ment flows over the past decades. At the same time, studies that specify the direct and indirect economic the frequency of natural catastrophes has increased benefits associated with an integrative risk management and associated economic losses have risen at an alarming process. 68 Building Safer Cities: The Future of Disaster Risk Annex 4.1 Reported catastrophe victims and economic losses, 1990­2000 People Total Affected/ Catastrophe GDP2 Losses/ affected population population1 losses PPP-basis GDP Country (#) (mill.) (%) ($ mill.) (# bill.) (%) Argentina 12,979,161 37.38 34.72 8,596 476.0 1.81 Bangladesh 329,090,346 131.27 250.70 10,579 203.0 5.21 Benin 844,000 6.59 12.81 3 6.6 0.05 Brazil 40,863,947 174.47 23.42 5,051 1130.0 0.45 Burkina Faso 2,823,293 12.27 23.01 0 12.0 0.00 Cameroon 810,070 15.80 5.13 2 26.0 0.01 Central Afr. Rep. 79,680 3.58 2.23 0 6.1 0.00 China 1,387,422,101 1,273.11 108.98 112,314 4500.0 2.50 Colombia 2,565,541 40.35 6.36 4,875 250.0 1.95 Costa Rica 1,179,442 3.77 31.26 848 25.0 3.39 Dominica 3,716 0.07 5.23 3 0.3 1.13 Dominican Rep. 2,368,827 8.58 27.61 2,264 48.3 4.69 Ecuador 577,276 13.18 4.38 801 37.2 2.15 Egypt, Arab. Rep. 199,331 69.54 0.29 292 247.0 0.12 Ethiopia 45,315,900 65.89 68.78 19 39.2 0.05 Haiti 4,089,855 6.96 58.73 211 12.7 1.66 Hungary 133,695 10.11 1.32 677 113.9 0.59 India 1,002,191,581 1,029.99 97.30 20,213 2200.0 0.92 Indonesia 7,198,000 228.40 3.15 17,879 654.0 2.73 Côte d'Ivoire 51 16.39 0.00 0 26.2 0.00 Jamaica 1,463,121 2.67 54.90 1,221 9.7 12.58 Jordan 198,744 5.15 3.86 400 17.3 2.31 Kenya 17,441,900 30.76 56.70 12 45.6 0.03 Malaysia 149,869 22.23 0.67 12 223.7 0.01 Mali 1,853,902 11.01 16.84 0 9.1 0.00 Mexico 3,515,773 101.88 3.45 10,203 915.0 1.12 Nepal 5,676,894 25.28 22.45 1,298 33.7 3.85 Nicaragua 2,439,230 4.92 49.60 2,044 13.1 15.60 Pakistan 26,206,278 144.62 18.12 1,341 282.0 0.48 Paraguay 760,652 5.73 13.27 84 26.2 0.32 Peru 5,325,635 27.48 19.38 1,549 123.0 1.26 Philippines 93,468,162 82.84 112.83 5,450 310.0 1.76 Rwanda 1,481,976 7.31 20.27 0 6.4 Thailand 25,849,910 61.80 41.83 4,218 413.0 1.02 Uganda 1,351,500 23.99 5.63 72 26.2 0.27 Uruguay 37,500 3.36 1.12 30 31.0 0.10 Venezuela, R. B. de 702,581 23.90 2.94 2,088 146.2 1.43 Zambia 4,306,218 9.77 44.08 21 8.5 0.24 Zimbabwe 10,861,153 11.37 95.57 2,598 28.2 9.21 1A ratio above 100 percent indicates that a number of people exceeding the entire population have been affected during the period. 2Based on purchasing power equivalents as opposed to factor cost converted at current foreign exchange rates. Sources: Centre for Research on the Epidemiology of Disasters (CRED) and World Factbook. Globalization and Natural Disasters: An Integrative Risk Management Perspective 69 . (%) 0 Change 1985­95 60.00- 71.95- 70.33- 74.89- 53.13- 43.44- 41.88- 46.13- 58.50- 42.56- 52.98- 65.89- 29.22- 44.93- 13.79- 17.78- 61.47- 52.69- 21.29- 40.76- 1.84- 65.74- 40.27- 23.35- 49.54- 51.58- 39.74- 14.68- 70.44- 38.13- 16.36 43.66- 48.00- 71.51- 59.16- 43.14- 133.70 fs %). arifT 1990­99 (agv 12.45 40.15 26.85 16.25 35.75 18.35 21.80 30.40 14.40 11.90 21.50 17.00 11.15 31.70 22.55 10.50 12.35 50.10 16.65 22.25 15.25 15.90 23.40 11.60 21.40 12.80 13.55 11.70 50.75 11.20 16.15 20.85 38.40 29.85 15.05 14.25 14.25 21.70 19.35 fs %) arifT 8.90 9.30 9.60 1995­99 (avg. 11.00 26.00 12.70 11.50 28.50 18.10 18.60 20.90 12.20 11.20 15.00 16.20 11.70 28.10 16.30 10.00 14.80 38.30 13.20 20.70 10.90 16.00 13.50 18.80 12.80 11.00 10.70 41.70 13.30 17.20 38.40 23.10 13.00 12.70 17.00 21.50 sf %) arifT 9.90 1990­94 (avg. 13.90 54.30 41.00 21.00 43.00 18.60 25.00 39.90 16.60 12.60 28.00 7.801 10.60 35.30 28.80 11.00 61.90 20.10 23.80 19.60 15.80 33.30 14.30 24.00 12.80 16.10 12.70 59.80 13.10 19.00 24.50 38.40 36.60 17.10 18.90 15.80 26.40 17.20 sf %) . arifT 9.20 1985­89 (avg. 7.502 2.709 2.804 5.804 0.806 2.003 2.003 8.803 9.402 19.50 31.90 4.303 9.703 9.602 1.601 8.001 9.409 7.902 6.302 18.40 16.30 9.403 14.90 6.701 1.802 2.102 9.206 0.901 5.004 7.802 3.003 1.004 5.002 3.703 1.103 9.902 p.a.) rade/GDPT 1990­99 (% 8.302 2.652 8.804 5.701 0.504 6.756 4.504 2.153 1.955 4.307 5.005 0.656 4.152 2.958 5.806 9.551 9.404 2.207 6.055 1.455 1.704 1.155 6.807 4.703 8.358 8.504 90.80 1.954 9.608 7.253 5.357 0.955 8.607 8.256 118.20 115.40 117.55 164.20 196.85 p.a.) 32.90 26.70 45.80 17.90 37.90 65.00 42.10 34.20 58.90 92.30 57.50 59.70 25.40 98.90 74.00 22.10 50.20 76.40 60.70 51.30 49.90 60.30 85.10 34.50 99.40 52.70 37.40 94.60 43.10 84.30 54.70 78.30 77.10 2001. rade/GDPT 1995­99 (% 128.10 112.30 125.90 166.20 219.80 106.10 Bank, orld W fs p.a.) 23.70 18.60 51.80 13.50 43.10 68.50 46.90 30.10 45.00 56.30 52.50 61.60 22.90 67.00 57.60 17.00 48.60 68.00 51.40 51.60 33.50 42.00 68.50 34.90 77.30 44.30 75.50 46.50 84.60 31.40 66.40 47.20 78.90 59.40 oup, tarif rade/GDPT 1990­94 (% 108.30 118.50 109.20 162.20 173.90 Gr ch and Resear owth trade, Gr p.a.) 6.00 3.55 2.00 1.25 1.15 1990­99 (% 2.55- 1.50- 8.20 1.50 0.95 1.60 3.70 0.85 1.70 1.90 3.80- 0.25 3.50 4.40 0.05- 1.75- 0.10- 0.00 5.60 0.25 3.30 2.60 1.10- 0.25 0.40 2.85 1.25 7.30- 3.75 2.90 4.60 1.25 0.50 1.75 GDP Development, national owth inter Poverty Gr p.a.) 5.20 3.70 2.60 1.60 3.20 2.10 1995­99 (% 0.20- 7.80 0.60 0.10- 1.80 5.60 0.60 3.30 5.00 0.30- 3.30 4.40 4.50 3.30 2.70- 1.60- 0.60 5.40 2.30 4.20 2.20 0 0.30- 0.20- 3.00 3.10 0.30 1.50 4.20 4.30 0.20 2.40 3.10 and GDP owth, owth, gr Gr owth radeT, income Gr p.a.) 6.80 3.40 1.40 0.90 8.60 2.40 2.00 1.40 1.80 1.10 0.10 2.60 4.30 1.40 5.80 2.40 3.00 0.80 1.00 2.70 6.00 1.60 4.90 2.30 0.40 1990­94 (% 0.90- 7.20- 2.80- 1.20- 7.30- 2.80- 3.40- 0.80- 0.60- 1.80- 2.20- 0.60- 14.90- 1.40- Kraay GDP A. de and National Rep.. Rep. B. Rep. e R. 4.2 Faso Dollar Afr oon Rica Arab. y D. d'Ivoir gentina dan ce: Annex Country Ar Bangladesh Benin Brazil Burkina Camer Central China Colombia Costa Dominica Dominican Ecuador Egypt, Ethiopia Haiti Hungar India Indonesia Côte Jamaica Jor Kenya Malaysia Mali Mexico Nepal Nicaragua Pakistan Paraguay Peru Philippines Rwanda Thailand Uganda Uruguay enezuela,V Zambia Zimbabwe Sour 70 Building Safer Cities: The Future of Disaster Risk Annex 4.3 National income, export concentration, and foreign direct investment Foreign Direct Investment GDP1 Ln(GDP) Export2 PPP-basis Logarithmic Concentration 1996 1999 Avg. 1996­99 Avg. 1996­99 Country ($ bill.) transformation index ($ mill.) ($ mill.) ($ mill.) (pct. of GDP) Argentina 476.0 6.17 0.13 6,900.0 24,000.0 15,450.0 3.25 Bangladesh 203.0 5.31 0.24 13.5 179.7 96.6 0.05 Benin 6.6 1.89 . 36.0 40.7 38.4 0.58 Brazil 1130.0 7.03 0.09 11,200.0 28,600.0 19,900.0 1.76 Burkina Faso 12.0 2.48 . 17.0 13.0 15.0 0.13 Cameroon 26.0 3.26 0.40 35.0 40.0 37.5 0.14 Central Afr. Rep. 6.1 1.81 . 5.0 13.0 9.0 0.15 China 4500.0 8.41 0.07 40,200.0 38,800.0 39,500.0 0.88 Colombia 250.0 5.52 0.29 3,100.0 1,500.0 2,300.0 0.92 Costa Rica 25.0 3.22 0.39 427.0 669.0 548.0 2.19 Dominica 0.3 . 0.50 17.8 18.0 17.9 5.97 Dominican Rep. 48.3 3.88 . 96.5 1,300.0 698.3 1.45 Ecuador 37.2 3.62 0.40 491.0 690.0 590.5 1.59 Egypt, Arab. Rep. 247.0 5.51 0.27 636.0 1,100.0 868.0 0.35 Ethiopia 39.2 3.67 . 22.0 90.0 56.0 0.14 Haiti 12.7 2.54 0.29 4.1 30.0 17.1 0.13 Hungary 113.9 4.74 0.12 2,300.0 2,000.0 2,150.0 1.89 India 2200.0 7.70 0.16 2,400.0 2,200.0 2,300.0 0.10 Indonesia 654.0 6.48 0.15 6,200.0 2,745.0 4,472.5 0.68 Côte d'Ivoire 26.2 3.27 . 269.0 323.7 296.4 1.13 Jamaica 9.7 2.27 0.64 183.7 523.7 353.7 3.65 Jordan 17.3 2.85 0.16 15.5 158.0 86.8 0.50 Kenya 45.6 3.82 0.28 12.7 13.8 13.3 0.03 Malaysia 223.7 5.41 0.22 5,100.0 1,000.0 3,050.0 1.36 Mali 9.1 2.21 0.67 84.0 19.0 51.5 0.57 Mexico 915.0 6.82 0.12 9,200.0 11,900.0 10,550.0 1.15 Nepal 33.7 3.52 0.45 19.2 4.4 11.8 0.04 Nicaragua 13.1 2.57 0.30 97.0 300.0 198.5 1.52 Pakistan 282.0 5.64 0.23 922.0 532.0 727.0 0.26 Paraguay 26.2 3.27 0.41 149.0 87.3 118.2 0.45 Peru 123.0 4.81 0.24 3,200.0 2,400.0 2,800.0 2.28 Philippines 310.0 5.74 0.50 1,500.0 573.0 1,036.5 0.33 Rwanda 6.4 1.86 . 2.2 1.7 2.0 0.03 Thailand 413.0 6.02 0.11 2,300.0 6,200.0 4,250.0 1.03 Uganda 26.2 3.27 0.56 11.20 222.0 171.5 0.65 Uruguay 31.0 3.43 0.18 136.8 235.0 185.9 0.60 Venezuela, R. B. de 146.2 4.98 0.79 2,200.0 3,300.0 2,750.0 1.88 Zambia 8.5 2.14 . 117.0 163.0 140.0 1.65 Zimbabwe 28.2 3.34 0.31 81.0 59.0 70.0 0.25 1 Based on purchasing power equivalents as opposed to factor cost converted at current foreign exchange rates. 2 The index is calculated as the square root of the sum of the ratios of each of the 239 three-digit SITC product groups over total export raised to the power of two. The results are normalized by dividing by the square root of 1 over 239 to create a numerical index range from 0 to 1 (this is a modified Hirschmann index). Sources: United Nations Conference on Trade and Development (UNCTAD), World Factbook, and World Bank data. Globalization and Natural Disasters: An Integrative Risk Management Perspective 71 Annex 4.4 Correlation Analysis (Pearson correlation coefficients) 1. 2. 3. 4. 5. 6. 7. GDP Growth Ln(GDP) Pct. Losses Trade/GDP Tariff ratio Tariff change Conc. index 1. GDP Growth . . . . . . . 2. Ln(GDP) 0.655** . . . . . . 3. Pct. losses -0.070 -0.141 . . . . . 4. Trade/GDP -0.073 -0.251 0.162 . . . . 5.Tariff rate -0.021 0.218 -0.200 -0.405* . . . 6. Tariff change -0.287+ -0.230 0.237 0.201 -0.034 . . 7. Conc. Index -0.453** -0.675** 0.210 0.189 -0.184 0.063 . 8. FDI/GDP 0.133 0.099 0.223 0.308+ -0.337* -0.194 0.149 Statistical significance levels: **p > 0.01; *p > 0.05; p > 0.10 + Notes 6. Benson and Clay (2001) observe that major disasters influ- 1. This may well underestimate actual damage, since available ence the composition of public spending and funding patterns, information on losses associated with natural catastrophes is less distort short- and medium-term investment plans, and hence than perfect and somewhat incompatible. The key sources for adversely affect economic growth potential, particularly in the loss data, e.g., Swiss Re, Munich Re, and CRED, often use economies that are dependent on public investment. different cutoff points in their definition of a catastrophe and they 7. The countries studied in this paper constitute a sample of all rely on different external informants such as newspaper arti- 39 developing countries that maintain relatively high global cles, news agencies, various multilateral organizations, insurance trade activities. The sample is taken from Dollar and Kraay reports, reinsurance periodicals, and specialist publications. (2001). 2. Total reported direct economic losses from natural catastro- 8. There is a negative correlation between economic growth phes reached close to $190 billion in 1996 and $28 billion in and catastrophe losses as a percentage of GDP in the sample, 2000. Source: Centre for Research on the Epidemiology of Dis- but the correlation coefficient is not statistically significant. asters (CRED), International Disaster Database, Université 9. The correlation coefficient between the percentage of the pop- Catolique de Louvain, Belgium. Secondary economic effects go ulation affected and annual economic growth is positive in this unreported although they can be substantial. Reported losses sample, although not statistically significant. refer to direct damage inflicted on private homes, commercial 10. The World Bank has extended more than $7 billion in post- assets, and public infrastructure. Natural catastrophes cause disaster loans and credits over the past 20 years (Gilbert and additional indirect damage due to reduced economic activities, Kreimer 1999). lost market opportunities, distortion of commercial working 11. In the Red Cross World Disasters Report (2001), there is a vivid relationships, disruption of educational efforts, research and description of how international relief organizations have per- development initiatives, strained public finances, contraction formed in particular disaster situations that prompted serious of capital investments, etc. questions about "whose needs are best served by aid--those of 3. True insurance coverage in developing countries is consider- the donor agencies or their beneficiaries." ably lower, because natural catastrophe statistics often do not 12. This is referred to as a "moral hazard" issue because author- include loss estimates (less than one in three registered natural ities are often less proactive in managing risk exposures when catastrophes in developing countries reported any loss figures). there is an expectation that international organizations will extend 4. The loss factors roughly correspond to annual percentage emergency assistance. See, for example, Financial Markets Trends, increase in catastrophe losses of around 25 percent in developed No. 76, 2000, OECD. economies and 30 percent in developing economies (calculated 13. Global competitive advantage can be achieved when an on a compound rate basis). economic entity is able to provide superior value to customers 5. The Red Cross World Disasters Report (2001) refers to these more efficiently than international competitors. The sustain- losses as emanating from "un-natural" catastrophes as they ability of competitive advantage depends on the specificity with escalate due to a lack of focused risk mitigation. which an entity is able to create superior value based on unique 72 Building Safer Cities: The Future of Disaster Risk product/service features or firm-specific processes that are dif- 21. Trade and Development Report, 2002, United Nation Con- ficult to emulate. This makes an advantage more sustainable. ference on Trade and Development (UNCTAD). 14. Various negotiations of the General Agreement on Tariffs and 22. See, for example, Mishra, Mody, and Murshid (2001) and Trade (GATT) have reduced import tariffs on manufactured goods Loungani and Razin (2001). over the years. Trade agreements on goods, services, and intel- 23. See also Financial Market Trends, Recent Trends in Foreign lectual property are now administered by the World Trade Orga- Direct Investment, No. 76, June 2000, OECD, Paris. nization (WTO), which also settles trade disputes. The WTO is 24. The world market prices for several commodities, such as organizing a series of ongoing negotiations to extend trade cotton, soybeans, and wheat, have arguably been depressed by rules on such things as agricultural products, textiles, and public the agricultural subsidies provided within the European Union, procurement. the United States, and Japan. 15. WTO and United Nations Conference on Trade and Devel- 25. There is a positive correlation between the export concentra- opment (UNCTAD). tion index and the ratio of registered direct catastrophe losses over 16. There has been a significant increase in FDI in developed GDP, although it is not statistically significant in the current sample. economies over the past decade stemming primarily from the 26. There is a positive correlation between the level of foreign ongoing integration of economic activities within the European direct investment, both in relative and absolute terms, and the Union. annual GDP growth, although it is not statistically significant in 17. Comparative advantage can arise from relative differences in this sample. production efficiencies. In an open economy, resources would 27. There is a positive and statistically significant correlation gradually be channeled to companies that maintain relatively high between the level of foreign direct investment over GDP and efficiencies when producing particular products and services and the export concentration index. they would become a source of exports to global markets, whereas 28. See Prahalad and Doz (1987), Bartlett and Ghoshal (1989), other product and service areas where companies have relatively Reich (1991), Markusen (1995). lower efficiencies would face stiff competition from global imports. 29. This type of foreign direct investment activity is not con- Relative production efficiencies can stem from the endowment fined to large multinational corporations. It is also a part of many of production factors in the economy such as oil, minerals, land, small to medium-sized companies with the flexibility to take agricultural raw materials, and labor, but they can also be rooted advantage of a global network. in superior skills, capabilities, and knowledge - specialized 30. Washington Post, "Mexican workers pay for success: with pharmaceuticals, telecommunications and engineering know- labor costs rising, factories depart for Asia," June 25, 2002. how, manufacturing and management capabilities. Therefore, in 31. See e.g., Barney (2002) and Saloner, Shepard, and Podolny a dynamic economic environment, comparative advantage can (2001). also be created from innovations in product and service offer- 32. The new partnership for Africa's development (NEPAD) pro- ings, technological inventions, and continuous improvements moted by South Africa's President Thabo Mbeki is a notewor- in operational processes that create better value for end-users or thy attempt to further such an approach. provide value more efficiently. The latter competency-based sources 33. The major causes of vulnerability to contagion are external of comparative advantage are proving to be more profitable, imbalances, unrealistic foreign exchange rate regimes, lax fiscal resilient, and durable in the contemporary economic environ- policies, non-credible monetary policies, unhealthy financial ment compared to pure factor-based advantages. sector, the quality of financial reporting, etc., OECD, Financial 18. The sample is based on Dollar and Kraay (2001). This Market Trends, June 2000. paper analyzes variables on a cross-section of developing coun- 34. Financial Times, "Coffee republics see their `grain of gold' tries over certain time periods whereas Dollar and Kraay ana- lose its luster," June 26, 2002. lyze internal country variations over time. 35. The Least Developed Countries Report 2002: Escaping the 19. There is a negative correlation between trade/GDP and annual Poverty Trap, United Nations Conference on Trade and Devel- GDP growth, but it is not statistically significant in this sample. opment (UNCTAD). 20. Collier and Gunnin (1999) find that African countries that 36. Wall Street Journal, "Weather condition El Niño may be return- have pursued open trade policies have generally achieved higher ing this year: companies that depend on domestic demand in economic efficiencies. Asia could be hurt," March 27, 2002. Globalization and Natural Disasters: An Integrative Risk Management Perspective 73 37. World Investment Report, 2001, Promoting Linkages, United 46. Hence, these real options provide firms with the right, but Nations Conference on Trade and Development (UNCTAD). not the obligation, to pursue new business opportunities. 38. See e.g., Wood (2000). 47. See Boer (2002). 39. If funding becomes scarce, new and possibly profitable proj- ects will not be funded. As a consequence, a firm will follow a Bibliography less than optimal growth path. In the country context, a credit Albala-Bertrand, J. M. 1993. Political Economy of Large Natural crunch will prevent a government from investing in important Disasters. Oxford: Clarendon Press. long-term economic development programs that could other- Andersen, T. J. 1993. Currency and Interest Rate Hedging, wise improve the competitive position of economic entities oper- Second Edition. New Jersey: Prentice-Hall. ating in the country. See Froot, Scharfstein, and Stein (1994). Andersen, T. J. 2000. "Real Options Analysis in Strategic Deci- 40. If the economic performance of a firm (or a country) becomes sion Making: An Applied Approach in a Dual Options excessively volatile, for example, due to uncontrolled risk Framework." Journal of Applied Management Studies 9(2):235­55. exposure, the credit risk of a firm (or country) increases and it Andersen, T. J., and P. Masci. 2001. "Economic Exposures to Nat- will be considered a more risky counterpart. It is more risky not ural Disasters, Public Policy, and Alternative Risk Management only for purposes of credit extension, but also in general eco- Approaches." Infrastructure and Financial Markets Review 7(4). nomic interactions because the entity's ability to fulfill its future Barney, J. B. 2002. Gaining and Sustaining Competitive Advan- commitments to creditors, customers, and suppliers, is jeop- tage. Second Edition. New Jersey: Prentice Hall. ardized. See Miller (1998). Bartlett, C. A., and S. Ghoshal. 1989. Managing Across Borders. 41. See Bernstein (1996). Cambridge, MA.: Harvard Business School Press. 42. See Barton, Shenkir and Walker (2002) and Dogherty (2000). Barton, T. L., W. G. Shenkir, and P. L. Walker. 2002. Making 43. Financial futures, forwards, and options make it possible to Enterprise Risk Management Pay Off. New Jersey: Prentice-Hall. lock in future market rates typically for periods of 6­18 months Benson, C., and E. Clay. 2001. "Bangladesh: Disasters and Public (Andersen 1993). This hedging technique cannot bypass the Finance." London: Overseas Development Institute. consequences of lower-than-expected market rates or continu- Bernstein, P. L. 1996. Against the Gods: The Remarkable Story of ously deteriorating terms-of-trade, but it can smooth the Risk. New York: Wiley. volatility of earnings flows. Boer, F. P. 2002. The Real Options Solution: Finding Total Value in 44. A financial option is a right, but not an obligation, to buy a High-Risk World. New York: Wiley. (call) or sell (put) a particular traded asset at a predetermined Charveriat, C. 2000. "Natural Disaster Risk in Latin America and price at a future time. Hence, options gives the holder the flex- the Caribbean." Washington, D.C.: IDB. ibility to utilize a favorable market situation, or let the option Collier, P., and J. W. Gunnin. 1999. "Explaining African Eco- lapse rather than incur a loss. This flexibility has value, which nomic Performance." Journal of Economic Literature 37(1). can theoretically be estimated based on the characteristics of Culp, G. L. 2002. The ART of Risk Management: Alternative Risk market price development of the underlying asset. A real option Transfer, Capital Structure, and the Convergence of Insurance represents that same formal structure except the underlying and Capital Markets. New York: Wiley. assets are not traded. They typically constitute an investment Doherty, N. A. 2000. Integrated Risk Management: Techniques opportunity underpinned by firm-specific and unique capabil- and Strategies for Reducing Risk. New York: McGraw-Hill. ities. Hence, a new economic venture or business opportunity Dollar, D., and A. Kraay. 2001. Trade, Growth, and Poverty. Wash- represents a real option because the firm can utilize the real ington, D.C.: World Bank. option when market conditions are favorable and leave or post- Dollar, D., and A. Kraay. 2002. "Spreading the Wealth." Foreign pone it if conditions are not yet favorable. The flexibility of this Affairs 81(1). options structure has value like that of a financial option. An Froot, K. A., D. S. Scharfstein, and J. C. Stein. 1994. "A Framework economy where economic entities have the ability to develop for Risk Management." Harvard Business Review (November- many different types of real options will have more alternatives December). for expanding economic activity and become more responsive Gilbert, R. and A. Kreimer. 1999. Learning from the World Bank's to changes in market conditions. Experience of Natural Disaster Related Assistance. Washington, 45. See Andersen (2000). D.C.: World Bank. 74 Building Safer Cities: The Future of Disaster Risk ICRC (International Red Cross and Red Crescent Societies). 2001. OECD (Organisation for Economic Co-operation and Develop- World Disasters Report: Focus on Recovery. Geneva. ment). 2000. Financial Market Trends No. 76. Paris. Kleindorfer, P. R., and H.C. Kunreuther. 1999. "Challenges Facing Oxfam International. 2001. "Eight Broken Promises: Why the the Insurance Industry in Managing Catastrophe Risks." In WTO Isn't Working for the World's Poor." Oxfam Briefing K.A. Froot, ed., The Financing of Catastrophe Risk. Chicago: Paper. London. University of Chicago Press. Prahalad, C. K., and Y. L. Doz. 1987. The Multinational Mission. Lall, S. 2000. "The Technological Structure and Performance of New York: Free Press. Developing Country Manufactured Exports, 1985­98." Oxford Reich, R. B. 1991. The Work of Nations. New York: Vintage. Development Studies, 28(1). Saloner, G., A. Shepard, and J. Podolny. 2001. Strategic Man- Loungani, P., and A. Razin. 2001. "How Beneficial Is Foreign agement. New York: Wiley. Direct Investment for Developing Countries?" Finance & Devel- UNCTAD (United Nations Conference on Trade and Develop- opment 38(2). ment). 2001. World Investment Report, 2001. Geneva. Markusen, J. R. 1995. "Boundaries of Multinational Enterprises UNCTAD. 2002. Trade and Development Report, 2002. Geneva. and the Theory of International Trade." Journal of Economic [CK text] Perspectives 9(x). UNCTAD. 2002. The Least Developed Countries Report 2002: Escap- Miller, K. D. 1998. "Economic Exposure and Integrated Risk ing the Poverty Trap. Geneva. [CK text] Management." Strategic Management Journal 19(5). Wood, E. 2000. "Aligning Innovation for Dynamic Capabilities Mishra, D., A. Mody, and A. P. Murshid. 2001. "Private Capital and Sustainable Growth in South African Manufacturing." Flows and Growth." Finance & Development 38(2). Working Paper. Cape Town: University of Cape Town. Chapter 5 Urban Disasters and Globalization J. M. Albala-Bertrand Disasters and their short- and long-term impacts upon impacts, since smaller ones are less destructive and more economies and societies have been studied extensively. easily absorbed, though much of the discussion is also As globalization spreads, this movement affects global applicable for smaller disasters. First, a disaster impact and local economies, implying changes in disaster is generally defined as major if estimated direct losses impacts. This paper looks at the impacts of disasters in approach or exceed the average GDP growth rate of an urban areas in light of globalization. While not all affected country and/or the damage seriously affects agree that globalization benefits developing countries, economic activity, even if direct losses from the event this movement allows certain assumptions to be made are not a significant portion of GDP. based on the performance of industrialized economies Second, the overwhelming majority of disasters, and linkages that impact developing countries. Fur- whether large or small, are geographically and/or eco- thermore, it appears that the effect of disaster situa- nomically localized, and a disaster impact is unlikely to tions on the macroeconomy is negligible; therefore, have negative consequences for the economy as a whole. disaster management activities should focus on com- Third, unless indicated, this paper takes into consider- munities and their resilience. ation only sudden types of urban natural (earthquakes, floods, hurricanes, etc.) and technological (engineering, chemical, biological, etc.) disasters, with the latter traced Urban Area and Urban Hazards mainly to institutional failure. Many sudden types of tech- nological disasters are confined to small areas, but they Urban Area. A city or urban area is a set of infrastruc- often have larger consequences for the surrounding pop- ture, other structures, and buildings that create an envi- ulation and areas not directly affected. The direct effects ronment to serve a population living within a relatively of these disasters can generally be treated similarly to small and confined geographic area. The lives and liveli- those resulting from natural disasters, but the responses hoods of the population are supported by interrelated tend to be different, since the institutional and internal systems around which the urban area and society political processes also affect the response. function. Capital stock makes up much of the infra- Furthermore,ithasbeenshownthaturbannaturaldis- structure that includes business fixed-capital machinery asters,andbyextensiontechnologicalones,normallyhave and equipment, structural capital (plant, infrastructure, low direct impacts on secondary (industrial) and tertiary overheads, offices, and social-physical capital), and res- (services) sectors. Residential capital tends to bear the idential capital (dwellings). A disaster can directly affect brunt of damage and destruction. Social capital, associ- people and capital stock and, as a consequence, the sys- ated with education and health services, does not nor- temic functioning of both. mally suffer high negative impacts. Damage to service Urban Hazard. An urban hazard is a risk that threat- sources and networks is not normally high; therefore, it ens a city, its population, and related socioeconomic is short-lived. Social overhead, including infrastructure, activities. If a risk threatens a capital or large city, the risk is frequently damaged, though transport infrastructure is may resonate beyond the area of impact. In the discus- the most affected, followed by water supply and sewage sion that follows, the focus is mostly on major disaster systems.Withincommunities,thosemostfrequentlykilled 75 76 Building Safer Cities: The Future of Disaster Risk andinjuredaretheurbanpoor,whooftenlosetheirhomes, Opponents, on the other hand, do not doubt that many of which are constructed with low-quality materi- more integration might be economically beneficial, als on unstable and disaster-prone land. When accounted but they have doubts about globalization's pacing, for as a percentage of GDP, such losses can be minimal, sequencing, and policies. Their concerns arise from wor- thoughthelossesaredevastatingformanyfamilies.Rebuild- ries that quick, unregulated, and socially unaccommo- ingtheseareasmaynotbeexcessivelycostlyonapercapita dating transitions to advanced stages of globalization basis, but commitment to helping this vulnerable section can have, and have had, deleterious consequences for of the population requires political will. This same pop- the economies of developing countries and the general ulationmayalsosufferwhendailywagejobsarelostduring population, particularly the most vulnerable people. a disaster. Ironically, particularly in urban areas, job cre- Thus, issues about institutional pre-conditions associ- ation often occurs following a disaster, as emergency and ated with the protection, pace, and sequencing of reforms reconstruction activities bring employment, especially in to prevent increases in poverty, larger inequalities, and the construction sector. socioeconomic instability cannot be ignored (Stiglitz Considering the above and including substitutions, 2002; Nayyar 2002; Mansoob 2002; World Bank 2001). market reactions, and other endogenous and exogenous The business cycle is a sequence of sustained upturns responses, the net effect of a disaster situation (consider- and downturns in GDP and employment associated ing both impact and responses) normally appears to have with economic shocks and/or agents' decisions that affect no negative macroeconomic effects. Expected negative aggregate demand that in turn is mediated by a collec- effects on GDP and investment, employment and infla- tion of not-well-understood societal factors and expec- tion,thetradedeficitandforeignreserves,arelargelyabsent tations of an economic and political nature. Until recently, and in the short term there is often an economic upturn the seeming absence of a synchronic cycle in OECD of some two years due to the reconstruction and business countries smoothed and softened the international busi- opportunitiesbroughtbyadisaster.Thepublicdeficitmay ness cycle, though recently the domestic cycles of OECD rise as the government finances rehabilitation and recon- countries have become more synchronic and depend- struction activities, but this rarely results in a long-term ent upon phases in the U.S. economy. Business cycles problem. Hence, one conclusion is that disasters may be of open developing economies have always been syn- "a problem of development, but not necessarily a problem chronic with that of the main OECD trading partner, for development" (Albala-Bertrand 1993a).1 The question but the possibility of diversifying trading partners has then is whether globalization affects this pattern. existed over time and this has reduced vulnerability to single-partner economic cycles. Globalization could provide similar stability through the diversification of Globalization and the Business Cycle exports and market transparency. It seems, however, that such diversification could become less successful Globalization involves a wider and deeper merging and as the cycle becomes more synchronic due to global- interdependence of domestic economies into a world- ization and hence more countries are economically wide arena of competitive multimarkets and exchanges. linked; then a major urban disaster in a developing It encompasses the flows of goods (trade), capital (finance country might have greater domestic impacts than had and direct), information (transparency and access), labor previously been experienced. The next section looks at (including human capital), and culture (institutional this hypothesis within a disaster situation framework. patterns and dynamics). The expectations of the advo- cates of globalization are that as more countries, and agents within countries, join this movement, obstacles Disaster Situation in Urban Areas to access will be reduced, and transportation and information costs will decrease. Participants will have A disaster situation is an analytical framework in which a wider and more penetrating reach in the worldwide disasters are composed of three interdependent processes arena, and globalization will benefit all. of one and the same phenomenon: a disaster impact, a Urban Disasters and Globalization 77 disaster response, and a societal interference wrought access, and rights) and environmental degradation (pol- by the former two. Thus, globalization and business lution and overcrowding). cycle synchronization may affect any of these phases. The lack of political influence and economic alter- Their potential effects can be analyzed by focusing on natives, poverty, and overall disenfranchisement may each of them in turn. A society is a living organism, be at the root of vulnerability, urban or rural. Increased which by its very nature responds endogenously when vulnerability and disaster risk can often be attributed disaster strikes. Human and societal responses, associ- to the wholesale policy rearrangements demanded and ated with in-built institutions, will then occur follow- imposed by a globalization process indifferent to soci- ing a disaster (see Albala-Bertrand 1993a). eties. This indifference is manifested in policy incon- sistencies in which institutional rearrangements are imposed with a pace and depth that are faster and Disaster Impact and Effects more far-reaching than the ability of people, especially the vulnerable, and activities to adapt and re-accom- A disaster impact is normally the result of a physically modate with a minimum of stability. This often places or socially uncompensated tension that leads to fatali- people and their livelihoods in a precarious condition ties, damage, destruction, and the disarticulation of soci- and safety vacuum. etal frameworks. In the case of natural disasters, the Disaster Impact Effects. For all kinds of disasters, once uncompensated tension is due to the physical weakness a disaster impact has occurred, two main types of effects of structures and societal processes that fail to take into ensue: direct or stock effects and indirect or flow effects. account and plan for extreme natural events. As such, Direct effects impact human populations (injury and even if the natural event were fully exogenous to soci- deaths) and physical and animal stocks (damage and ety, which might not be the case, physical resistance to destruction). In turn, indirect effects derive from the the geophysical phenomenon would not be. Resis- former, affecting the interrelation between physical struc- tance depends upon disaster-proof technology and tures and between people. These two types of effects sociopolitical access to it, a mostly endogenous process cause losses to society's stocks and flows. For socially to society at large. It is similar for technological disas- made disasters such as complex emergencies and tech- ters, but here the inducing phenomenon is also fully nological hazards, however, there is a third type of effect. endogenous, in-built in social structure and location This is an institutional effect where institutional failure and management of industrial facilities. Technological builds up vulnerability until a triggering event unleashes failures are associated with institutional failure to ensure a devastating breakdown. the safe production, containment, and use of risky Indirect effects can be broken down into four frame- technology. This brings to the forefront the issue of works: householdconditions (homelessness, shortages, dis- vulnerability. placement, livelihood erosion); the states of health and Generally, vulnerability is the exposure of physical nutrition of the population (environmental degradation, and societal frameworks to violent events. Exposure is hygiene problems, disease increase, food scarcity); the in turn associated with the risk of item or framework economic circuit (effects on intermediate markets, final failure. Societal vulnerability can then be defined more markets, policy, and expectations); and public activities specifically as the exposure of institutions and organ- (overburden, discontinuities, fragmentation, politiciza- ized people to violent and extreme events. The main tion). These frameworks are not fully independent of each factors influencing vulnerability to natural and tech- other,buttheformertworelatemoredirectlytothehuman nological hazards are unsafe living quarters (building condition, including basic needs and welfare, while the quality and location) and unsafe economic activities (engi- latter two relate indirectly to people, but directly to the neering quality and location of structures and risky social system as a whole. With some qualification, these processes). In turn, the main societal factors that may general effects are common to all types of disasters. increase the likelihood and destructiveness of disasters Regarding indirect impacts, disasters appear to be areentitlementerosion(economicandpoliticalpossessions, more remarkable for the effects they do not have than for 78 Building Safer Cities: The Future of Disaster Risk those they do. Especially in urban areas, there is little evi- One of the main planks of globalization is the support dence that a disaster's negative impacts are long-lasting and development of domestic financial markets to receive even if the direct effects dramatically affect some vul- foreign investment and tap international funds. Setting nerable social strata. Still, only an effective emergency aside the issue of domestic regulation of foreign finan- response may guarantee that the disruptions to society cial flows, a more developed financial market would and the negative impacts and their potential effects are include mechanisms to spread and reduce risk and not only short-lived but the emergency itself is not waste- vulnerability. If so, the impact of a natural or techno- ful (see Albala-Bertrand 1993a; 1993b; 2000a). logical disaster could be reduced or even eliminated How could globalization affect the above potential through improved loan access and insurance coverage. effectsinurbanareas?First,urbandisastersdonotstrongly These mechanisms may help larger businesses and impact exports, since export activities in most develop- wealthier individuals, though not everyone has the ing countries are overwhelmingly associated with pri- means to borrow following a disaster. Insurance and mary activities such as agriculture, mining, and fishing. reinsurance industries, however, are likely to develop An urban disaster cannot affect these exports directly, significantly due to globalization and risk can be spread and there is little reason to assume that indirect effects more evenly. Even more important, wider availability on this sector would be significant except in the very of insurance would act as a check on construction activ- short term. The weakness or lack of manufacturing ities, location, and technology use, as these are gener- exports in most developing countries is not due to ally preconditions for insurance coverage. Corruption backwardness or general deindustrialization in favor of in the use of land, materials, and the design of struc- the service sector, but to clear-cut globalization policies tures might also be reduced. The same would also be that lead to this result, at least for the foreseeable future true for riskier technologies, where discipline and mon- (Weiss 2002; Albala-Bertrand 1999; Chang 1996). There- itoring could be increased. Globalization therefore might fore, if globalization maintains the same trends, we should bring a wealth of mechanisms to increase prevention expect countries that join globalization to become even and distribute risk. more primary-sector dependent than they are now. In But these positive developments are hindered by terms of this increase in dependency, the issue is whether the ability of all people to afford insurance. Currently, globalization can create conditions to reduce some of a the majority of vulnerable people are excluded from the disaster's direct and indirect effects through the reduc- insurance market. Much of this exclusion is created by tion of losses to private and public buildings, dwellings, the transition costs of globalization, which may not be structures, and machinery and equipment. prepared to protect the real-time losers at any one The main mechanisms to reduce the physical vul- time. In addition, even if no one were excluded, the nerability of a built-up environment are regulations, experience with natural disaster insurance shows that building codes, land use restrictions, and the siting of people and small firms rarely take it up, except after risky technology. Assuming that the knowledge of these major disasters. People do not normally reflect on the aspects is sound, their effectiveness will depend upon possibility of large losses coming from very unlikely legislation and transparency. Globalization, via dissem- events (Kunreuther 1997; Giarini 1984). ination of information about best practices, may con- Synchronization of the business cycle may cause some tribute positively to a better understanding and drafting unintended negative effects by reinforcing insurance of disaster-related legislation. Globalization also assumes weaknesses in recessions and ignoring them in upturns. an increasingly more open society and greater trans- During recessions, access to insurance may be easy, parency, which may favor the observance and applica- but the capacity to pay policy premiums is reduced. In tion of legislation to reduce unchecked corruption. market booms, the insurance industry may down- Adoption and implementation of new legislation, how- grade de facto their standards so as to make inroads in ever, depend more on a society than on globalization. a tight market. Therefore, globalization notwithstand- There is another, safer area where globalization may ing, the potentially good results in this story would all play a positive role in reducing potential impact effects. depend on state backing, tight regulations, and penalty Urban Disasters and Globalization 79 enforcement. In other words, the freer and wealthier emergent coalitions appear to lie somewhere in between. the market, the stronger and more all-embracing should Exogenous mechanisms, in turn, are those channeled be the probity of the state and its regulation of economic via ad hoc, unpatterned, unguaranteed, and irregular activity. processes. They are expressed through actions, meas- ures, and policies that formally fill gaps left by in-built responses, by-pass endogenous channels, shift initia- Disaster Response and Mechanisms tives away from regular actors, and superimpose alter- native structures. This normally implies private and Disaster Response. Disaster response can be defined as a public interventions that go beyond in-built actions and widearrayofendogenousandexogenousreactions,meas- international assistance and aid that goes beyond ures, and policies that mitigate, counteract, and prevent existing guarantees. In the long run, however, these two disaster impacts and effects. Response to a disaster can response types might not necessarily be independent. be described as follows: once a disaster has occurred, This is because the endogenization of societally useful the impacts stimulate the unfolding of systemic response exogenous initiatives, actions, and behaviors, via edu- mechanismsandthecreationofspeciallydesignedresponse cation, policy, and social interaction, is the normal way measures. These two sets of responses aim temporarily in which society strengthens and develops (see Albala- to counteract functioning flow losses through emer- Bertrand 1993a; Cuny 1983; Davis 1981; Quarantelli, gency relief and rehabilitation activities and permanently 1978; White 1974; Barton 1970; Dynes 1970; Sorokin compensate stock losses and institutional insufficien- 1942; Prince 1920). cies through reconstruction activities. The impact effects Compensatory and Anticipatory Responses. In sudden and derived compensatory responses also stimulate an natural disasters and technological failures, reversing anticipatory responseaimed at the prevention and mit- the negative impacts is possible once the emergency igation of future potential disasters (Albala-Bertrand response has contained the spread and worsening of 1993a). These responses generate three main areas of indirect effects. These responses normally require a attention, which make up the response side of a disaster significant amount of public involvement and public situation: (i) response mechanisms; (ii) compensatory finance, and in developing countries, foreign aid and response; and (iii) anticipatory response. In addition, credits are also required (Kunreuther 1997; Albala- as a disaster situation always generates varying degrees Bertrand 1993a). These responses and the inflow of of societal interference, we should also focus on response- funds, materials, and finances interfere with normal induced interfering effects. activities that compete for the same resources. In addi- Response Mechanisms. As introduced earlier, response tion, given the societal endogeneity of technological fail- mechanisms refer to endogenous and exogenous response ure, intense critique and reassessment of the control processes. Endogenous response mechanisms are and handling of risky technology, which also prompts those channeled through society's in-built institutional an institutional reaction, will also arise. Anticipatory processes. These processes represent a series of formal actions in turn may also have strong societal implica- and informal feedback mechanisms that are part of the tions as they aim to modify behavior and institutions existing self-regulatory social machinery (e.g., the family, either to prevent disasters or to respond efficiently when informal finance, the informal sector, formal markets, prevention fails. Prevention also includes formal insur- political and administrative frameworks, cultural norms ance, mortgage, and taxation systems that encourage and customs, psychological attitudes and habits). These less risky behavior and a better distribution of risk to involve a wide array of activities that range from highly reduce negative disaster impacts (Kunreuther 1997; automatic to nonautomatic in-built responses. For Albala-Bertrand 1993a; Giarini 1984; Cochrane 1975; example, extended family solidarity represents a highly Dacy and Kunreuther 1969). These actions also include automatic endogenous reaction, while the use of the the monitoring of markets, migration, and reactions to hazard reserve item of the public budget is a planned maximize response effectiveness and minimize antiso- and calculated response. Likewise, market reactions and cial and speculative behavior. 80 Building Safer Cities: The Future of Disaster Risk How can globalization affect the above response mech- is a concerted international effort to improve the sound- anisms and their effectiveness? Community, defined as ness and safety of globalization policies as an aim in a stable array of institutions that set useful societal inter- itself, so that vulnerable people do not suffer as a result action and hierarchies within and between particular of globalization and a disaster. identity groups, like family, neighborhood, workplace, On the other hand, if globalization makes the busi- and formal and informal working relationships, might ness cycle synchronic and there is a recession in the be the first casualty of fast and unfettered globalization United States and other industrialized countries, then (Stiglitz 2002). There is growing evidence that quick transitional economies would suffer in three ways. First, trade liberalization makes small and struggling urban the recession would reinforce the negative situation firms uncompetitive and unviable. These would affect described earlier. Second, as export demand, commodity formal firms and their workers as well as the informal prices, and capital flows fall off, there would be fewer economic activity that depends on these firms, which domestic financial resources and already-depressed com- may represent the overwhelming majority of urban eco- munities would be further harmed, weakening endoge- nomic activity in many developing countries (Thomas nous response mechanisms. And third, with a worldwide 1990). In addition, the current globalization push for recession, the availability and willingness of countries privatization, deregulation of labor markets, and the to give foreign aid might be reduced. (Although if the general restructuring of firms seeking efficiency and world economy were in an upturn, international donors productivity improvements, without heeding the might be more generous, but developing countries might ensuing social costs, will make matters even worse for also be in better financial positions.) Reconstruction a precarious social fabric. If there are neither alterna- requirements, however, might put a strain on other activ- tive livelihoods nor public protection for potentially ities, exportable or not, which might affect the econ- affected people and the transition to higher employ- omy indirectly. ment and stability is slow, the informal and formal A country in recession might have more idle capital endogenous mechanisms might be badly impaired at resources to serve rehabilitation and reconstruction. the time of a disaster impact. Such activities might create a domestic demand stim- In addition, the other plank of globalization, the lib- ulus, not limited to domestic activities related to dis- eralization of capital, is now known to create negative aster response, and positively affect the whole economy. economic shocks and instability associated with This possibility is not always the case, however, since unregulated capital flows. If this also comes to the fore, most capital is not fungible and cannot be switched to then any respite from poverty will be short-lived. other types of production in the short run or even the There will also be little in the way of reestablishing the medium term. Second, if the economy were significantly stability and strength of endogenous response mecha- open, then most output would be geared for exports. nisms, let alone of improving them. Therefore, global- The export sector and its backward and forward link- ization as it has been conducted to date may significantly ages would then be operating with significant idle capac- weaken endogenous response mechanisms at the time ity. Domestic demand might normally be satisfied with of disaster, thus demanding a stronger exogenous only a small fraction of these exportable goods, but then response from domestic and foreign sources when dis- the level of domestic demand would also be affected. aster strikes. This would make disaster-induced expenditure less For example, if a country is facing economic diffi- effective as a mechanism to compensate the economy culties due to the requirements imposed upon its than it would otherwise have been. economy and society by the globalization process, Mostdisasterlegislationandarrangementsfollowmajor then it is unlikely that the domestic economy and budget disasters. Fast globalization, of the current type may would have enough laxity to respond appropriately to make the social fabric more unstable and fluid than it a hazard. In an ideal world, we would then expect that would otherwise have been. Fluidity may offer the oppor- foreign aid would flow from institutions in support of tunity and impetus to restructure institutions. Under globalization. This will not happen, however, unless there what type of social contract would these rearrangements Urban Disasters and Globalization 81 be incorporated? This would depend most impor- dependence upon foreign aid would also be less nec- tantly on the type of society in question, followed by essary than in downturns. the disaster response. Instability and its results might make it more difficult to set up an organized type of institutional rearrangement, so authoritarian rule might Conclusions come to the fore; this form of governance does not normally favor the most vulnerable people, hence the In the context of a disaster situation, as described above, overall welfare may be worsened for a considerable the following conclusions can be reached: period of time. · Urban disasters affecting large cities and megacities Visible Response Type and Origin. Notice first that a may impose large residential and infrastructure losses, good deal of the endogenous response is not visible or as well as large death tolls and injuries. amenable to quantification, which may normally under- · Losses of capital and activities, deaths, and injuries estimate the energy and effort made by society itself in are unlikely to affect the macroeconomy negatively. the wake of a disaster (Albala-Bertrand, 1993a). The · It is unlikely that this pattern would significantly visible response usually comes in the form of finance, change by virtue of the negative features of global- materials, technical expertise, labor, and organizations ization, but this does not mean that the people and to manage such resources. The main response sources activities affected by urban disasters would not be are domestic and foreign. The former can be subdivided victimized. into local and national, and the latter into bilateral, mul- · The negative features of globalization might make a tilateral, private financial institutions, NGOs, and remit- significant difference for increased victimization, as tances. Commonly, the majority of resources are domestic the endogenous mechanisms of response are likely in origin. However, significant expertise and financial to be badly impaired. resources from international sources are sometimes piv- · Given that globalization appears to make the cycle otal in disciplining the general response. synchronic and dependent on the U.S. economy, When disaster strikes, the stage of globalization and financing disaster response might become procycli- phase of the business cycle can affect funding sources cal, more adversely affecting a disaster-struck coun- andmechanismsinvariousways.First,acycle-synchronic try in the event of a recession. recession might significantly affect local and domestic · Globalization could provide new opportunities for resources. This would also affect informal financial mar- diversifying risk and improving prevention, but the kets, which might become less flexible and effective in useful incorporation of these opportunities into the the wake of fast globalization, impairing recovery. Second, economy and polity would depend more on the remittances from abroad would be strongly affected. type of domestic society than on globalization itself. As this is usually a very important type of informal finan- Furthermore, given that the macroeconomy might cial response at the family level, recovery of household not suffer as a result of an urban disaster, communities and individual livelihood conditions would likely be and activities directly affected should be the main target further impaired. Third, both bilateral and multilateral of response policies. International response organiza- sources might become strongly procyclical, significantly tions can also participate, as they often attempt, in the reducing their role in recessions. Fourth, NGOs depend endogenization of initiatives dealing with preventive on donations from a variety of people and organiza- and compensatory mechanisms at grassroots and national tions. Donations might dry up with a synchronic reces- levels. These initiatives might bring strong political sion. Finally, international private funding sources might opposition from entrenched political elites who may be undergoing excess liquidity, which could con- perceive them as interference, but this should not be a tribute to easy, but risky, lending. The latter could be a reason to drop them. In addition, organizations that short-term blessing but a long-term disaster if additional make up part of the globalization movement should debt were taken on. On the other hand, during upturns, take into consideration the social and economic prob- the concerns mentioned might be insignificant, but lems that arise as a result of their policies, so that they 82 Building Safer Cities: The Future of Disaster Risk can design safer and more stable approaches to global- Dynes, R.R. 1970. Organized Behaviour in Disaster. Lexington, ization in general and to hazards in particular. MA: Heath & Co. ECLAC (Economic Commission for Latin America and the Caribbean). 1999. Manual for Estimating the Socio-Economic Notes Effects of Natural Disasters. Santiago, Chile. 1. For an analytical explanation of this situation illustrated by Giarini, O., ed. 1984. The Geneva Papers on Risk and Insurance. Latin American countries see Albala-Bertrand 1993b. Geneva: The Geneva Association. Kunreuther, H. 1997. "Rethinking Society's Management of Cat- astrophicRisks."GenevaPapersonRiskandInsurance83:151­176. Bibliography Mansoob, M., ed. 2002. Globalization, Marginalization and Devel- Albala-Bertrand, J.M. 1993a. The Political Economy of Large Nat- opment. London: Routledge. ural Disasters. Oxford: Clarendon Press. Nafziger, W., and R. Vayrynen, eds. 2002. ThePreventionofHuman- Albala-Bertrand, J.M. 1993b. "Natural Disaster Situations and itarian Emergencies. Houndmills: Palgrave. Growth: A Macroeconomic Model for Sudden Disaster Impacts." Nayyar, D., ed. 2002. Governing Globalization: Issues and Institu- World Development 21(9):1417­1434. tions. Oxford: Oxford University Press. Albala-Bertrand, J.M. 1999. "Industrial Interdependence Change Prince, S.H. 1920. "Catastrophe and Social Change." Studies in in Chile 1960­90. A Comparison with Taiwan and South Korea." History, Economics and Public Law 94/1. International Review of Applied Economics 13(2):161­191. Quarantelli,E.L.1978.DisasterTheoryandResearch.London:SAGE. Albala-Bertrand, J.M. 2000a. "Complex Emergencies versus Nat- Rossi, P.H. and others. 1982. Natural Hazards and Public Choice. ural Disasters. An Analytical Comparison of Causes and Effects." New York: Academic Press. Oxford Development Studies 28(2):187­204. Smith, K. 1992. Environmental Hazards: Assessing Risk and Reduc- Albala-Bertrand, J.M. 2000b. "Responses to Complex Human- ing Disaster. New York: Routledge. itarian Emergencies and Natural Disasters. An Analytical Com- Sorokin, P.A. 1942. ManandSocietyinCalamity. New York: Dutton. parison." Third World Quarterly 21(2):215­227. Stiglitz, J. 2002. Globalization and its Discontents. London: Allen Barton, A. 1970. Communities in Disaster. A Sociological Study Lane. of Collective Stress Situations. Garden City, MI: Anchor, Stewart, F., F.P. Humphreys, and N. Lea. 1997. "Civil Conflict Doubleday. in Developing Countries Over the Last Quarter of a Century: Chang, H. 1996. The Political Economy of Industrial Policy. London: An Empirical Overview of Economic and Social Consequences." Macmillan. Oxford Development Studies 25(1):11­41. Cochrane, H.C. 1975. Natural Disasters and their Distributive Thomas, J. 1990. Informal Economic Activity. London: LSE Effects. Boulder, CO: Institute of Behavioral Science. Publications. Cuny, F.C. 1983. Disaster and Development. Oxford: Oxford Weiss, J. 2002. Industrialisation and Globalization. London: University Press. Routledge. Dacy, D.C., and H. Kunreuther. 1969. The Economics of Natural White, G.F., ed. 1974. Natural Hazards: Local, National, Global. Disasters. New York: Free Press. Oxford: Oxford University Press. Davis, I., ed. 1981.DisasterandSmallDwelling.Oxford: Pergamon World Bank. 2001. Globalization, Growth, and Poverty. Washington, Press. D.C. Chapter 6 Interdependent Disaster Risks: The Need for Public-Private Partnerships Howard Kunreuther Do individuals or businesses have economic incentives to Suppose there are two identical adjoining homes H1 carryoutsociallyappropriatelevelsofmitigationforreduc- and H2 where there is a probability p that a severe earth- ing future disaster losses? This paper shows that property quake will occur in their region. Currently both homes owners will systematically underinvest in protection due are not mitigated, so that if an earthquake occurs, the to the presence of interdependent disaster risks1 damage to each of the structures will be L. The cost of These are situations where damage to property or a mitigation measure is c. If both homes invest in pro- infrastructure creates direct or indirect losses to tection, then the damage from an earthquake to each others. For example, the collapse of a building from an of them is assumed to be 0. On the other hand, if only earthquake could cause severe damage to an adjoining H1 invests in mitigation, then there is a probability q < p structure. Kunreuther and Heal (forthcoming) have that H2 will suffer earthquake damage, which causes a demonstrated that this type of interdependency may loss L to H1. This might occur, for example, if a water negate much of the benefits a property owner would heater in one house was not bolted, an earthquake caused otherwise obtain from adopting mitigation measures. it to topple over, and the resulting fire spread to the The following three questions related to this interde- neighboring house. The decision facing each family is pendency are addressed below: whether to invest in mitigation, knowing that the other · What incentives do property owners have for house may not be protected. investing in protection when their neighbors have Assume that each homeowner has initial assets of A, not done the same? perfect information on the risks and costs of protection, · What types of public-private-sector partnerships and has to make a choice between investing in protec- are necessary for the adoption of cost-effective mit- tion, Y, or not to do so, N. Table 6.1 shows the outcomes igation measures? to each individual A based on the four possible strategies: · How can one deal with problems of interdependen- The rationale for these payoffs is straightforward: if cies in emerging economies such as Turkey's? both invest in protection (Y, Y), then each incurs a cost of c and faces no losses, so each homeowner's net assets Incentives for Investing in Protection is A - c. If H1 invests and H2 does not [(Y, N)--top right entry], then H1 incurs a cost of c and also runs the A family is considering investing in a structural miti- risk of a loss emanating from H2. The probability of H2 gation measure to reduce losses from a future earth- contaminating H1 is q, so that H1's expected loss from quake. In making this decision, the family needs to damage to H2 is qL. This cost represents the negative balance the cost of the mitigation measures with the reduction in earthquake damage to its house. As Table 6.1 Expected outcomes associated with investing shown below, the incentive to invest in mitigation is and not investing in protection greatly diminished if surrounding homes fail to adopt Homeowner 1 (H1) Homeowner 2 (H2) protective measures. The challenge is to find ways to Y N Y N convince all property owners that it is in their best inter- A - pL, A - [pL + (1 - p)qL], A - c, A - c A - c - qL est to invest in protection. A - c - qL, A - [pL + (1 - p)qL], A - pL 83 84 Building Safer Cities: The Future of Disaster Risk externality imposed by H2 on H1. The lower left pay- Table 6.2 Illustrative example: Expected costs associ- offs (N, Y) are just the mirror image of these. If neither ated with investing and not investing in protection house invests in protection, (N, N), then both have an Homeowner 1 ( H1) Homeowner 2 (H2) expected payoff of A - pL - (1 - p)qL. Y N Y N Now that the outcomes have been specified, one A - 100, A - 145, A - 98, A - 148 can ask the natural question: under what conditions A - 148, A - 145, A - 98, A - 100 will the homeowners invest in protection? It is clear from table 6.1 that, for investment in protection to be loss even if it protects itself. The expected benefits to H1 a dominant strategy so each homeowner will want to of investing in security will now only be pL(1 - q) = 95, invest in protection no matter what the other home- which is less than the cost of the security measure. owner does, we need: Hence H1 will not want to invest in protection. In other words, either both agents invest in security or neither A - c > A - pL of them do so. These are the two Nash equilibria. and A - c - qL > A - pL - (1 - p) qL Kunreuther and Heal (forthcoming) show that, as The first inequality just says that c < pL:, the cost of invest- one increases the number of homes subject to damage, ing in mitigation, must be less than the expected loss, the incentive for any house to protect itself against a the normal benefit-cost condition for an isolated loss gets progressively worse. Imagine many homes, home. The second inequality is more interesting: it each one of which could cause a fire if it were unmiti- reduces to c < pL - pqL = pL(1 - q). This is clearly a gated that would spread to all the other structures in tighter inequality, reflecting the possibility of damage the neighborhood whether or not they had invested in from the other home to yours. This possibility reduces a loss-reduction measure. If there are enough homes the incentive to invest in protection. Why? Because in that are unprotected, then there is a high likelihood that isolation, investment in mitigation buys the homeowner a protected home may still be damaged from one of complete freedom from risk; with the possibility of con- these structures. Unless c is very low, there will be no tagion, it does not. Even after investment in mitigation incentive for any homeowner to invest in mitigation if there remains a risk of loss emanating from the other he knows that most neighbors have not done so. unprotected home. Investing in mitigation buys you less A similar interdependency exists between the damage when there is the possibility of contagion from others. to a region's infrastructure from a disaster and the deci- In the two-agent problem with identical costs, one sions by property owners served by these lifelines to can determine the optimal behavior of each homeowner invest in mitigation measures. Suppose that a business if they both make decisions simultaneously without any knows that there is a good chance that if a hurricane communication. In this noncooperative environment wreaks havoc in its area, the electric utilities and/or water if c < pL(1 - q), then both agents will want to invest in supply will be nonfunctional for a considerable period protective measures (Y, Y); if c > pL then neither agent of time. Even if the business invests in mitigation to will want to invest in protection (N, N). If pL(1 - q) < reduce physical damage to the structure, the owners of c < pL then there are two Nash equilibria (Y, Y) and the firm know that it faces the risk of business inter- (N, N) and the solution to this game is indeterminate. ruption due to damage to the community's infrastruc- The solution concept for two agents with identical ture. This type of contagion will discourage it from costs and risks is illustrated below with a numerical exam- investing in some protective measures that it would ple. Suppose that p = .1, q = .05, L = 1000 and c = 98. The otherwise undertake had the community lifelines been matrix in table 6.1 is now represented as table 6.2. If H2 sufficiently protected to be functional after a disaster. invests in security (Y), then it is worthwhile for H1 also to invest in security, since without protection its expected Need for Public-Private Partnerships losses will be pL = 100, and it will only have to spend 98 to eliminate this risk. If H2 does not invest in security One way to encourage property owners to invest in secu- (N), then there is still a chance that H1 will experience a rity when they face the possibility of contagion from Interdependent Disaster Risks: The Need for Public-Private Partnerships 85 others is to internalize the externalities. This section of decisionmaking from the regulator to the property briefly examines the roles that different policy tools-- owners, who are now required to do their own plan- ranging from private market mechanisms to govern- ning as to how they will meet a set of standards or reg- ment regulations to collective choice--can play in ulations (Coglianese and Lazer 2001; Kunreuther, encouraging agents to adopt protective measures for McNulty, and Kang 2002). interdependent security problems. Kunreuther and Heal (forthcoming) examine the role thatdifferentprivatesectormechanismscanplayinencour- Dealing with Interdependencies in Emerging aging investment in mitigation. They show that private Economies insurance cannot encourage investment in protection because the insurer is only responsible for the damage to Emerging economies face significant challenges in deal- its own insured property. One reason for this contractual ing with problems of interdependencies because many arrangement between insurer and insured is the diffi- buildings have been too poorly designed to withstand culty in assigning causality for a particular event. On the the impacts of natural disasters, and the limited govern- other hand, consider a social insurance program that mental personnel are not adequate to enforce building provided coverage to all property owners. It would want codes and other regulations for reducing future losses. to develop a premium structure that encouraged invest- These points are clearly illustrated by the case of Turkey. mentinmitigation,sincethegovernmentwouldberespon- The city of Istanbul is very likely to experience strong sible for covering damage to all insured property. shaking from a large earthquake in the Marmara Sea The possibility of contagion from other units pro- during the next 30 years (Parsons and others 2000). vides a rationale for well-enforced building codes that Subsequently, most of the current citizens of Istanbul require property owners to adopt cost-effective protec- (the majority of the city's population is under the age tive mechanisms when they would not do so voluntar- of 35) are likely to experience this event. Without prepa- ily. More specifically, building codes solve the coordination ration and prevention, the people of Istanbul face a high problem in the sense that all individual property owners risk of suffering significant losses from earthquake are better off by adopting these loss-reduction meas- damage, including loss of life. The Greater Metropoli- ures. When asked whether they would do so voluntar- tan Municipality is taking steps to address mitigation ily, they say "no" because they either know or assume for critical infrastructure, as well as providing public that others will be unprotected. education concerning nonstructural approaches to There may also be a need for well-enforced regula- risk reduction in homes. However, there are no steps tions if there were externalities to other parties in addi- in place to address the structural fragility of the city's tion to the contagion effects between the agents. For thousands of residential apartment buildings. example, when a building collapses it may create Based on an incomplete engineering inventory, approx- externalities in the form of dislocations and other imately 5,000 of these buildings have been assessed as social costs that are beyond the economic losses suf- likely to experience complete structural failure when fered by the owners. These may not be taken into account subjected to strong shaking, risking total loss of life of when the owners or developers evaluate the importance occupants of these buildings. Another class (40,000+) of adopting a specific mitigation measure and, hence, are likely to experience significant structural damage, may justify the need for building codes (Cohen and Noll also with the potential to cause death or serious injury. 1981; Kleindorfer and Kunreuther 1999). Poor performance of buildings in areas east of Istanbul, One way for the government to enforce its regula- with similar construction design and quality, was demon- tions is to turn to the private sector for assistance. strated in 1999 by two very severe earthquakes. A com- More specifically, third-party inspections coupled with pulsory earthquake insurance program for homes has insurance protection can encourage property owners to been initiated in an attempt to share the expected cost reduce their risks from accidents and disasters. Such a of reconstruction among the population. This financial management-based regulatory strategy shifts the locus instrument, however, only deals with losses following a 86 Building Safer Cities: The Future of Disaster Risk disaster and does not have a premium structure that pro- equity considerations. This issue is particularly impor- vides economic incentives for property owners to invest tant in emerging economies, where explicit ways of deal- in measures to reduce the risk (Gulkan 2001). The chal- ing with the poverty problem and how to deal with lenge facing Turkey and other emerging economies is low-income families who cannot afford to invest in mit- how to finance and enforce these mitigation measures. igation measures unless they are highly subsidized by Benefit-cost analyses of retrofitting measures suggest funds from other sources must be considered. that it will benefit the potential victims and lead to an improvement in social welfare (Smyth and others forth- coming). Institutions such as the World Bank have a Notes stake in promoting cost-effective mitigation measures 1. Partial support for this research comes from the U.S. Envi- because such measures promise to save them consid- ronmental Protection Agency under Cooperative Agreement erable expenses that they would otherwise incur fol- C R 826583 with the University of Pennsylvania, the Wharton lowing a future disaster. Developed countries also have Risk Management and Decision Processes Center, and the Colum- a stake in the outcome because of their interest in ensur- bia University Earth Institute. ing that emerging economies such as Turkey's are on a sound economic footing. Bibliography Future Research Coglianese, Cary and David Lazer. 2001. "Management-Based Regulation: Using Private Sector Management to Achieve Public Theissuesdiscussedabovealsosuggestanumberofempir- Goals." Regulatory Policy Program Working Paper RWP01- ical studies on interdependent security. Given the concern 047. Cambridge, MA: John F. Kennedy School of Government, with terrorism both in the United States and the rest of Harvard University. the world, it would be interesting to learn more about Cohen, Linda and Roger Noll. 1981. "The Economics of Build- what factors lead some individuals and organizations to ing Codes to Resist Seismic Shocks." Public Policy Winter:1­29. invest in security and why others are reluctant to do so. Gulkan, Polat. 2001. "Revision of the Turkish Development Law What institutional mechanisms would aid the decision- No. 3194 Governing Urban Development and Land Use Plan- makingprocessofagentsconsideringprotectivemeasures ning." In Paul Kleindorfer and Murat Sertel, Mitigation and in situations where others would be affected? What are Financing of Seismic Risks: Turkish and International Perspectives. the appropriate roles of the public and private sectors in Dordrecht: Kluwer Academic Publishers. developing strategies that include economic incentives Kleindorfer, Paul and Howard Kunreuther. 1999. "The Com- (fines or subsidies), third-party inspections, insurance plementary Roles of Mitigation and Insurance in Managing coupled with well-enforced regulations, and standards? Catastrophic Risks." Risk Analysis 19:727­38. Proposed solutions depend on the nature of the dis- Kunreuther, Howard, Patrick McNulty, and Yong Kang. 2002. aster and the current institutional arrangements in the "Improving Environmental Safety Through Third Party Inspec- country or region of interest. They also require data on tion." Risk Analysis 22:309­18. thelikelihoodoftheeventoccurringanditsconsequences, Kunreuther, Howard and Geoffrey Heal. Forthcoming. "Inter- as well as the uncertainties and ambiguities associated dependent Security." Journal of Risk and Uncertainty. with these estimates. In developing strategies for dealing Parsons,T., S. Toda, R. Stein, A. Barka, and J. Dieterich. 2000. with these problems, one needs to consider the type of "Heightened odds of large earthquakes near Istanbul: An available information, the nature of the decisionmaking interaction-based probability calculation." Science 288:661­65. processes of the key stakeholders, and how their behav- Smyth, Andrew and others. Forthcoming. "Probabilistic Benefit- ior differs from normative models of choice. Cost Analysis for Earthquake Damage Mitigation: Evaluat- Finally, one needs to balance how one allocates ing Measures for Apartment Houses in Turkey." Earthquake resources in an efficient manner with distributional and Spectra. Traffic drives down flooded Manila street, 2002. PART II ENVIRONMENT, CLIMATE VARIABILITY, AND ADAPTATION Chapter 7 Cities and Climate Change Anthony G. Bigio This was prepared as a background paper for the World 0.5°C for the last 1,000 years. The latest report from Development Report 2003 and was presented at the World the Intergovernmental Panel on Climate Change (IPCC), Bank's Urban Forum in April 2002. Its four key messages "Third Assessment Report: Climate Change 2001" (IPCC are the following: 2001b), establishes a direct correlation between the sus- a) Cities in developing countries are severely threatened tained use of fossil fuels (that occurred primarily in by the impacts of climate change that already cost them bil- industrialized countries), the resulting accumulation of lions of dollars and compound other ongoing developmen- CO2 and other gases in the atmosphere, and global tal challenges; warming. b) Given the limited role of the developing world in the Global warming manifests itself in many different historical accumulation of greenhouse gases, there are increas- ways in the biosphere. The two most relevant to the sub- ing expectations that financial assistance for adaptation to ject of this paper are the progressive rise in sea level and climate change should be provided by OECD countries; the increased intensity and frequency of climatic episodes c) There is emerging evidence that some of The World leading to natural disasters. Both represent a signifi- Bank's infrastructure projects are vulnerable to the impacts cant threat to urban areas in developing countries. of climate change and as a consequence may be unable to According to several projections, the sea level is meet their objectives and provide the full range of benefits expected to increase anywhere from 8 to 88 centime- expected when they were financed; and ters during the twenty-first century, mostly due to d) Given that some 30 percent of The World Bank's invest- thermal expansion and the loss of mass from glaciers ments are in urban areas, an increased ability by the Bank's and ice caps (IPCC 2001b: 16). The frequency and inten- infrastructure staff to incorporate climate change consider- sity of natural disasters are also growing rapidly world- ations into project design would play a positive role in assist- wide. A recent analysis of great natural catastrophes ing cities. since 1960 shows an increase in the 1990s by a factor of three, and this seems to be directly correlated with global warming. Economic losses adjusted for inflation Impacts of Climate Change on Cities rose by a factor of nine (Munich Re 1999). The rapid pace of urbanization, compounded with In the latter part of the 20th century, the earth's aver- an ever-increasing population burden, has also signif- age temperature rose 0.6° Celsius. Projections for fur- icantly increased the overall vulnerability of urban areas ther increases in the 21st century vary considerably, to natural disasters. The location of many urban areas from a minimum of 1.4°C to a maximum of 5.8°C. with large populations and critical economic assets in The increase will depend upon the level of stabiliza- high-risk zones contributes to the increased attention tion of carbon emissions, the pace of decarbonization given to impacts in urban areas of disasters induced or of the global economy, and the patterns of demographic enhanced by climate change (Munich Re 2000). In the and economic development. Such increases represent case of cities in developing countries, the size and vul- a dramatic shift with regard to natural variability in the nerability of informal settlements, generally built in planet's mean temperature, which has remained within unstable areas such as coastal zones, flood-prone planes 91 92 Building Safer Cities: The Future of Disaster Risk and ravines, and geologically unstable slopes, greatly results of modeling simulations, and other available increase their susceptibility to risk. The impacts of a information on the projected impacts of climate rising sea level and more frequent and intense natural change on urban areas (IPCC 2001a). According to the disasters in rural areas will likely generate an addi- IPCC report, the main threats to the urban populations tional influx of people to cities. Such "environmental and physical assets of developing cities, impacted with refugees" often become permanent, thus increasing the more or less intensity based on the actual climate changes vulnerability of cities yet again (Hardoy and others 2001). that unfold, are the following: Developing countries are considered to be particu- · A rise in sea level: This is the most fundamental chal- larly vulnerable to climate change, as many are in trop- lenge that urban settlements face from global warm- ical and sub-tropical zones with economies and societies ing. The threat will likely increase due to the ongoing highly dependent upon the climate and heavily impacted influx of people and economic assets into coastal by its variations. Many of the largest cities in Africa, zones. At risk are entire sections of coastal cities and Asia, and Latin America are port cities, historically linked their infrastructure, beaches subject to erosion, to a colonial past, and directly subjected to the impacts river floors in estuarine zones subject to sedimenta- of a rising sea level. The high cost of land in a central tion, and wetlands and tidal flats subject to flooding city and around ports has often encouraged major com- (box 7.1). Furthermore, groundwater risks increased mercial developments on land reclaimed from sea and salinization,andcoastalaquifersriskdiminishing,which river estuaries that are especially vulnerable to a rise in affect fresh water supplies and peri-urban agriculture. sea level (Hardoy and others 2001: 205). · Tropical cyclones: Increasingly frequent and intense The IPCC Third Assessment Report Volume 2: tropical and extra-tropical cyclones will likely cause "Impacts, Adaptation and Vulnerability," updates exist- severe wind damage and storm surges which, com- ing knowledge and provides integrated field studies, pounded with a rise in sea level, are expected to Box 7.1 Coastal cities and small island states By mid-century, more than 70 percent of the population in settlements that could be flooded due to a rise in sea level are likely to be located in West and East Africa, along the southern coast of the Mediterranean, and South and South- east Asia. With a 40-centimeter rise in sea level, the midpoint of the IPCC projection ranges for the end of the century, the world population at risk from annual flooding is expected to increase from the current 10 million to 22­29 million by the 2020s, to 50­80 million by the 2050s, and to 88­241 million by the 2080s (Nicholls and others 2001). The biggest impacts, however, are expected in the small island states of the Atlantic, Pacific, and Indian Oceans. Cities such as Alexandria, Egypt; Banjul, The Gambia; Tianjin, China; Jakarta, Indonesia; and Bangkok, Thailand will be affected. A 50-centimeter sea-level rise along Egypt's coastal zones would affect 2 million people and 214,000 jobs, and cause land and real estate losses worth US$35 billion (El-Raey 1997). Alexandria's Old City, 12 meters above sea level, is safe from the direct effects of a rise in sea level. However, the port area and newer suburbs are at risk since, with the aid of flood defenses, they were constructed on lowlands. Low marshes and lagoons that surround the city could be lost or seriously contaminated with saltwater due to a rise in sea level. Ultimately, the city could become a peninsula, surrounded by the Mediterranean and only accessible by bridges and causeways (Turner and others 1990). The average number of people in Africa affected by coastal flooding could increase from the one million of 1990 to a worst-case scenario of 70 million by 2080 (Nicholls 2001: 515). Through coastal erosion and a rise in sea level, Banjul, capital of The Gambia, could disappear by mid-century. East African coastal settlements are also at risk. In most small island states, coastal planes have provided the best locations for urban centers and population con- centrations, physical assets, economic activities, and services. On most Caribbean islands, for instance, more than half of the population lives within two kilometers of the coast. On atolls, the most important infrastructure and popula- tion clusters are less than 100 meters from the shoreline. The threats induced or enhanced by climate change will severely affect this group of nations and their cities (IPCC 2001a: 847, 864), which are becoming acutely aware of their vulnerability. Cities and Climate Change 93 become a severe problem for low-lying coastal regions · worsening urban air pollution exacerbated by and cities. Ports and other coastal infrastructure are increased ground ozone formation; especially at risk. · enhanced effects of urban heat islands due to higher · Flooding and landslides: Expected increases in the overall temperatures. scale, intensity, and frequency of rainfall in most The indirect impacts of such climatic threats are, of developing countries will severely strain or over- course, much wider. They include environmental health whelm the storm drainage systems of many urban problems due to the expected changes in geographic centers. This could lead to periodic flooding of low- ranges and the incidence of vector-borne and infectious lying areas as well as landslides and mud-slips on diseases, allergic and respiratory disorders, nutritional geologically unstable slopes, often subject to infor- disorders due to climate-related food shortages, and the mal settlements. Cities built next to rivers and on physical damage to and institutional strains imposed reclaimed lands in riverbed planes will be prone to upon the health care system (Hardoy and others 2001: additional inundations. 203). Where impacts are felt, urban economic activi- · Water quality and shortage: Urban flooding dam- ties will likely be affected by physical damage to infra- ages water treatment works and flood wells, pit latrines, structure, services, and businesses. There will also be and septic tanks. Sewage treatment systems and solid repercussions on overall productivity, trade, tourism, waste disposal areas can also be affected, contami- and the provision of public services. nating water supplies. Where overall rainfall decreases, droughts will likely compromise the replenishment of the water tables, the normal sources of water supply. Economic Valuation of Climate Change Impacts · Heat and cold waves: Intense episodes of thermal on Cities variability could severely strain urban systems by adding an environmental health risk for more vul- Of the major threats to cities described above, only the nerable segments of the population, imposing extraor- rising sea level has an exclusive correlation with climate dinary consumption of energy for heating and air change. The other four threats are natural disasters conditioning where available, and disrupting ordi- and climatic episodes that are exacerbated in their inten- nary urban activities. sity and frequency by climate change (box 7.2). Thus, In addition to these major threats, the IPCC report the major attempts at valuating, in economic terms, indicates additional risks related to: the impacts of climate change on physical infrastruc- · increased possibility of urban fires, severe hail, and ture generally refer to a rise in sea level and are based windstorms; on the funds needed to defend the coastal areas and · negative impacts on the productivity of fisheries replace physical assets. For instance, the cost of pro- and agriculture, on which some urban economies tecting port facilities and coastal structures, raising partially depend; wharves and quays, and reconstructing water gates Box 7.2 Natural disasters: what percentage is due to climate change? In 1982, Peru's GDP declined by 12 percent, half of which was attributable to the El-Niño-related floods of that year. Between 1989 and 1996, China experienced annual losses from natural disasters averaging 3.9 percent of GDP. In 2000, flooding in Mozambique resulted in direct and indirect losses of some 6 percent of GDP (Mimura and Harasawa 2001). Damage and losses from Hurricane Mitch in 1999 equaled 80 percent of GDP in Honduras and 49 percent in Nicaragua (FAO 1999: 409). These figures have not been disaggregated to show losses that occurred in cities as opposed to rural areas. Nor can we ascertain how many of these disasters or how much of their intensity was due to the inducing or enhancing effects of climate change. However, they help us understand the order of magnitude of the problems faced by cities in devel- oping countries due to climate change. Further detailed analysis is needed, however. 94 Building Safer Cities: The Future of Disaster Risk and pumping stations for a one-meter rise in sea level with a proven vulnerability to climate change, invest- (the high point of the IPCC projection ranges for the ments will likely require: end of the century) in 39 prefectures in Japan has been · "hardening up" of the infrastructure systems, including estimated at US$194 billion, or about 7 percent of annual storm-drainage systems, water supply and treatment GDP (Mimura and Harasawa 2001). plants with protective physical improvements; Conversely, the incremental nature of the intensity · protection or relocation of solid waste management and frequency of violent climatic episodes due to cli- facilities, energy generation and distribution systems; and mate change is hard to disaggregate from the "baseline," · consolidation of hydro-geologically fragile areas. even if scientific evidence has proven the correlation of Coastal cities will likely need to invest in heavy phys- intensity and frequency with the recent increase in atmos- ical infrastructure projects specifically related to sea- pheric temperature. Data on economic losses associated level rise, such as: with catastrophic events show that in constant 1999 · sea-surge protective barriers and dams; U.S. dollars, they have increased worldwide from an · reconstruction of harbor facilities; and average of US$71.1 billion in the 1960s to an average · flood barriers and Tsunami-prevention facilities (IPCC of US$608.5 billion in the 1990s (Freeman 2000). 2001a: 405). Although the economic impacts of these catastrophic Such expenditures represent a significant burden episodes are relatively evenly split between developed for the public sector, for private utility companies, and and developing countries, when related to respective indirectly for urban economies as a whole. GDP, the economic losses are significantly higher in Permanent changes to local ecosystems induced by developing than in developed countries (Sharma and climate change such as the salinization of ground-water others 2000). Average economic losses in developing and river estuaries might also alter the local economic countries due to climate change were tentatively esti- base. Adaptation to such changes may be difficult to mated at 2 to 9 percent of GDP by the IPCC in its Second achieve, but it is essential. An integrated adaptation Assessment Report, but variability in the underlying response might involve coastal zone protection, the cre- assumptions is considered high. ation of new breeding grounds for fish, the expansion of irrigation agriculture, the implementation of new public health measures, appropriate land-use planning, and building codes that internalize climate change con- Adaptation of Cities to Climate Change straints (Burton and Van Aalst 1999: 6). While private investment must become part of the strategy, the Throughout history, cities and human settlements public sector will have overall responsibility for the have adapted to climate variability, but the intensity and adaptation plans and managing such transitions. pace of the present and forthcoming climate changes In some cities in developing countries, relocation or induced by the continued and ongoing use of fossil fuels "managed retreats"(IPCC 2001a: 405) of resident pop- are already and will increasingly be a major challenge to ulations and economic activities to less vulnerable sec- many of them. Urban adaptation to climate change can tions of urban areas will have to take place over time. be defined as the sum of all physical and organizational This would require a mix of market incentives and public adjustments to urban life that will be required to cope sector planning and investments. Increased awareness with the profound and durable changes in weather ofvulnerabilitytoclimatechangecaninduceprivatefirms and climatic patterns. Determinants of adaptive capac- to relocate, especially when the differential costs for insur- ity include the availability of financial resources, tech- anceandre-insurance,whereapplicable,representapow- nology, specialized institutions and human resources, erful incentive to do so. However, this would not apply access to information, and the existence of legal, to the informal sector or to small businesses that do not social, and organizational arrangements (Burton and have the necessary resources. Replacement of ozone- Van Aalst 1999), all of which are assets that are typi- harming assets at the end of their productive lifecycles cally scarce in developing countries and cities. In cities maycreateopportunitiesforadaptationthroughupgrades Cities and Climate Change 95 or relocation. Land-use planning should channel new estimated at some 21 percent of the total, or 35 percent residential developments and industrial investments if the transition countries of the former Soviet Union toward less vulnerable areas. However, residents of poor are included. and informal settlements and slums--unless assisted-- Discussions surrounding mitigation have led to a would in all likelihood lack the tenure and resources to heated political debate between developing and indus- vacate the vulnerable areas in exchange for safer ones. trialized countries over bearing responsibility for past In the context of scarce financial and technical emissions and regulating current emissions. Develop- resources and of competing developmental priorities, ing countries insist that industrialized countries assume investments related to the adaptation of cities--their the burden of past emissions that led to global warm- physical infrastructure, local economy, basic urban serv- ing and that developing countries be allowed unre- ices, and residential settlements--to the impacts of cli- stricted access to energy generation and consumption, mate change will not be popular. They will only be which is considered critical for their future economic possible if there is an increased understanding of growth. They argue that costly mitigation measures urban vulnerability and the ability to demonstrate that are now required on account of the historical accu- the investments will be targeted at aspects of urban mulation of carbon emissions and the high level of development that require urgent attention. Informal set- fossil fuels currently consumed by industrialized tlements and critical infrastructure such as water supply countries that should henceforth pay the costs of and drainage systems may be already stretched to the mitigation. limit and would be first priorities. The reduction of GHG emissions has a globally ben- Management and institutional aspects of climate eficial effect wherever it occurs. The commitment of change adaptation can be just as challenging, if not more some industrialized countries to reduce their national so, than financial ones. These include: global emissions in the future (subject to ongoing nego- · generation of reliable and comprehensive assessments tiations at the international level) and the relatively lower of risk vulnerabilities for exposed cities and the dis- costs of GHG emissions reductions in developing coun- semination of such information; tries provide the rationale for international carbon emis- · establishment of early warning systems and evacu- sions trading. This represents an opportunity for ation plans, including emergency preparedness and "win-win" investments in cities in developing countries, neighborhood response systems; provided that the projects financed have sound local · improved efficiency of water supply management, and global objectives. by minimizing leakages and instituting marked-based In cities, the direct sources of global emissions include pricing mechanisms; energy generation, vehicle use, industrial and point- · improving health education and institutional capac- source use of fossil fuels, and burning of biomass. ity in urban environmental management; Indirect sources include electrical energy produced for · regularizing property rights for informal settle- public lighting, transportation, and industrial, com- ments and other measures to allow low-income groups mercial, and household consumption. Added together, to buy, rent, or build good-quality housing on safe these determine overall urban energy demand. Exam- sites (IPCC 2001a: 405­406). ples of interventions that combine significant local ben- efits with GHG emissions abatement that can generate revenues for developing cities are: Mitigation of Greenhouse Gas Emissions in Cities · Improved building materials and energy efficiency: Reduce energy requirements for heating, lighting, Cities in developing countries can contribute to the mit- and cooling and can increase efficiency in the use of igation of global greenhouse gas (GHG) emissions by building materials and the building cycle itself. reducing the volume of CO2, methane, and other gases · Transport demand management: Reduces the total that they release. Overall, however, the contribution volume of CO2 emissions of vehicles by promoting of developing countries to global emissions is low, greater ridership in mass transportation systems, 96 Building Safer Cities: The Future of Disaster Risk pedestrian zones, nonmotorized transportation, and of adaptation should be presented as "reparation costs" the use of more fuel-efficient vehicles and environ- for which industrialized countries must pay. This is not mentally friendly fuels. reflected in the letter or spirit of the UNFCCC agree- · Methane recuperation from landfills: Channels harm- ments, however, and many industrialized countries do ful methane gas emissions that would normally be not want to address this issue. released into the atmosphere into power generation-- Financial provisions for vulnerability and adaptation a viable investment with good economic returns. in the UNFCCC are currently limited to 2 percent of · Cleaner energy generation: Uses carbon sequestra- project investments in developing countries, resulting tion to reduce pollution by switching power plants from the trade of Carbon Emission Reduction (CER) from coal to natural gas, promoting the use of gas certificates. Such trade will start only when the Kyoto and clean energy sources to replace biomass, and Protocol is ratified and the Clean Development Mecha- cogenerating heat and electricity--results in signif- nism (CDM), conceived as the certification and clearing- icant local pollution abatement. house for the trading of emissions between developing and transitional economies on one side, and industri- alized countries on the other, becomes functional. Rat- Financing of Adaptation and Mitigation Projects ification of the protocol is expected in 2002 or 2003, in Developing Cities after which national regimes for emissions reductions will be adopted. The United Nations Framework Convention on Climate Several industrialized countries (Canada, Iceland, Change (UNFCCC) was established to forge interna- Norway, New Zealand, Australia, Switzerland, and mem- tional cooperation in the mitigation of further climate bers of the European Union) are considering the estab- change; therefore, its provisions and implementation lishment of a Special Climate Change Convention Fund mechanisms primarily address opportunities for GHG or Adaptation Fund to be administered by the GEF. emissions abatement. The Global Environment Facil- The fund would make limited financial resources ity (GEF) was set up as the financial mechanism of the available to developing countries for climate change UNFCCC to assist developing and transitional coun- adaptation purposes. The overall endowment of the tries with four global challenges: biodiversity conser- fund would be some 450 million annually, and it is vation, ozone depletion, international waters protection, expected to become effective in 2005.1 Least-developed and climate change. In the latter area, GEF has so far countries would be the primary beneficiaries as their exclusively focused on the implementation of projects GHG emissions (1 percent of the total) would not entitle with GHG emissions abatement or mitigation objectives. themtoanysignificantgainsfromtheCleanDevelopment During the 1990s, the climate change community Mechanism for emissions trading, while their adaptation paid relatively less attention to adaptation, while on needs are often the largest. The World Bank's Environ- the other hand, awareness of and preparedness for nat- ment Strategy also refers to the creation of a Variability ural disasters have significantly increased. This may be and Adaptation Facility (VAP), to be financed by donor changing rapidly, however, as the impacts of climate governments in the near future (World Bank 2001). change are beginning to be felt more strongly. Devel- In the case of mitigation, on the other hand, UNFCCC oping countries are now increasingly asking for proj- financial resources and mechanisms that can be accessed ect and financial assistance from industrialized countries by the public and the private sector to contribute to for adaptation projects and their related costs. the abatement of GHG emissions are already in place. Providing funding to developing countries is contro- Targeted interventions include land-use planning, urban versial, however. Many of the problems relating to global transport, energy generation and efficiency, and urban warming were caused by industrial countries' sustained environmental management. use of fossil fuels over the past 150 years, during which The UNFCCC has established Activities Implemented accumulated carbon emissions became trapped in the Jointly (AIJ), also known as the Joint Implementation atmosphere. Consequently, it can be argued that the costs (JI), as the formal mechanism for collaboration between Cities and Climate Change 97 industrialized and developing countries to implement incorporating climate change objectives--mitigation or projects that have GHG-abatement objectives. Under adaptation--to be supported by concessional financ- AIJ, an industrialized country partners with a develop- ing at no cost to the client. ing country to provide financial and technical assistance for a project.2 GEF Operational Programs (OP) in the climate change Implications for The World Bank's Infrastructure area are all highly relevant to cities. These programs Projects in Urban Areas co-finance development projects that combine signifi- cant local objectives with global ones for which GEF A number of key strategies call for the integration of provides grants equivalent to the "incremental costs." climate change in The World Bank Group's work: the These are: 1999 Energy and Environment Strategy, "Fuel for · OP 5: Removal of Barriers to Energy Efficiency and Thought;" the 2001 Environment Strategy, "Making Sus- Energy Conservation; tainable Commitments;" and the 2002 Urban Transport · OP 6: Promoting the Adoption of Renewable Energy Strategy, "Cities on the Move." Climate change is also by Removing Barriers and Reducing Implementation classified as a corporate public good priority and forms Costs; part of the Millennium Development Goal4 of ensuring · OP 7: Reducing the Long-Term Costs of Low Green- environmental sustainability. While the Bank is active house Gas Emitting Energy Technologies; and through its lending operations in the mitigation of future · OP 11: Promoting Environmentally Sustainable climate change, however, vulnerability and adaptation Transport.3 objectives are rarely found as part of the developmen- The World Bank Group-GEF portfolio consists of tal objectives of its projects or as project risks that need 62 projects, for which $6.2 billion has been mobilized-- to be mitigated through given interventions. $730 million from the GEF and the remainder from The The most comprehensive review to date of the Bank's World Bank Group, donors, private investors, and performance in this area is to be found in "Come Hell or governments (World Bank 2001). High Water--Integrating Climate Change Vulnerability Finally, in the wake of the Kyoto Protocol ratifica- and Adaptation into Bank Work," by Ian Burton and tion and the effectiveness of the CDM, the Prototype Maarten van Aalst, Environment Department Paper No. Carbon Fund (PCF) was established by The World Bank. 72, 1999. The review finds the level of integration modest The PCF is a pilot project to test carbon emissions and provides numerous examples of projects that have trading, channel financial resources and information, failed to internalize obvious climate change risks, to the and build capacity in developing countries on GHG detrimentoftheBank'sclientsandtheinstitution'sintegrity. emissions abatement issues. PCF emphasizes the devel- Among the examples provided in the study, an emer- opment of renewable energy projects such as wind, small gency loan for potable water service recovery and restora- hydropower, solar direct and photovoltaic, landfill gas, tion and flood protection for the city of Georgetown, and refuse-derived fuel. Additionally, energy-efficiency Guyana, after the 1997­98 El Niño episode, was projects involving buildings and appliances are also reviewed. Despite the fact that a significant portion of financed under the PCF, which has a significant share the city is below sea level and the emergency loan of such projects in its portfolio (PCF 2001). The capi- addressed damage caused by a climatic episode, the tal of the PCF is currently $145 million. intensity of which was enhanced by global warming, AnanalysisoftheWorldBank/GEFportfoliointhearea no new assessment of the risks was made in project of climate change provides an encouraging example of design. This is often the case in hurried emergency oper- the proactive work in which the Bank could further ations, making the new investment equally vulnerable engage itself should the Adaptation Funds mentioned to further climate-induced damage. earlier become available. The Bank could focus on assist- In the absence of concessional financing for adapta- ing its client cities by designing programs and projects tion, it is understood that Bank clients would be unwill- aimed at relevant local developmental priorities and ing to borrow for climate change, but the authors 98 Building Safer Cities: The Future of Disaster Risk argue that the Bank--in addition to promoting more About 30 percent of the Bank's overall investments, or awareness of climate change concerns--should at the $6­7 billion, take place in urban areas. Only 5 percent very least (Burton and Van Aalst 1999: 35): of this is handled directly by urban sector units. Urban · Address climate change concerns in the Country projects, in addition to internalizing climate change mit- Assistance Strategies (CAS): Currently CASs do not igation and adaptation concerns in the design of their generally cover such risks. Especially for those coun- physical components, could also provide opportunities tries that are known to be particularly vulnerable, to ensure that such integration occurs as well in all urban the CAS could highlight activities to be avoided or projects, regardless of sector. discouraged as well as new activities designed to take Some short-term measures to begin to address these advantage of opportunities opened up by climate issues could include: change if milder climates or increased precipitation · Collection and validation of information on specific affect previously cold or dry regions (Burton and Van vulnerability to climate change in urban areas of coun- Aalst 1999: 13). tries at risk · Minimize the vulnerability of the Bank's portfolio · Review of a sample of active projects in urban areas to climate change: Based on examination of the Bank's of at-risk countries to identify opportunities for min- portfolio in countries with known vulnerability,5 an imizing risks and including "no regrets" measures average of 37 percent of projects are at some degree · Dissemination of practical information on mitigation of risk due to their location, design, or for failing to measures and financial mechanisms, in an urban con- have taken into account climate change impacts on text, to urban and infrastructure staff. water flows, sea-level rise, and related issues. As climate change dramatically continues to manifest · Maximize the role of the Bank's portfolio on vul- itself with loss of life and property and affect profound nerability and adaptation: As part of the same port- changes to livelihoods in cities of the developing world, folio review, the authors found that 43 percent of The World Bank has the obligation to assess its impor- the projects could help reduce vulnerabilities to cli- tance and assist its client cities in addressing these spe- mate change by incorporating adaptive considera- cial challenges. While its agenda will still be driven by tions into their designs such as the heights of themissionofreducingpovertyandtheMillenniumDevel- bridges and embankments, the expected frequency opmentGoals,theincreasingprominenceofclimatechange of coastal storms, or other "no regrets" or noncon- mustbetakenfullyintoaccountandinternalizedtodesign troversial measures (Burton and Van Aalst 1999: 18). and implement development projects that target the spe- These recommendations are particularly relevant cial needs of populated and growing urban areas. for the Infrastructure (INF) Vice-Presidency of The World Bank, home to three sectors responsible for the major- Notes ity of the Bank's urban infrastructure investments: energy 1. Personal communication with Alan Miller, Senior Environ- and mining; transportation; and water, sanitation, and mental Specialist, GEF, January 28, 2002. flood protection. They are also relevant for urban devel- 2. See http://unfccc.int/program/coop/aij/index.html. opment, an INF theme that carries the overall respon- 3. See . strategies, improving urban livability and reducing urban 4. The eight Millennium Development Goals are: eradicate extreme poverty, through physical investments and policy poverty and hunger; achieve universal primary education; pro- guidance. The Disaster Management Facility (DMF) and mote gender equality and empower women; reduce child mor- the Cities Alliance are working on closely related issues: tality; improve maternal health; combat HIV/AIDS, malaria, and the first, as it addresses concerns of disaster awareness, other diseases; ensure environmental sustainability, and develop preparedness and reconstruction; the second, as it focuses a global partnership for development. on the elimination of urban slums and informal settle- 5. Bangladesh, Ecuador, Guyana, Papua New Guinea, Samoa, ments built in vulnerable locations. and Zimbabwe. Cities and Climate Change 99 Bibliography Working Group I to the Third Assessment Report of IPCC. Cam- Burton, I. and M. Van Aalst. 1999. "Come Hell and High Water-- bridge: Cambridge University Press. Integrating Climate Change Vulnerability and Adaptation into Mimura, N. and H. Harasawa, eds. 2001. Data-book of Sea-level Bank Work." ESSD Climate Change Series. Washington, D.C.: Rise 2000. Tsukuba: Center for Global Environmental Research, The World Bank. NIES as quoted in IPCC Third Assessment Report, Vol. 2:396. El-Raey, Mohamed. 1997. "Vulnerability Assessment of the Coastal Munich Re, Press Release March 15, 1999:1. Zone of the Nile Delta in Egypt to the Impacts of Sea-level Munich Re. 2000. "World of Natural Disasters." CD-ROM, Rise." Ocean and Coastal Management 37(1):29­40 as quoted Geospecials/Megacities. in IPCC Third Assessment Report Vol. 2, chapter 7. Nicholls, R. and others. 1999. "Increasing Flood Risk and Wet- FAO (Food and Agricultural Organization of the United Nations). land Losses due to Global Sea-level Rise: Regional and Global 1999. The State of Food Insecurity in the World as quoted in Analyses." Global Environmental Change 9:S69­S87 as quoted IPCC Third Assessment Report, Vol. 2:409. in IPCC Third Assessment Report Vol. 2, chapter 7. Freeman, K. 2000. "Infrastructure, Natural Disasters and Poverty." PCF (Prototype Carbon Fund). 2001. Annual Report 2001, In Alcira Kreimer and M. Arnold, eds., Managing Disaster Risk Washington, D.C.: The World Bank: 17. in Emerging Economies. Washington, D.C.: The World Bank. Sharma, M. and others. 2000. "Reducing Vulnerability to Envi- Hardoy, Jorge and others. 2001. Environmental Problems in an ronmental Variability." Environment Strategy Background Paper. Urbanizing World. London: Earthscan: 207. Washington, D.C.: The World Bank: 5. IPCC (Intergovernmental Panel on Climate Change). 2001a. Turner, R. and others. 1990. "Cities at Risk." London: BNA Inter- Climate Change 2001: Impacts, Adaptation and Vulnerability-- national as quoted in Hardoy and others, Environmental Prob- Contribution of Working Group II to the Third Assessment Report lems in an Urbanizing World. London: Earthscan. of IPCC. Cambridge: Cambridge University Press: 383­416. WorldBank.EnvironmentalandSociallySustainableDevelopment. IPCC (Intergovernmental Panel on Climate Change). 2001b. 2001. "Climate Change." Environment Strategy. Washington, Climate Change 2001: The Scientific Basis--Contribution of D.C.: World Bank, 175. Chapter 8 The Resilience of Coastal Megacities to Weather-Related Hazards Richard J. T. Klein, Robert J. Nicholls, and Frank Thomalla The 20th century has been characterized by a funda- According to Costanza and others (1995), coastal mental change to humanity: we have moved from being ecosystems are highly resilient because of the diversity a predominantly rural to a predominantly urban species of their functions and the linkages between these func- (UN/DESA 2001; UNEP 2002). Further, nearly all future tions. In the same manner, it is argued that coastal population growth is expected to occur in urban areas economies are more diverse and have multiple niches as this process of urbanization continues through the and that they are inherently more resilient because of 21st century. Part of this process has been the emer- the resilience of the ecosystems on which they depend gence of large cities--larger cities than have ever existed (Adger 1997). in the history of humanity--which are often termed This paper explores the concept of resilience in the megacities (defined in this paper as cities with a popu- context of coastal megacities, and particularly explores lation exceeding 8 million inhabitants). Most of these its value and utility in the context of hazard manage- megacities are found in coastal locations (Nicholls 1995), ment and reduction. First the development of coastal and their development is raising a range of fundamental zones and megacities is explored, in order to define the management questions (Timmerman and White 1997). context of the question. This includes defining an explicit History shows us that some of the world's large set of "coastal megacities" by drawing on data of UN/DESA cities have a long history of continuous occupancy and (2001). Then weather-related hazards in coastal zones importance, even though they have adjusted to chang- and relevant hazard-reduction strategies are considered. ing circumstances. Cairo, Istanbul (Constantinople), This analysis focuses on the hazards that are specific to and Baghdad began the second millennium as they coastal locations, with particular emphasis on megac- ended it--among the world's largest cities. Other major ity and other large city locations. It also includes an cities in 1000 A.D. are now of relatively minor impor- analysis of sea-level rise and climate change, as these tance (e.g., Kaifeng, China; Nishapur, Persia; Córdoba, are beginning to modify weather-induced hazards in Spain) or have even been abandoned (Angkor, Khmer coastal areas, and they are encouraging new thinking Empire) (Harrison and Pearce 2000), although all of on hazard-mitigation strategies. Note that in a parallel these examples of abandonment are noncoastal cities. paper, Bigio (2002) provides an overview of the chal- Weather-induced hazards are one potential reason for lenges that climate change poses to cities. The academic such decline. While the new coastal megacities can com- debate over the last 30 years on the meaning of resilience mand large resources and significant political influence, in the context of hazard mitigation and, more gener- they are also concentrating increasing populations in ally, environmental management, is then considered. potentially hazardous locations, particularly in the Last, all of this information is synthesized using three new emerging megacities in the developing world. Thus, questions that are core to this paper: the development of these cities requires successful · Is resilience a desirable attribute of megacities? mitigation of these hazards. · Does enhanced resilience reduce the vulnerability System resilience is seen as a desirable property in of megacities to weather-related hazards? the face of a range of potential stresses, including weather- · Is resilience a useful concept in the hazard manage- induced hazards anddisasterreduction(UN/ISDR2002a). ment of megacities? 101 102 Building Safer Cities: The Future of Disaster Risk Coastal Zones and Megacities be concentrated in coastal zones, as has been the case in most industrialized countries. It is projected that by Although there is a long history of human settlement 2015 there will be 33 cities with a population of more in coastal zones, until the 20th century, the level of dis- than 8 million (UN/DESA 2001). As shown in table 8.1, turbance to natural processes did not appear to be crit- 21 of these megacities are located in coastal zones. ical. During the 20th century, coastal populations grew This table, which is an update of the overviews pro- rapidly around the globe because of the many economic vided by WCC'93 (1994) and Nicholls (1995), also opportunities and environmental amenities that coastal shows that 17 of the 21 largest megacities are coastal zones provide (Turner and others 1996). Low-lying areas and that, with the exception of Tokyo, New York, Los near coasts now have the largest concentrations of people Angeles, Osaka, Paris, and Moscow, all the projected on earth (Small and Cohen 1999). The population in megacities are situated in developing countries. Con- the "near-coastal zone" (defined as areas both within tinued growth of urban areas can be expected after 2015, 100 meters elevation and 100 kilometers distance of especially in Africa and Asia (UNEP 2002), resulting in the coast) in 1990 was estimated at 1.2 billion (thou- the development of additional coastal megacities to those sand million), or 23 percent of the world's population shown in table 8.1. (Nicholls and Small 2002). Some care should be taken in interpreting the data Nicholls and Small (2002) also showed that most of presented in table 8.1. A certain degree of subjectivity the near-coastal zone is sparsely inhabited, with the is inevitable in labeling a megacity as "coastal," especially human population being concentrated in a few spe- because there are no straightforward definitions of a cific areas of the world's coast. These areas correspond coastal zone. São Paolo, for example, is not considered mainly to near-coastal plains in Europe and parts of coastal because it is situated at an elevation of 800 meters Asia, and to a lesser extent to densely populated urban above sea level. However, its proximity to the Atlantic areas. Hence, there are wide variations in coastal pop- Ocean and the port of Santos has yielded benefits that ulations among nations. In many small island nations, would not have been available in places further inland all land suitable for human habitation is coastal and also (e.g., as a coffee trading capital). On the other hand, in some large countries, most or all major urban cen- cities like Dhaka, Calcutta, and Cairo are situated at some ters are located near the coast (e.g., Australia). In other distance from the sea, yet in table 8.1 they are consid- countries, such as Mexico, Colombia, Russia, and Iran, ered coastal because of their deltaic setting. Other many larger cities are found further inland, in spite of cities, such as Los Angeles, Seoul, and Istanbul, have the countries' long coastlines. developed on coasts with steeper gradients and hence, The United Nations medium projection for popula- parts of their agglomerations extend outside the coastal tion growth suggests that the world's population will plain. The predominant criterion to classify a city as reach 7.2 billion by the year 2015, 7.9 billion by 2025, coastal in table 8.1 is whether the city has economic and 9.3 billion by 2050 (2000: 6.1 billion; UN/DESA and geomorphic characteristics that are typically or exclu- 2001). Growth rates in individual countries will be sively coastal (e.g., seaport, deltaic, or estuarine setting). largely determined by the country's current demographic Some coastal agglomerations with populations exceed- patterns and fertility rate. Age structures in most devel- ing 8 million people may have been omitted because oping countries are such that during the coming decades of the dataset utilized in table 8.1. These include Greater greater numbers of people will come into their prime London in the United Kingdom and the Hong Kong- reproductive years than in industrialized countries. Fur- Shenzhen-Guangzhou conurbation in China (Nicholls ther, fertility rates are generally higher in developing 1995). More disperse agglomerations are not consid- countries, albeit declining. As a result, all projected pop- ered either, such as the Amsterdam-Brussels axis in the ulation growth until 2050 is expected to occur in the Netherlands and Belgium, the Osaka-Nagoya-Tokyo developing world (UN/DESA 2001). axis in Japan, and "Megalopolis" in the United States, Most of the population growth in developing coun- which stretches over 600 kilometers from Boston to tries will occur in urban settings and much of this will Washington, D.C., and has a collective population The Resilience of Coastal Megacities to Weather-Related Hazards 103 Table 8.1 The world's largest cities, with projected populations in 2015 exceeding 8 million Expected Population size (million) growth (%) Rank in Rank in Rank Agglomeration Country 1975 2000 2015 2000­2015 1975 2000 1 Tokyo Japan 19.771 26.444 26.444 0.00 1 1 2 Bombay India 6.856 18.066 26.138 44.68 15 3 3 Lagos Nigeria 3.300 13.427 23.173 72.59 33 6 4 Dhaka Bangladesh 2.172 12.317 21.119 71.46 43 11 5 São Paolo Brazil 10.047 17.755 20.397 14.88 5 4 6 Karachi Pakistan 3.983 11.794 19.211 62.89 16 12 7 Mexico City Mexico 11.236 18.131 19.180 5.79 4 2 8 New York United States 15.880 16.640 17.432 4.76 2 5 9 Jakarta Indonesia 4.814 11.018 17.256 56.62 22 14 10 Calcutta India 7.888 12.918 17.252 33.55 12 8 11 Delhi India 4.426 11.695 16.808 43.72 23 13 12 Metro Manila Philippines 5.000 10.870 14.825 36.38 21 16 13 Shanghai China 11.443 12.887 14.575 13.10 3 9 14 Los Angeles United States 8.926 13.140 14.080 7.15 8 7 15 Buenos Aires Argentina 9.144 12.560 14.076 12.07 7 10 16 Cairo Egypt 6.079 10.552 13.751 30.32 20 19 17 Istanbul Turkey 3.601 9.451 12.492 32.18 32 22 18 Beijing China 8.545 10.839 12.299 13.47 10 17 19 Rio de Janeiro Brazil 7.854 10.582 11.905 12.50 13 18 20 Osaka Japan 9.844 11.013 11.013 0.00 6 15 21 Tianjin China 6.160 9.156 10.713 17.01 19 24 22 Hyderabad India 2.086 6.842 10.457 52.84 45 31 23 Bangkok Thailand 3.842 7.281 10.143 39.31 28 27 24 Lahore Pakistan 2.399 6.040 9.961 64.92 41 35 25 Seoul Republic of Korea 6.808 9.888 9.923 0.35 16 20 26 Paris France 8.885 9.624 9.677 0.55 9 21 27 Lima Peru 3.651 7.443 9.388 26.13 30 25 28 Kinshasa Dem. Rep. of the Congo 1.735 5.064 9.366 84.95 n/a 40 29 Moscow Russian Federation 7.623 9.321 9.353 0.34 14 23 30 Madras India 3.609 6.648 9.145 37.56 31 32 31 Chongqing China 2.439 5.312 8.949 68.47 40 38 32 Teheran Islamic Republic of Iran 4.274 7.225 8.709 20.54 25 28 33 Bogotá Colombia 3.036 6.288 8.006 27.32 35 34 Total (coastal agglomerations) 150.6 254.1 324.1 27.53 Total (all agglomerations) 217.4 368.2 467.2 26.88 n/a: not available. Note: Arrows indicate coastal agglomerations. [Arrows TK: ck. with author] Source: Population data from UN/DESA (2001). approaching 50 million people. Such disperse coastal of fisheries potential because of habitat loss and pollu- agglomerations may also emerge in the developing world, tion of rivers and coastal waters from urban and indus- such as from Accra, Ghana to Lagos, Nigeria, embrac- trial waste. ing parts of four countries. In addition, the large populations in many coastal Urban populations tend to have higher consumption areas around the world are, to a greater or lesser levels than their rural counterparts, as well as different extent, vulnerable to hazardous events associated with consumption patterns. The increasing demand for food natural coastal dynamics such as storm surges, floods, requires increasing productivity in fisheries and agri- and tsunamis. Human-induced climate change and sea- culture. However, this is often impeded by the loss of level rise will further increase this vulnerability. These agricultural land to urban expansion and the reduction hazards are discussed in the next section. 104 Building Safer Cities: The Future of Disaster Risk Weather-Related Hazards in Coastal Zones Coastal storms can cause significant damage at the coast, including beach erosion (i.e., the physical removal Coastal zones and any cities so located are threatened of sediment). These impacts are normally confined to by a wide variety of weather-related hazards, and human quite narrow areas adjacent to the coast, but erosion activities often modify these hazards. Some of these haz- can trigger catastrophic flood events if natural barriers ards, such as drought, river floods, and poor air qual- such as dunes are breached. Long-term erosion, which ity enhanced by stagnant air masses and inversion, can is occurring widely around the world's coasts, can cause affect the entire terrestrial landscape. Other weather- land loss and will often degrade beach resources that related hazards are more specific to coastal locations are used for tourism (Bird 1985). The resulting changes and are emphasized here. These hazards are often related in coastal configuration also change the risk of flood- and include: ing and storm damage. Erosion has been an issue in · Erosion most coastal cities on the open coast (15 out of 21 megac- · Storm and wind damage ities in table 8.2). Beach nourishment is becoming a · Flooding widespread response to this problem (Davison and others · Salinization of surface waters. 1992; Hamm and others 1998). Table 8.2 summarizes the occurrence of these haz- Flooding of low-lying coastal areas can occur for a ards for the coastal megacities identified in table 8.1, range of sometimes interacting reasons, including storm as well as the occurrence of major human-induced surges induced by tropical and extra-tropical storm sys- subsidence within the city during the 20th century. tems, high river flows, and, generally less importantly, Table 8.2 Summary of the major weather-related hazards and the occurrence of subsidence during the 20th century for coastal megacities as forecast in 2015 Storm and wind damage Flooding Agglomeration Erosion Hurricane landfall1 Extra-tropical storms River Surge Salinization Major subsidence Tokyo Y Y (3) -- Y Y ? Y Bombay Y Y (<1) -- -- Y ? -- ? Lagos Y -- -- -- Y ? ? Dhaka -- Y (<1) -- Y -- Y Y ? Karachi Y Y (<0.1) -- -- Y ? -- ? New York Y Y (<1) Y -- Y ? -- Jakarta Y -- -- Y -- ? Y Calcutta -- Y (<1) -- Y -- Y Y ? Metro Manila Y Y (>3) -- -- Y ? Y Shanghai -- Y (1) -- Y Y Y Y Los Angeles Y -- Y -- Y -- Y2 Buenos Aires Y -- Y Y Y ? -- ? Cairo -- -- Y Y -- -- -- ? Istanbul Y -- Y -- Y ? -- Rio de Janeiro Y -- -- -- Y -- -- Osaka Y Y (3) -- Y Y ? Y Tianjin -- Y (<0.1) Y Y Y Y Y Bangkok Y Y (<1) -- Y Y ? Y Seoul -- Y (1­3) Y Y -- -- -- ? Lima Y -- -- ? Y -- -- Madras Y Y (<1) -- -- Y ? ? Y: yes, --: no, ?: uncertain. 1 The relative frequency of hurricane landfall is indicated by the annual occurrence of tropical storms and cyclones (Beaufort force 8 and above). 2 Due to oil and gas extraction, rather than groundwater withdrawal. Sources: Synthesized from Nicholls (1995) and expert judgment. The Resilience of Coastal Megacities to Weather-Related Hazards 105 intense local precipitation in areas with poor drainage. cities often triggers accelerated groundwater exploita- Hurricanes, tropical cyclones, and typhoons (which are tion. As the water table beneath the city drops, formerly different names for the same meteorological phenom- saturated sediments can irreversibly consolidate, the ena) have produced some of the most dramatic coastal bulk density rises, and the ground surface rapidly sub- floods over the last few years. These include the cyclones sides (Nicholls 1995). In low-lying coastal areas this of April 1991 in coastal Bangladesh (139,000 fatali- leads to a relative rise in sea level as the land sinks. As ties) or October 1999 in Orissa, India (10,000 people a result, the erosion, flood, and salinization hazards killed and 15 million people made homeless) (UNEP already described will all be exacerbated. In addition, 2002). These coastal storm events comprise more than rapid differential subsidence can cause damage and storm-surge flooding, as they are associated with intense destruction of infrastructure such as pipelines and build- precipitation and strong winds. Hurricane Mitch, for ings. In the extreme, complete submergence of low- example, caused great damage on the coasts of Hon- lying coastal areas could be the result of subsidence. duras and Nicaragua in 1998, but most of the 17,000 At least 8 to 10 of the coastal megacities in table 8.2 fatalities occurred inland of the coastal zone due to flash subsided during the 20th century with adverse eco- river floods and landslides. Winds rather than surges nomic consequences. Subsidence rates can be rapid, were the major cause of property damage during the locally reaching one meter per decade (Nicholls 1995). landfall of Hurricane Andrew near Miami, Florida in For instance, land subsidence around Tianjin, China 1992. These examples emphasize the need to consider was up to 5 centimeters per year in the late 1980s and all weather-related hazards and their interrelationships. locally up to 11 centimeters per year in some periods. Variations in river flow and sea level also lead to On the other hand, in Japan subsidence has been largely movement of the saltwater interface in surface waters. In stopped with better groundwater management, although the strongly seasonal rivers of South, Southeast, and East large areas in the Osaka-Tokyo conurbation, which are Asia, saltwater penetrates far upstream in the dry season home to 2 million people, are now beneath high-water with even greater penetration during droughts. In cities levels and would be submerged but for the extensive where the river is the main water supply, salinization can flood defense systems (Mimura and others 1992). The endanger water supplies, although as shown in table 8.2, success of better groundwater management in Osaka, only limited information is available on this factor. Tokyo, and Shanghai has not prevented similar prob- It should be noted that the aforementioned weather- lems developing in other "younger" megacities (e.g., related hazards are usually directly modified by other Bangkok, Metro Manila, and Jakarta), and it is likely human activities in and around urban areas, including: that during the 21st century other coastal cities will · Changing sediment supply due to changing land use, experience similar problems (e.g., Rangoon, Myanmar hydrologicalmodificationorcoastalprotectionandthe and Hanoi, Vietnam). Given that the physical causes consequent influence on erosion and deposition (e.g., and socioeconomic triggers of human-induced subsi- rapid land loss in the Mississippi delta is increasing dence are now well understood, such problems are thefloodriskinNewOrleans;Boeschandothers1994); avoidable with appropriate planning and management · Land claim of intertidal areas and deepening of chan- (Nicholls 1995). nels for navigation, which often increase extreme Changing land use also exacerbates human exposure water levels and hence flood risk (e.g., London; Kelly to weather-related hazards. Initial settlements are often 1991); and builtonthemostsuitableareasandasacityexpands,new · Enhanced subsidence due to groundwater withdrawal, development has to occur on more hazardous ground. which has reduced land elevation in many large coastal Good examples are the expansion of Alexandria, Egypt cities, particularly those in deltaic settings in Asia and Dhaka, Bangladesh onto low-lying flood-prone areas. (Nicholls 1995). Dakar, Senegal evolved on the relatively high and non- Land subsidence due to excessive groundwater with- erosive land on the Cap Vert peninsula but recent growth drawal is a particular problem in many coastal cities has taken place on lower-lying land. Erosion is now built on geologically recent deposits. Rapid growth of becoming an issue and this newly developed land may 106 Building Safer Cities: The Future of Disaster Risk be flood-prone (Dennis and others 1995). Often the that are twice and four times preindustrial levels, respec- poorest people in cities are found in unauthorized shan- tively (Church and others 2001). tytown areas (UN/ISDR 2002b), including Bombay, A major uncertainty is how the global mean sea-level Lima, Calcutta, Manila, Karachi, Lagos, Bangkok, Jakarta, rise will manifest itself on a regional scale. All models and Rio de Janeiro (Devine 1992). These shantytowns analyzed by the IPCC show a strongly nonuniform tend to be built in the least desirable or secure areas, pattern of sea-level rise. Some regions show a sea-level which are often also those areas most exposed to weather- rise substantially more than the global average (in many related hazards. For example, in Recife, Brazil the illegal cases of more than twice the average) and others a sea- shantytown areas (or favelas) are expanding in man- level fall (Church and others 2001).1 However, the grove areas, which are by definition at or below patterns given by the different models are not similar normal high water (Muehe and Neves 1995). in detail. This lack of similarity means that confidence It is important to note that weather-related hazards in projections of regional sea-level changes is low, vary with climate variability, such as the El Niño although it is clear that they are important with respect Southern Oscillation (ENSO) and the North Atlantic to impacts of sea-level rise on coastal zones. For indi- Oscillation (NAO), and will be modified by climate vidual cities, rates of subsidence need to be consid- change. Both climate variability and climate change need ered, as subsidence will add to global mean sea-level to be taken into account when planning hazard responses rise and in some cases these changes could exceed the (see next section). For example, on the eastern seaboard changes due to climate change (see above). of the United States, significant coastal development A rise in mean sea level will lead to a decrease in the and urbanization occurred from 1966 to 1989, increas- return period of storm surges; however, it is unclear ing exposure during a period when hurricane activity whether the variability of storm surges itself will change. was well below average. The new inhabitants were often Changes that are considered possible over some areas ignorant of the hurricane risk, which became manifest are increases in peak wind intensities and mean and with the more frequent and stronger hurricanes that peak precipitation intensities of tropical cyclones. began with Hugo in 1989 (Pielke and Landsea 1998). Changes in tropical cyclone location and frequency are Climate change will be of increasing concern to coastal still uncertain, but it is clear that any changes can have megacity populations, as existing experience of haz- important consequences for many coastal cities, includ- ardous events will cease to be a guide to future events. ing 16 coastal megacities in table 8.2. Based on the most recent projections of future climate, Changes in wind patterns and precipitation in higher the Intergovernmental Panel on Climate Change (IPCC) latitudes also remain uncertain at local and regional in its Third Assessment Report projects global mean scales. To date, no reliable projections of coastal storm sea level to rise by 9 to 88 centimeters between 1990 characteristics and tidal range at scales useful for coastal and 2100, with a central value of 48 centimeters. The impact analysis exist. The effect of storm surges can also projected sea-level rise is due primarily to the thermal be compounded by increased precipitation on land and expansion of ocean water (11 to 43 centimeters), fol- associated increased river runoff. Runoff projections of lowed by contributions from mountain glaciers (1 to 23 selected rivers are becoming available, allowing their centimeters) and ice caps (­2 to 9 centimeters for Green- consideration in future impact analysis. land and ­17 to 2 centimeters for Antarctica) (Church The IPCC Third Assessment Report concludes that and others 2001). The projected central value of 48 cen- recent trends for El Niño are projected to continue in timeters gives an average rate of sea-level rise of 2.2 to many models (Houghton and others 2001). These trends 4.4 times the observed rate over the 20th century. include a stronger warming of the eastern tropical Pacific Even with drastic reductions in greenhouse-gas emis- relative to the western tropical Pacific, with a corre- sions, sea-level rise will continue for centuries beyond sponding eastward shift of precipitation. Changes in El 2100 because of the long response time of the global Niño patterns are likely to be relevant in a number of ocean system. An ultimate sea-level rise of 2 to 4 meters coastal locations, but they are still too uncertain to be ispossibleforatmosphericcarbondioxideconcentrations considered in coastal impact studies. The Resilience of Coastal Megacities to Weather-Related Hazards 107 Owing to the great diversity and variation of natu- rise and climate change. While it is recognized that ral coastal systems and to the local and regional differ- impacts of climate change on coastal megacities are ences in relative sea-level rise, the occurrence and prompted by more than sea-level rise alone, consider- response to sea-level rise will not be uniform around ation of other factors is constrained by the uncertain- the globe. Coastal environments that are especially at ties surrounding them, particularly at local and regional risk include deltaic and low-lying coastal plains, sandy scales. Nonetheless, a range of strategies to reduce coastal beaches and barrier islands, coastal wetlands, estuaries hazards due to all causes can be implemented, also in and lagoons, and coral reefs and atolls (Bijlsma and the face of uncertainty, as discussed in the next section. others 1996); cities in these locations are similarly threat- ened. The common association of megacities with deltas in Asia suggests that these cities are particularly threat- Hazard Reduction Strategies ened. Regional and global analyses suggest that, for the regions containing coastal megacities, increased There is a wide range of different strategies for dealing coastal flooding will be most severe in South and South- with the weather-induced hazards faced by coastal cities east Asia, Africa, the southern Mediterranean coasts, as defined in table 8.4. These can be applied from the and to a lesser extent, East Asia (Nicholls 2000), which level of the individual up to the entire city. Given the includes 16 of the coastal megacities in table 8.1. focus of the paper on the resilience of coastal megaci- However, these general conclusions need to be sup- ties, the focus here is on the city-scale response. ported by more detailed city-scale assessments, such In table 8.4, choosing change means accepting the as the recent New York assessment (Rosenzweig and hazard and changing land use, or even the relocation Solecki 2001). of exposed populations. Reducing losses includes trying Table 8.3 lists the most important socioeconomic sec- to reduce the occurrence of the hazardous event, or, tors in coastal zones and indicates from which of the more commonly, reducing the impacts of a hazardous aforementioned weather-related hazards they are expected event when it occurs. Accepting losses includes bear- to suffer direct impacts. Indirect impacts, for example, ing the loss, possibly by exploiting reserves, or sharing impacts on human health resulting from deteriorating the loss through mechanisms such as insurance. water quality, are also likely to be important to many Hence the ability to recover from the disaster is of most sectors but these are not shown in table 8.3. importance if losses are accepted. Note that these strate- In summary, coastal megacities are threatened by a gies are not mutually exclusive, and hazard reduction range of specific weather-related hazards, and during the within any coastal city might include elements of all 21st century these will likely be exacerbated by sea-level three approaches. Over time, technology is increasing the options that are available for hazard reduction, particularly those Table 8.3 Qualitative overview of direct socioeconomic strategies that reduce losses (Klein and others 2000). impacts of weather-related hazards and climate change Given the large populations and economic values in on a number of sectors in coastal zones Storm Socioeconomic and wind Table 8.4 Generic approaches to hazard reduction sector Erosion damage Flooding Salinization based on purposeful adjustment Water resources -- -- Y Y Purposeful adjustment Option Agriculture -- Y Y Y Human health -- Y Y -- Choose change Change location Fisheries Y Y Y Y Change use Tourism Y Y Y -- Reduce losses Prevent effects Human Y Y Y Y Modify event settlements Accept losses Share loss Bear loss Y: yes, --: no. Source: adapted from Klein and Nicholls 1999. Source: Burton and others 1993. 108 Building Safer Cities: The Future of Disaster Risk cities, there is usually a bias toward loss reduction, and to reduce losses (e.g., flood defense) only reduces rather it can be argued that large coastal cities would not than removes the risk, the measures are often seen as have evolved without the availability of these hazard- invulnerable and hence encourage further development reduction strategies. These might include warning sys- in what remain potentially hazardous areas (e.g., Parker tems, defense works, and resistant infrastructure. This 2000). Boxes 8.1 and 8.2 discuss examples of hazard- approach is most developed in coastal cities around the reduction strategies employed over the 20th century for North Sea, Europe, and in Japan, where flooding claimed London and Shanghai. many lives up to the middle of the 20th century. Most Article 3.3 of the United Nations Framework Con- of the coastal megacities identified in table 8.1 have seen vention on Climate Change (UNFCCC) suggests that significant action to reduce the effects of weather-related proactive adaptation (as well as mitigation to reduce hazards. A particular problem is that, while any strategy greenhouse gas emissions) deserves particular attention Box 8.1 Responding to coastal flooding in London, United Kingdom One million Londoners are potentially exposed to coastal flooding. They are defended by a complex system involving fixed flood defenses, the mobile Thames Barrier, which is closed before a surge arrives, and a suite of warning systems that are used to decide when to close the barrier (Gilbert and Horner 1984). While the possibility of a barrier was dis- cussed earlier in the 20th century, the decision to build the present defenses was made in direct response to the 1953 storm surge that killed 300 people and flooded 65,000 hectares of farmland on Britain's east coast. The barrier was completed in 1983--i.e., 30 years after the decision to build was made. Parallel developments in storm surge warn- ing were fundamental to the operation of the barrier. Since the barrier's completion, London's derelict docklands have been regenerated with new homes and businesses, including the new financial district around Canary Wharf. Signif- icant future development is expected in potentially flood-prone sites alongside the tidal Thames. The design life of the barrier is until 2030, when rising flood levels due to a combination of global sea-level rise and more local changes will reduce the residual flood risk to below a 1-in-1,000-year standard. Given the long lead- time to upgrade the defenses, planning of the flood defenses to 2100 is already in its early stages. This includes con- sideration for the first time of inland realignment of the flood defense line as a complimentary strategy to raising defenses, and as such reverses a long-term trend of encroachment and land claim into the tidal Thames (Shih 2002). While the fact that there is residual flood risk even with the existing defenses is not widely appreciated, all insured properties in London still receive flood cover. Therefore, losses would be shared in the unlikely event of failure. How- ever, the U.K. insurance industry is concerned about its exposure to flooding and that flood insurance might be selec- tively withdrawn: a situation that would arguably reduce the resilience of Britain to major flood events (Clark 1998). Box 8.2 Responding to human-induced subsidence in Shanghai, China Shanghai subsided as much as 2.8 meters during the 20th century due to groundwater withdrawal (Han and others 1995; Wang and others 1995). The subsidence was triggered by the growing city and economy in the 1920s and it continued until the 1960s, when groundwater withdrawal was regulated and subsidence rates were reduced to 3 to 4 millimeters per year--rates of subsidence one would expect in a deltaic setting. Therefore, while human action trig- gered the subsidence problem, it also made it possible to reduce the subsidence to a large degree. Shanghai was flood- prone before the subsidence occurred, but a substantial increase in both the incidence of flooding and the area affected occurred due to the subsidence. A range of new flood protection measures was required, culminating in large new floodwalls built in the early 1990s, which protect the main city to a 1-in-1,000-year standard. However, future subsi- dence problems remain possible. Anecdotal reports suggest that illegal groundwater withdrawal has increased in Shang- hai over the last 10 years, increasing the rate of subsidence again. This illustrates the ongoing nature of managing weather-related hazards. The Resilience of Coastal Megacities to Weather-Related Hazards 109 from the international climate change community given stresses and/or removing barriers to migration (such the threat of human-induced climate change: as establishing eco-corridors). · Reversing trends that increase vulnerability ("mal- The Parties should take precautionary measures to antic- adaptation")--for example, by introducing setbacks ipate, prevent or minimize the causes of climate change and mitigate its adverse effects. Where there are threats for new development or relocation of existing of serious or irreversible damage, lack of full scientific cer- development in vulnerable areas such as flood- tainty should not be used as a reason for postponing plains and coastal zones. such measures, taking into account that policies and meas- · Improving societal awareness and preparedness-- ures to deal with climate change should be cost-effective for example, by informing the public of the risks so as to ensure global benefits at the lowest possible cost. and possible consequences of climate change and/or The threat of climate change is extending the scope setting up early-warning systems. of reduction strategies for weather-related hazards and As with the approaches listed in table 8.4, these in particular focusing attention over many decades approaches are not mutually exclusive. into the future. The Intergovernmental Panel on Cli- Each of these five objectives of adaptation is rele- mate Change, Third Assessment Report includes a vant for hazard reduction in coastal megacities. How- dedicated chapter on adaptation for the first time ever, for coastal zones, another classification of adaptation (Smit and others 2001). Coastal zones are an area where options is often used. This classification, introduced there has been particular interest in adaptation, given by IPCC CZMS (1990) and still the basis of many coastal the inevitability of global-mean sea-level rise (Klein and adaptation analyses, distinguishes between the follow- others 2000; 2001; Tol and others forthcoming), includ- ing three basic hazard management strategies: ing some international projects funded by the Global · Protect--to reduce the risk of coastal hazards by Environmental Facility and others: Caribbean Planning decreasing their probability of occurrence; for Adaptation to Climate Change (CPACC); and The · Retreat--to reduce the risk of coastal hazards by lim- Pacific Islands Climate Change Assistance Programme iting their potential effects; and (PICCAP) (see Nicholls and Mimura 1998). Some coastal · Accommodate--to increase society's ability to cope countries such as Britain and Japan are at the forefront with the effects of coastal hazards. of planning for climate change, including the implica- Klein and others (2001) discuss these three strate- tions for their major coastal cities (e.g., box 8.1). gies in detail and provide examples of technologies for In general, proactive adaptation is aimed at reduc- implementing each of them. While the main hazard ing a system's vulnerability by either minimizing risk considered is sea-level rise, their approach is relevant or maximizing adaptive capacity. Five generic objectives to all weather-related hazards in coastal areas. of anticipatory adaptation can be identified (Klein and The three coastal adaptation strategies roughly Tol 1997; Klein 2002 forthcoming): coincide with the first three of the five proactive adap- · Increasing robustness of infrastructural designs tation objectives listed above. Protecting coastal zones andlong-terminvestments--forexample,byextend- against sea-level rise and other climatic changes would ing the range of temperature or precipitation a system involve increasing the robustness of infrastructural can withstand without failure and/or changing a designs, and long-term investments such as seawalls system's tolerance of loss or failure (e.g., by increas- and other coastal infrastructure. Efficient management ing economic reserves or insurance). of beach and coastal sediments is also an important strat- · Increasing flexibility of vulnerable managed egy to maintain and enhance soft defenses in many systems--for example, by allowing mid-term adjust- coastal cities, which can also sustain recreational and ments (including change of activities or location) other functions (although this could also been seen as and/or reducing economic lifetimes (including increas- enhancing adaptability). ing depreciation). A retreat strategy would serve to enhance the adapt- · Enhancing adaptability of vulnerable natural sys- ability (or resilience) of coastal wetlands by allowing tems--for example, by reducing other (nonclimatic) them space to migrate to higher land as sea level rises. 110 Building Safer Cities: The Future of Disaster Risk Historically coastal ecosystems in urban areas have great deal of uncertainty about the future, one approach suffered encroachment and destruction, but in the future, that is increasingly advocated is adaptive management, proactive adaptation could allow other development or learning by doing (Green and others 2000; Nicholls pathways to occur. A strategy to accommodate sea-level and others 2000). First suggested by Holling (1973) and rise could include increasing the flexibility or coping developed in subsequent papers (Holling 1986; 1997), capacity of managed systems. Examples include rais- it proposes that any intervention should be treated as ing buildings above flood levels to minimize flood an experiment from which lessons should be drawn and damage, as is already practiced in the United States, future interventions made using these lessons. This phi- and/or sharing losses via insurance mechanisms. losophy has been embraced in other decisionmaking While protection has historically dominated the response contexts such as coastal management (National Research to hazards in urban areas, proactive adaptation will pres- Council 1995). This approach explicitly acknowledges ent opportunities to retreat or accommodate in some our incomplete knowledge of the systems that we are urban settings. trying to manage and can easily incorporate change from The other two objectives of proactive adaptation, whatever cause. It also accepts that the systems that we reversing maladaptive trends and increasing awareness are managing are open and, hence, optimal solutions and preparedness, are relevant for coastal zones and in an engineering sense are difficult to define. cities as well. However, given that a whole city built in The experimental approach of adaptive management a low-lying location might be considered maladaptive, makes it explicit that there are risks and uncertainties the scope for relocation in cities is more limited. Once associated with any decision, including those concern- a large city emerges, there is a large inertia against relo- ing hazard reduction. The strength of the adaptive man- cation due to the high level of investment (see box agement philosophy is that this fact is not suppressed 8.1). Hence, cities tend to develop increasing depend- or ignored: rather, it is turned into a learning opportu- ence on artificial measures as they evolve and grow, nity, which increases knowledge and improves subse- although some will argue that this is nonsustainable. quent decisions. To date, the authors are unaware of Therefore, relocation should be integrated into plan- adaptive management being applied to hazard-reduc- ning future development and exploiting redevelopment tion strategies in coastal cities, and there is scope for opportunities, including proactively exploiting disaster exploring how this might improve existing practice. recovery (UNCHS 2001). In this way, disaster recovery Note that many of the ideas concerning adaptive man- and long-term disaster prevention and preparedness agement are linked to the conceptualization of resilience can be linked for coastal cities. as discussed in the next section. Sustaining the functioning of natural coastal systems This brief review shows that there is a range of hazard around cities could also have many benefits, including reduction strategies available for coastal megacities. mitigating weather-related hazards. Thus, hazard mit- While continued technology development may further igation needs to be seen as one goal within integrated increase the detailed options that are available, new coastal zone management, and all hazard mitigation problems and issues will continue to emerge: climate strategies need to address the wider coastal zone con- change is only one example of this. Reducing losses/ text in which they are implemented (cf. Bijlsma and protection has been the main response in the past and others 1996; Klein and others 1999). this seems likely to continue. However, the implemen- Another key point about the effective implementa- tation of proactive adaptation and the utilization of adap- tion of hazard reduction strategies is that they are more tive management principles raise opportunities to use than implementing a set of technical measures and need other approaches as cities develop or are redeveloped, to be thought of as an ongoing process, including including recovery from disaster. This stresses that hazard planning, design, implementation, and monitoring (Klein reduction is an ongoing process rather than a simple set and others 1999; 2001). The examples given for London of technical measures, and it needs to be implemented and Shanghai support this point (boxes 8.1 and 8.2). on this basis. Last, hazard reduction strategies need to Given that we are discussing a process, and there is a be implemented in the wider coastal context and, hence, The Resilience of Coastal Megacities to Weather-Related Hazards 111 comprise one issue within the broader goal of integrated normal quickly, while a highly resilient system may be coastal zone management. quite unstable, in that it may undergo significant fluc- tuation (Handmer and Dovers 1996). Since the seminal work by Holling (1973), resilience Resilience Conceptualized has become an issue of intense conceptual debate among ecologists. The literature provides many perspectives The previous sections of this paper have provided an and interpretations of ecological resilience and, despite overview of the context and importance of coastal megac- 30 years of debate, there appears to be no consensus ities worldwide, the diverse weather-related hazards that on how the concept can be made operational or even coastal megacities face both now and in the future, and how it should be defined. Alternative definitions have the range of strategies that are available to reduce these been provided, focusing on different system proper- weather-related hazards. This section introduces the ties. For example, Pimm (1984) defines resilience as the concept of resilience, as it is assumed that more resilient speed with which a system returns to its original state megacities (as well as other human and natural systems) following a perturbation. are less vulnerable to weather-related and other hazards Other ecologists question the core assumption that (United Nations International Strategy for Disaster Reduc- underpins the concept of resilience, namely, that ecosys- tion 2002a). However, for this assumption to be valid tems exist in an equilibrium state to which they can return and useful, one needs to have an understanding and after experiencing a given level of disturbance. They argue clear definition of resilience, including by which fac- that ecosystems are dynamic and evolve continuously in tors it is determined, how it can be measured, and, most response to external influences taking place on a range importantly how it can be maintained and enhanced. of different time scales. Attempts of ecosystem manage- As stated in the introduction of this paper, there is ment at achieving stability are therefore bound to fail. no literature that deals specifically with resilience in In spite of the relative lack of specificity of ecologi- the context of megacities. However, the concept of cal resilience (or perhaps as a result of it), the concept resilience has been analyzed for a range of other sys- has also gained ground in social science, where it is tems and this literature is reviewed here. The next sec- applied to describe the behavioral response of com- tion discusses the usefulness of the concept in the context munities, institutions, and economies. Timmerman of megacities and weather-related hazards. (1981) has been one of the first to discuss the resilience The Oxford English Dictionary defines resilience as: of society to climate change. In so doing, he links resilience (1) the act of rebounding or springing back; and (2) to vulnerability. He defines resilience as the measure of elasticity. The origin of the word is in Latin, where resilire a system's or part of a system's capacity to absorb and means "to jump back." In a purely mechanical sense, recover from the occurrence of a hazardous event. the resilience of a material is the quality of being able Dovers and Handmer (1992) distinguish between to store strain energy and deflect elastically under a load the reactive and proactive resilience of society. A soci- without breaking or being deformed (Gordon 1978). ety relying on reactive resilience approaches the future However, since the 1970s, the concept has also been by strengthening the status quo and making the pres- used in a more metaphorical sense to describe systems ent system resistant to change, whereas one that devel- that undergo stress and have the ability to recover. ops proactive resilience accepts the inevitability of change Holling (1973) coined the term resilience for ecosys- and tries to create a system that is capable of adapting tems as a measure of the ability of these systems to absorb to new conditions and imperatives. This is an impor- changes and still persist. As such, it determines the tant broadening of the traditional interpretation of persistence of relationships within an ecosystem. This resilience, which is based on the premise of resilience is contrasted with stability, which is defined by Holling being affected by an initial perturbation. The distinc- (1973) as the ability of a system to return to a state of tion by Dovers and Handmer (1992) is based on the equilibrium after a temporary disturbance. Thus, a very major difference between ecosystems and societies: the stable system would not fluctuate greatly but return to human capacity of anticipation and learning. 112 Building Safer Cities: The Future of Disaster Risk Dovers and Handmer (1992) thus link resilience to However, Adger (2000) observes that, while resilience planning and adaptation to hazards. In a later paper is certainly related to stability, it is not clear whether they develop a typology of institutional resilience, which this characteristic is always desirable (cf. Handmer and provides a framework for considering the rigidity and Dovers 1996). inadequacy of present institutional responses to global The above overview of conceptual development of environmental change (Handmer and Dovers 1996). resilience shows that what was once a straightforward They argue that current institutions and policy processes concept used only in mechanics is now a complex multi- appear to be locked into a type of resilience that is char- interpretable concept with contested definitions and acterized by change at the margins. Responses to envi- even relevance. Nonetheless, the concept of resilience ronmental change are shaped by what is perceived to is now used in a great variety of interdisciplinary work be politically and economically palatable in the near term concerned with the interactions between people and rather than by the nature and scale of the threat itself. nature, including disaster reduction (UN/ISDR 2002a). This type of resilience, as well as a type that is char- The most important development over the past 30 years acterized by resistance to change, provides some level is the increasing recognition across disciplines that of stability in society, although there is a potentially large human and ecological systems are interlinked and that risk that this apparent stability is not sustainable and their resilience relates to the functioning and interac- could lead to collapse if society cannot make the social, tion of the systems rather than to the stability of their economic, and political changes necessary for survival. components or the ability to maintain or return to some The third type of resilience described by Handmer and equilibrium state. Dovers (1996), one that is characterized by openness For coastal systems, the self-organizing capacity and adaptation, is more likely to deal directly with the implied by resilience is considered a desirable prop- underlying causes of environmental problems and erty, given the uncertainty about the future (Nicholls reduces vulnerability by having a high degree of flexi- and Branson 1998). However, resilience is most easily bility. Its key feature is a preparedness to adopt new conceptualized in the case of natural systems. The role basic operating assumptions and institutional struc- of social systems in contributing to resilience and its tures. However, there is also a potentially large risk relationship to the resilience of natural systems is less involved in moving toward this type of resilience. Change clear. In a conceptual study on the resilience of the Dutch deemed as necessary could turn out to be maladaptive, coast, Klein and others (1998) focus on the function- rendering a large cost to society. ing of morphological, ecological, and socioeconomic Adger (2000) investigates the links between social processes in determining coastal resilience. These resilienceandecologicalresilience.HefollowsTimmerman processes produce a coastal system that is continu- (1981) in his definition of social resilience: the ability ously changing, so no original or equilibrium state can of human communities to withstand external shocks be identified. Moreover, perturbations are not isolated or perturbations to their infrastructure, such as envi- events from which a coastal system may or may not ronmental variability or social, economic, or political recover, but are ever-present and occur at different tem- upheaval, and to recover from such perturbations. Social poral and spatial scales. Klein and others (1998) define resilience is measured through proxies of institutional coastal resilience as the self-organizing capacity of the change and economic structure, property rights, access coast to preserve actual and potential functions under to resources, and demographic change (Adger 1997). the influence of changing hydraulic and morphologi- It is argued by many ecologists that resilience is the cal conditions. This capacity is based on the (potential) key to sustainable ecosystem management and that dynamics of morphological, ecological, and socioeco- diversity enhances resilience, stability, and ecosystem nomic processes and the demands for these processes functioning (e.g., Schulze and Mooney 1993; Peterson made by the functions to be preserved. The key ques- and others 1998; Chapin and others 2000). Ecological tion about the relationship and exchangeability between economists also argue that resilience is the key to sus- the different types of resilience remains to be resolved. tainability in the wider sense (e.g., Common 1995). A conceptual model of vulnerability to sea-level rise, The Resilience of Coastal Megacities to Weather-Related Hazards 113 which combines the concept of resilience with that of operational tool for policy and management purposes: resistance, is also presented (see also Klein and Nicholls a challenge that 30 years of academic debate does not 1999). This implies that resilience and resistance seem to have resolved. might be exchangeable, if, for example, a natural soft coast were armored with a seawall, or the opposite process, which is termed managed retreat. Discussion The recognition that human and ecological systems are interlinked has led to the establishment of the The vulnerability of megacities to hazards and disas- Resilience Alliance, a network of scientists with roots ters has been a subject of increasing academic interest, mainly in ecology and ecological economics, which aims with recent special issues of GeoJournal (Parker and to stimulate academic research on resilience and inform Mitchell 1995), Applied Geography (Mitchell 1998), and the global policy process on sustainable development Ocean & Coastal Management (Barbière and Li 2001), (Folke and others 2002). The Resilience Alliance con- as well as influential publications by Timmerman and sistently refers to social-ecological systems and defines White (1997), Rakodi and Treloar (1997), Mitchell their resilience by considering three distinct dimensions (1999), and Cross (2001). This academic interest has (Carpenter and others 2001): complemented the increasing policy interest, as reflected · The amount of disturbance a system can absorb by initiatives of the International Decade for Natural and still remain within the same state or domain of Disaster Reduction (and now the International Strat- attraction; egy for Disaster Reduction--ISDR) and the Disaster · The degree to which the system is capable of self- Management Facility of the World Bank. organization; and It is clear that climate change will increase the hazard · The degree to which the system can build and increase potential in many cities (Scott and others 2001). In a par- the capacity for learning and adaptation. allel paper, Bigio (2002) provides an overview of the chal- This definition is an amalgamation of the aforemen- lenges that climate change poses to cities. This paper has tioned definitions of ecological, social, and institutional focused on coastal megacities, as they concentrate a grow- resilience. ing part of the world's human population in potentially The United Nations International Strategy for Dis- hazardous locations. In 2000, the cities with projected aster Reduction (UN/ISDR 2002a) uses the term resilience populationsexceeding8millionby2015containedabout as follows: 250 million people (table 8.1), which is about 4 percent of the world's population. Similarly, using 1990 data, The capacity of a system, community or society to resist or to change in order that it may obtain an acceptable level Nicholls and Small (2002) identify about 120 million in functioning and structure. This is determined by the people living in the near-coastal zone at the high- degree to which the social system is capable of organis- population densities associated with large urban areas ing itself and the ability to increase its capacity for learn- (>10,000 people/square kilometer). The strong global ing and adaptation, including the capacity to recover from tendency to urbanization, both in these large coastal a disaster. megacities and in the more numerous smaller cities, com- However, resilience remains at the conceptual level bined with a general tendency for migration toward the throughout this document and operationalizing the con- coast, suggests that coastal urban centers will contain a cept is not discussed. significant proportion of the world's population by 2050, Hence, both in the academic realm of the Resilience if not earlier (although it is important not to forget the Alliance and in the applied realm of the ISDR, the large nonurban coastal population (Nicholls and Small same problems as with previous definitions persist: there 2002)). Most of these urban centers will be in the devel- is a danger that there is a "motherhood and apple pie" oping world, including at least 17 coastal megacities. view on resilience, which leaves limited scope for meas- Population density and urban centrality make it likely urement, testing, and formalization. The challenge that the disaster problems of large cities will be different remains to transform the concept of resilience into an from those of rural areas. In large cities, events with 114 Building Safer Cities: The Future of Disaster Risk relatively small areas of impact affect great numbers of are competing processes influencing resilience and vul- people (Kelly 1995). Nonetheless, the scale, speed, nerability, which are dynamic and not fully understood. and complexity of urbanization are neglected topics in Resilience is seen as an important characteristic of most investigations of hazard prediction and response. megacities that helps to reduce the vulnerability of its In addition, there is a lack of historical evidence about citizens to weather-related hazards. However, as shown disasters in very large cities, mainly because such cities in the section "Resilience Conceptualized," resilience are a recent phenomenon (Mitchell 1993). is a relatively poorly defined concept that has not yet In spite of the high hazard potential of megacities in been made operational for policy and management. Fol- general and coastal megacities in particular, there is no lowing Timmerman (1981), there seems to be a con- compelling evidence that megacities are more vulnerable sensus that a resilient city is less vulnerable to hazards to hazards than are smaller cities and towns. Handmer but no systematic and reproducible analysis exists to (1995) argues that major cities have inherent features date as to what makes cities resilient and how resilience that enable them to deal with hazards more effectively can be enhanced. than smaller settlements. The immense power and In this section, the authors revisit this consensus view resources of large cities confer considerable capacity to by asking the following questions: respond (i.e., resilience). Most major cities are able to · Is resilience a desirable attribute of megacities? harness massive financial resources and expertise from · Does enhanced resilience reduce the vulnerability within the city, the country, and the rest of the world to of megacities to weather-related hazards? combat disaster and aid recovery. Parker (1995) sup- · Is resilience a useful concept in the hazard manage- ports this view and argues that the built-in complexities ment of megacities? and redundancies characteristic of very large urban sys- tems and the modern global electronic trading systems Is Resilience a Desirable Attribute of Megacities? of which they are part may also enhance resilience. Cross (2001) also emphasizes the greater resilience Whether or not resilience is a desirable attribute of of megacities compared to small towns. He argues that megacities depends on the definition of the concept. the different response capacities of smaller communi- The traditional definitions that assume some equilib- ties profoundly influence the long-term consequences rium state may be outdated according to some, but they of a disastrous event on the individual victims and still tend to capture the imagination of many when whether they receive timely or adequate emergency resilience is mentioned. It is clear that megacities are in assistance. Individuals in both small communities and a continuous state of flux and that "bouncing back" to megacities are vulnerable to hazard losses, but losses the original state after a disaster is impossible. More for residents of large cities are more easily reduced by importantly, if a megacity is struck by a disaster, it fol- the warning and protection systems that the cities' lows that the original state was one in which it was vul- concentrated wealth can justify. nerable to disaster in the first place. Going back to this On the other hand, as discussed in the earlier sec- original state is undesirable, as it would leave the city tion "Weather-related Hazards in Coastal Zones," urban- just as vulnerable to the next disaster. ization in the developing world is also concentrating Later definitions of resilience focus on the function- poor populations in potentially hazardous areas and ing of systems, including their self-organizing capacity. thus raising the vulnerability of these groups and Resilience interpreted in this manner is desirable in hence the city as a whole to hazardous events and dis- megacities. While it does not help to prevent disasters asters. This increase and concentration of vulnerability or reduce their immediate impact, once a disaster hap- attracts considerable attention in the literature (UNCHS pens, it facilitates and contributes to the process of recov- 2001) and is a process likely to reduce the resilience of ery. A resilient megacity thus would be less likely to cities. Hence, while coastal megacities in the devel- experience a severe lasting impact from a disaster. oped world might be seen as more resilient than smaller Recently, resilience has also been interpreted as includ- settlements or rural areas, in the developing world there ing the degree to which a system can build and increase The Resilience of Coastal Megacities to Weather-Related Hazards 115 the capacity for learning and adaptation (Carpenter and on the assumed definitions of resilience, none of others 2001; cf. proactive resilience as defined by Dovers which are operational. After 30 years of academic analy- and Handmer 1992). The capacity for learning and adap- sis and debate, the definition of resilience has become tation is clearly a desirable attribute, although few would so broad as to render it almost meaningless. The afore- associate intuitively the ability to increase this capacity mentioned definition by Carpenter and others (2001) with resilience. This interpretation of resilience relates to includes many issues that are currently en vogue in adaptive management and adaptive capacity: two con- discussions on sustainable development and hazard cepts with their own literature and interpretations, but management. Rather than the definition providing an perhaps more operational for policy and management. explanation of a system attribute that is observed and can be measured, resilience has become an umbrella con- Does Enhanced Resilience Reduce the Vulnerability of cept for a range of system attributes that are deemed desir- Megacities to Weather-Related Hazards? able. The system attributes and the umbrella concepts need to be kept distinct in a conceptual hierarchy. As it was for the previous question, this is a matter of It is uncertain whether more rigorous academic work definition. The early interpretations of resilience would will lead to a concept of resilience that does have direct not reduce vulnerability; later ones would. However, it relevance to hazard management. Alternatively, the term is also important to consider who or what would be vul- resilience should be used only to refer to a system's nerable and how this vulnerability would manifest itself. reactive response to a perturbation. This would corre- A megacity typically covers a large and often physio- spond to the first two dimensions of the definition by graphically heterogeneous area, with different exposure Carpenter and others (2001): the amount of disturbance and susceptibility to hazards. In addition, the population a system can absorb and still remain within the same will be diverse, as will be the conditions under which state or domain of attraction and the degree to which they live. As a result, while a megacity has a particular the system is capable of self-organization. For human vulnerability to hazards, some population groups systems, this is also referred to as "coping ability" in within the city may be particularly vulnerable due to their the hazard management literature. high exposure and unfavorable socioeconomic situation. The proactive type of resilience corresponds to what Resilience interpreted as facilitating and contribut- hasbeencoined"adaptivecapacity"bytheclimatechange ing to the process of recovery after a disaster is irrele- community. Adaptation in the context of climate change vant to those who lose their lives during a disaster. Those refers to any adjustment that takes place in natural or losing their marginal livelihoods in shantytowns may human systems in response to actual or expected impacts not benefit as much from being able to display resilience of climate change, aimed at moderating harm or exploit- as those who could afford insurance to cover any damage ing beneficial opportunities. Planning for adaptation is to their property. This shows that equity is an important becoming increasingly important, especially since it has factor in determining whether resilience reduces the beenestablishedthathumansare--atleastinpart--respon- vulnerability of megacities to weather-related hazards. sible for climate change and that impacts can no longer be avoided only by reducing greenhouse-gas emissions Is Resilience a Useful Concept in the Hazard (Arnell and others 2002). Adaptive capacity is defined as Management of Megacities? the ability to plan, prepare for, and implement adapta- The fact that, among others, the United Nations Inter- tion options. Factors that determine adaptive capacity to national Strategy for Disaster Reduction (2002a) has climate change include economic wealth, technology adoptedthetermresiliencewouldsuggestthatitisauseful and infrastructure, information, knowledge and skills, concept for hazard management, including that involv- institutions, equity, and social capital (Smit and others ing coastal megacities. However, the problems with 2001). Hence, resilience as defined in the previous para- resilience include the multitude of different definitions graph contributes to the overall adaptive capacity. and turning any of them into operational concepts. For two years, the concept of adaptive capacity has The answers to the previous two questions depended been the subject of worldwide interdisciplinary research 116 Building Safer Cities: The Future of Disaster Risk efforts, aimed at making it operational for the interna- These specific attributes are more amenable to meas- tional climate policy process, as well as for national plan- urement and monitoring, although questions about ning agencies (Smith and others 2002). The concept the relationship between natural system and social system has gained recognition in climate policy and science resilience remain to be fully explored. and is now used outside the climate community as In this conceptual framework, resilience is one prop- well (e.g., Turton 1999). Climate variability is increas- erty that contributes to the overall adaptive capacity of ingly considered along with climate change when the system in question. The narrower definition of planning for adaptation, as it is recognized that in many resilience proposed here differs from the approach fol- areas the most direct and immediate impacts of climate lowed by the ISDR, which uses it in a broader sense change will occur through changes in the frequency and that has similarities to the definition of adaptive capac- intensity of weather-related hazards. ity (UN/ISDR 2002a). The use of the concept of adaptive capacity by the In the case of coastal megacities, maintaining and natural hazard community would not only foster much- enhancing both resilience and adaptive capacity for needed communication between that community and weather-related hazards would be desirable policy and the climate community. More importantly, it would pro- management goals, although based on the hierarchy, vide hazard managers with a tool that is similar to maintaining and enhancing adaptive capacity is the over- resilience in its relationship to vulnerability, but offers all goal. The framework for resilience and adaptive capac- greater potential in application, especially when attempt- ity proposed here has benefits in terms of linking: ing to move away from disaster recovery to disaster pre- · Analysis of present and future hazardous conditions vention and preparedness. (being addressed by climate change); and · Enhancement of the capacity for disaster prevention and preparedness with disaster recovery. Conclusions This is consistent with the challenges for the future identified by United Nations International Strategy for While resilience is widely seen as a desirable property Disaster Reduction (2002a). of natural and social systems, including coastal megac- ities, the term has been used in a number of different Note ways. Some authors have used it to define specific system 1. This analysis only considers the thermal expansion compo- attributes, while others have used it to characterize the nent of sea level change. entire system as an umbrella concept. This leads to con- siderable confusion. Without an explicit operational definition, resilience only has the broadest meaning and Bibliography remains a concept rather than a policy or management Adger, W.N. 1997. Sustainability and Social Resilience in Coastal tool. At the same time, work on adaptation to climate Resource Use. Centre for Social and Economic Research on change has developed and is now operationalizing the the Global Environment, University of East Anglia, Norwich, concept of adaptive capacity, which is defined as the and University College London, London. ability to plan, prepare for, and implement adaptation Adger, W.N. 2000. "Social and Ecological Resilience: Are They options. Related?" Progress in Human Geography 24: 347­364. Based on the present knowledge, the authors feel that Arnell, N.W., M.G.R. Cannell, M. Hulme, R.S. Kovats, J.F.B. the definition of resilience is best used to define spe- Mitchell, R.J. Nicholls, M.L. Parry, M.T.J. Livermore, and A. cific system attributes, particularly: White. 2002. "The Consequences of CO2 Stabilisation for the · The amount of disturbance a system can absorb and Impacts of Climate Change. Climatic Change 53: 413­446. still remain within the same state or domain of attrac- Barbière, J. and H. Li. 2001. "Editorial, Third Millennium Special tion; and Issue on Megacities." OceanandCoastalManagement 44(4), v­ix. · The degree to which the system is capable of self- Bigio, A.G. 2002. Cities and Climate Change. Urban Development organization. Division, The World Bank, Washington, D.C. The Resilience of Coastal Megacities to Weather-Related Hazards 117 Bijlsma, L., C.N. Ehler, R.J.T. Klein, S.M. Kulshrestha, R.F. McLean, Davison, A.T., R.J. Nicholls, and S.P. Leatherman. 1992. "Beach N. Mimura, R.J. Nicholls, L.A. Nurse, H. Pérez Nieto, E.Z. Nourishment as a Coastal Management Tool: An Annotated Stakhiv, R.K. Turner, and R.A. Warrick. 1996. "Coastal Zones Bibliography on Developments Associated with the Artificial and Small Islands." In R.T. Watson, M.C. Zinyowera, and R.H. Nourishment of Beaches." Journal of Coastal Research 8(4): Moss, eds., Impacts, Adaptations and Mitigation of Climate Change: 984­1,022. Scientific-technical Analyses. Contribution of Working Group Dennis, K.C., I. Niang-Diop, and R.J. Nicholls. 1995. "Sea-level II to the Second Assessment Report of the Intergovernmen- Rise and Senegal: Potential Impacts and Consequences." tal Panel on Climate Change. Cambridge, U.K.: Cambridge Journal of Coastal Research Special Issue 14: 243­261. University Press: 289­324. Devine, N.P. 1992. Urban Vulnerability to Sea-level Rise in the Third Bird, E.C.F. 1985. Coastal Changes: A Global Review. Chichester, World. M.S. Thesis. New Brunswick, N.J.: Rutgers, The State U.K.: Wiley. University of New Jersey. Boesch, D.F., M.N. Josselyn, A.J. Mehta, J.T. Morris, W.K. Nuttle, Dovers, S.R. and J.W. Handmer. 1992. "Uncertainty, Sustain- C.A. Simenstad, and D.J.P. Swift. 1994. "Scientific Assessment ability and Change." Global Environmental Change 2: 262­276. of Coastal Wetland Loss, Restoration and Management in Folke, C., S. Carpenter, T. Elmqvist, L. Gunderson, C.S. Holling, Louisiana." Journal of Coastal Research, Special Issue 20; v B. Walker, J. Bengtsson, F. Berkes, J. Colding, K. Danell, M. and 103. Falkenmark, L. Gordon, R. Kasperson, N. Kautsky, A. Burton, I., R.W. Kates, and G.F. White. 1993. The Environment Kinzig, S. Levin, K.-G. Mäler, F. Moberg, L. Ohlsson, P. Olsson, as Hazard. Second Edition. New York: Guilford Press. E. Ostrom, W. Reid, J. Rockstroem, H. Savenije, and U. Svedin. Carpenter, S., B. Walker, J.M. Anderies, and N. Abel. 2001. "From 2002. "Resilience and Sustainable Development: Building Metaphor to Measurement: Resilience of What to What?" Adaptive Capacity in a World of Transformations." The Envi- Ecosystems 4: 765­781. ronmental Advisory Council to the Swedish Government, Chapin, F.S., E.S. Zavaleta, V.Y. Eviner, R.L. Naylor, P.M. Vitousek, Stockholm. H.L. Reynolds, D.U. Hooper, S. Lavorel, O.E. Sala, S.E. Hobbie, Gilbert, S. and R. Horner. 1984. The Thames Barrier. London: M.C. Mack, and S. Diaz. 2000. "Consequences of Changing Thomas Telford:182. Biodiversity." Nature 405: 234­242. Gordon, J.E. 1978. Structures. Harmondsworth, U.K.: Penguin Church, J.A., J.M. Gregory, P. Huybrechts, M. Kuhn, K. Lambeck, Books. M.T. Nhuan, D. Qin, and P.L. Woodworth. 2001. "Changes Green, C., R.J. Nicholls, C. Johnson, S. Schackley, and J.W. Hand- in Sea Level." In T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, mer. 2000. Climate Change Adaptation: a Framework for P.J. van der Linden, X. Dai, K. Maskell, and C.A. Johnson, eds., Analysis and Decision-making in the Face of Risk and Uncer- Climate Change 2001: The Scientific Basis. Contribution of Work- tainty Report No. 28. A report produced for NCRAOA as part ing Group I to the Third Assessment Report of the Intergov- of Environment Agency R and D Project E2-036. Environ- ernmental Panel on Climate Change. Cambridge, U.K.: ment Agency: 87. Cambridge University Press: 639­693. Hamm, L., H. Hanson, M. Capobianco, H.H. Dette, A. Lechuga, Clark, M. 1998. "Flood Insurance as a Management Strategy for and R. Spanhoff. 1998. "Beach Fills in Europe--Projects, Prac- U.K. Coastal Resilience." The Geographical Journal 164(3): tises and Objectives." Proceedings of the 26th Coastal Engineer- 333­343. ing Conference in Copenhagen. New York, N.Y.: American Common, M. 1995. Sustainability and Policy: Limits to Economics. Society of Civil Engineers: 3,060­3,073. Cambridge, U.K.: Cambridge University Press. Han, M., J. Hou, and L. Wu 1995. "Potential Impacts of Sea- Costanza, R., M. Kemp, and W. Boynton. 1995. "Scale and Bio- level Rise on China's Coastal Environment and Cities: A National diversity in Estuarine Ecosystems." In C. Perrings, K.G. Mäler, Assessment." Journal of Coastal Research Special Issue No. 14: C. Folke, C.S. Holling, and B.O. Jansson, eds., Biodiversity 79­95. Loss: Economic and Ecological Issues. Cambridge, U.K.: Cam- Handmer, J.W. 1995. "Managing Vulnerability in Sydney: Plan- bridge University Press: 84­125. ning or Providence?" GeoJournal 37(3): 355­368. Cross, J.A. 2001. "Megacities and Small Towns: Different Per- Handmer, J.W. and S.R. Dovers. 1996. "A Typology of Resilience: spectives on Hazard Vulnerability." Environmental Hazards 3: Rethinking Institutions for Sustainable Development." Indus- 63­80. trial and Environmental Crisis Quarterly 9: 482­511. 118 Building Safer Cities: The Future of Disaster Risk Harrison, P. and F. Pearce. 2000. AAAS Atlas of Population and be applied?" Mitigation and Adaptation Strategies for Global Environment (American Association for the Advancement of Change 4: 51­64. Science). Berkeley, CA: University of California Press, Berkeley: Klein, R.J.T., M.J. Smit, H. Goosen, and C.H. Hulsbergen. 204. 1998. "Resilience and Vulnerability: Coastal Dynamics or Holling, C.S. 1973. "Resilience and Stability of Ecological Sys- Dutch Dikes?" The Geographical Journal 164(3): 259­268. tems." Annual Review of Ecological Systems 4: 1­23. Klein, R.J.T., J. Aston, E.N. Buckley, M. Capobianco, N. Mizutani, Holling, C.S. 1986. "The Resilience of Terrestrial Ecosystems: R.J. Nicholls, P.D. Nunn, and Ragoonaden, S. 2000. "Coastal Local Surprise and Global Change." In W.C. Clark and R.E. Adaptation Technologies." In B. Metz, O.R. Davidson, J.W. Munn, eds., Sustainable Development of the Biosphere. Cam- Martens, S.N.M. van Rooijen, and L.L. Van Wie McGrory, eds., bridge, U.K.: Cambridge University Press: 292­320. IPCC Special Report on Methodological and Technological Issues Holling, C.S. 1997. "Regional Responses to Global Change." in Technology Transfer. Cambridge: Cambridge University Press: (Editorial) Conservation Ecology 1(1). Available from 349­372. . Klein, R.J.T., R.J. Nicholls, S. Ragoonaden, M. Capobianco, J. Houghton, T., Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, Aston, and E.N. Buckley. 2001. "Technological Options for X. Dai, K. Maskell, and C.A. Johnson, eds. 2001: Climate Adaptation to Climate Change in Coastal Zones." Journal of Change 2001: The Scientific Basis. Contribution of Working Coastal Research 17(3): 531­543. Group I to the Third Assessment Report of the Intergovern- Mimura, N., M. Isobe, and Y. Hosakawa. 1992. "Impacts of Sea mental Panel on Climate Change. Cambridge, U.K.: Cam- Level Rise on Japanese Coastal Zones and Response Strate- bridge University Press. gies." In J. O'Callaghan, ed. Global Climate Change and the IPCC CZMS (Intergovernmental Panel on Climate Change, Rising Challenge of the Sea. Proceedings of the Third IPCC CZMS Response Strategies Working Group, Coastal Zone Manage- Workshop, Margarita Island, Venezuela, March 9­13, 1992. ment Subgroup). 1990. Strategies for Adaptation to Sea Level National Oceanic and Atmospheric Administration. Silver Rise. Report of the Coastal Zone Management Subgroup. The Spring, MD: 329­349. Hague: Ministry of Transport, Public Works and Water Man- Mitchell, J.K. 1993. "Natural Hazard Predictions and Responses agement: x and 122. in Very Large Cities." In J. Nemec, J. M. Nigg, and F. Siccardi, Kelly, M.P. 1991. Global Warming: Implications for the Thames eds. Prediction and Perception of Natural Hazards. Dordrecht, Barrier and Associated Defences. In R. Frasetto, ed., Impact of The Netherlands: Kluwer Academic Publishers: 29­37. Sea Level Rise on Cities and Regions. Proceedings of the First Mitchell, J.K. 1998. "Introduction--Hazards in Changing Cities." International Meeting "Cities on Water" in Venice, Italy, Applied Geography Special Issue 18(1): 1­6. December 1989. Venice: Marsilio Editori: 93­98. Mitchell, J.K., ed. 1999. Crucibles of Hazard: Megacities and Dis- Kelly, C. 1995. "Assessing Disaster Needs in Megacities: Per- asters in Transition. Tokyo: United Nations University Press: spectives from Developing Countries." GeoJournal 37(3): 535. 381­385. Muehe, D. and C.F. Neves. 1995. "The Implications of Sea-level Klein,R.J.T.(Forthcoming)."AdaptationtoClimateVariabilityand Rise on the Brazilian Coast: A Preliminary Assessment." Jour- Change: What is Optimal and Appropriate?" In C. Giupponi nal of Coastal Research Special Issue 14: 54­78. and M. Schechter, eds., Climate Change and the Mediterranean: National Research Council. 1995. Science, Policy and the Coast: Socio-Economics of Impacts, Vulnerability and Adaptation. Chel- Improving Decision-Making. Washington, D.C.: National Acad- tenham, U.K.: Edward Elgar. emy Press. Klein, R.J.T and R.J. Nicholls. 1999. "Assessment of Coastal Vul- Nicholls, R.J. 1995. "Coastal Megacities and Climate Change." nerability to Climate Change." Ambio 28(2): 182­187. GeoJournal 37(3): 369­379. Klein, R.J.T. and R.S.J. Tol. 1997. "Adaptation to Climate Change: Nicholls, R.J. 2000. "An Analysis of the Flood Implications of Options and Technologies--An Overview Paper." Technical the IPCC Second Assessment Global Sea-level Rise Scenar- Paper FCCC/TP/1997/3. United Nations Framework Con- ios." In D.J. Parker, ed., Floods. London: Routledge: 148­162. vention on Climate Change Secretariat. Bonn, Germany: 33. Nicholls, R.J. and J. Branson, eds. 1998. "Enhancing Coastal Klein, R.J.T., R.J. Nicholls, and N. Mimura. 1999. "Coastal Adap- Resilience--Planning for an Uncertain Future." TheGeographical tation to Climate Change: Can the IPCC Technical Guidelines Journal 164(3): 255­343. The Resilience of Coastal Megacities to Weather-Related Hazards 119 Nicholls, R.J. and N. Mimura. 1998. "Regional Issues Raised by Adaptation and Vulnerability. Contribution of Working Group II Sea-level Rise and Their Policy Implications." Climate Research to the Third Assessment Report of the Intergovernmental Panel 11: 5­18. on Climate Change. Cambridge: Cambridge University Nicholls, R.J. and C. Small. 2002. "Improved Estimates of Coastal Press: 381­416. Population and Exposure to Hazards Released." EOS Trans- Shih, S. 2002. A Framework for Analysis of Managed Realignment actions 83(2): 301, 305. for the Thames Estuary. MSc Thesis. London: Middlesex Uni- Nicholls, R.J., C. Johnson, C.P. Green, and S. Shackley. 2000. versity: 88. "An Adaptive Management Framework for Climate Change." Small, C. and J.E. Cohen. 1999. "Continental Physiography, Cli- In A. de la Vega-Leinert, R.J. Nicholls, and R.S.J. Tol, eds., mate and the Global Distribution of Human Population." European Vulnerability and Adaptation to Impacts of Accelerated Proceedings of the International Symposium on Digital Earth. Bei- Sea-level Rise (ASLR). Proceedings of SURVAS Expert Work- jing: Science Press: 965­971. Available from . 19­21, 2000. Flood Hazard Research Centre, Middlesex Smit, B., O. Pilifosova, I. Burton, B. Challenger, S. Huq, R.J.T. University,London:125­139.Availablefrom. the Context of Sustainable Development and Equity." In J.J. Parker, D.J. 1995. "Hazard Transformation and Hazard Man- McCarthy, O.F. Canziano, and N. Leary, eds., Climate Change agement Issues in the London Megacity." GeoJournal 37(3): 2001: Impacts, Adaptation and Vulnerability. Contribution of 313­328. Working Group II to the Third Assessment Report of the Inter- Parker, D.J., ed. 2000. Floods. 2 volumes. London: Routledge: governmental Panel on Climate Change. Cambridge: Cam- 317, 431. bridge University Press: 877­912. Parker, D.J. and J.K. Mitchell. 1995. "Disaster Vulnerability of Smith, J.B., R.J.T. Klein, and S. Huq, eds. 2002. Enhancing the Megacities: An Expanding Problem that Requires Rethinking Capacity of Developing Countries to Adapt to Climate Change. and Innovative Responses." GeoJournal 37(3): 295­301. London: Imperial College Press, London. Peterson, G.D., C.R. Allen, and C.S. Holling. 1998. "Ecological Timmerman, P. 1981. Vulnerability, Resilience and the Collapse of Resilience, Biodiversity and Scale." Ecosystems 1: 6­18. Society: A Review of Models and Possible Climatic Applications. Pielke, R. and C. Landsea. 1998. "Normalised Hurricane Dam- Institute for Environmental Studies, University of Toronto, ages in the United States: 1925­1995." Weather and Forecast- Toronto. ing 13: 621­631. Timmerman, P. and R. White. 1997. "Megahydrolopolis: Coastal Pimm, S.L. 1984. "The Complexity and Stability of Ecosystems." Cities in the Context of Global Environmental Change." Global Nature 307: 321­326. Environmental Change 7(3): 205­234. Rakodi, C. and D. Treloar. 1997. "Urban Development and Coastal Tol, R.S.J., R.J.T. Klein, and R.J. Nicholls. Forthcoming. "Towards Zone Management. An International Review." Third World Plan- Successful Adaptation to Sea Level Rise Along Europe's Coasts." ning Review 19(4): 401­424. Journal of Coastal Research. Rosenzweig, C. and W.D. Solecki, eds. 2001. "Climate Change Turner, R.K., S. Subak, and W.N. Adger. 1996. "Pressures, Trends, and a Global City: The Potential Consequences of Climate and Impacts in Coastal Zones: Interactions Between Socioe- Variability and Change--Metro East Coast." Report for the conomic and Natural Systems." Environmental Management U.S. Global Change Research Program, National Assessment 20(2): 159­173. of the Potential Consequences of Climate Variability and Change Turton, A.R. 1999. "Water Scarcity and Social Adaptive Capac- for the United States, Columbia Earth Institute, New York: ity: Towards an Understanding of the Social Dynamics of Man- 224. aging Water Scarcity in Developing Countries." MEWREW Schulze, E.D. and H.A. Mooney. 1993. Biodiversity and Ecosys- Occasional Paper No. 9, University of London, School of tem Function. Berlin: Springer-Verlag. Oriental and African Studies (SOAS), Water Issues Study Group. Scott, M., S. Gupta, E. Jáuregui, J. Nwafor, D. Satterthwaite, UN/DESA (United Nations Department of Economic and Y.A.D.S. Wanasinghe, T. Wilbanks, and M. Yoshino. 2001. Social Affairs). 2001. World Population Prospects--the 2000 "Human Settlements, Energy and Industry." In J.J. McCarthy, Revision. Population Division, Department of Economic and O.F. Canziano, and N. Leary, eds., Climate Change 2001: Impacts, Social Affairs. New York: United Nations: xii, 64. 120 Building Safer Cities: The Future of Disaster Risk UNEP (United Nations Environment Program). 2002. Global Understanding the Links Between Development, Environment and Environment Outlook 3. London: Earthscan: 446. Natural Disasters. Background document for the World Summit UNCHS Habitat (United Nations Human Settlements Programme). on Sustainable Development (WSSD) No. 5. Geneva. Avail- 2001. Cities in a Globalizing World: Global Report on Human able from . Settlements 2001. London: Earthscan: 344. Wang, B., S. Chen, K. Zhang, and J. Shen. 1995. "Potential UN/ISDR (United Nations International Strategy for Disaster Impacts of Sea-Level Rise on the Shanghai Area." Journal of Reduction). 2002a: Living With Risk: A Global Review of Dis- Coastal Research Special Issue No. 14: 151­166. aster Reduction Initiatives. Prepared as an interagency effort and WCC' 93 (World Coast Conference). 1994. "Preparing to Meet coordinated by the UN/ISDR Secretariat, Geneva. Available the Coastal Challenges of the 21st Century." Report of the from . World Coast Conference, Noordwijk, The Netherlands, Novem- UN/ISDR(UnitedNationsInternationalStrategyforDisasterReduc- ber 1­5, 1993, Ministry of Transport, Public Works and Water tion). 2002b: Natural Disasters and Sustainable Development: Management, The Hague: 49, and appendices. Chapter 9 Flood Management and Vulnerability of Dhaka City Saleemul Huq and Mozaharul Alam Dhaka, the capital and largest city in Bangladesh, was estimated that about 85 percent of the city was inun- established by the Mughal Emperor Jahangir in 1608 dated at depths ranging from 0.3 to over 4.5 meters, on the banks of the river Buriganga. The city is surrounded and about 60 percent of city dwellers were affected. It by the distributaries of the two major rivers, the also disrupted city life, air travel, and communication Brahmaputra and the Meghna. The surrounding rivers from the capital city to the outside world. The 1998 are Buriganga to the south, Turag to the west, Tongi khal flood was most severe in terms of extent and duration. to the north, and Balu to the east. The city and adjoining It was estimated that about 56 percent of the city was areas are composed of alluvial terraces of the southern inundated, including most of the eastern and 23 per- part of the Madhupur tract and low-lying areas of the doab cent of the western parts of the city. The flood protec- of the river Meghna and Lakkha. The combined area of tion embankment and floodwalls along the Turag and Dhaka East and Dhaka West known as Greater Dhaka the Buriganga rivers protect the western part of the city covers an area of approximately 275 square kilometers from river flooding. ( JICA 1991). The elevation of Greater Dhaka is 2 to 13 The Buckland Flood Protection Embankment along meters above the mean sea level, and most of the urban- the river Buriganga was the first attempt to mitigate ized areas are at elevation of 6 to 8 meters above the flood damage in Dhaka City. Flood protection plans mean sea level. The land area above 8 meters above mean for Greater Dhaka have been under study and consid- sea level covers about 20 square kilometers. The land eration for many years, but the catastrophic flooding ranging from 6 to 8 meters above mean sea level covers that occurred in 1987 and 1988 brought into focus the 75 square kilometers, while 170 square kilometers of urgent need to proceed with immediate action. In 1989, Greater Dhaka is below 6 meters above mean sea level construction activities commenced on a "crash program" ( JICA 1987). The highest lands are located at Mirpur. defined as Phase I work to embank the western part of The present population of the Dhaka Statistical the city. This has been completed. Flood protection Metropolitan Area is more than 10 million. The last infrastructure for the eastern zone is under active con- decadal growth rate was about 70 percent, though the sideration by the government of Bangladesh. population growth rate was even higher, more than 100 This report on the flood vulnerability of Dhaka city percent in the previous decade (1981 to 1991). Popu- and flood management is organized into five main sec- lation statistics of Dhaka city show that the annual growth tions. An introductory section briefly explains the origin, rate was 2.9 percent from 1951 to 1961, 10.2 percent population, flood proneness, and mitigation measures from 1961 to 1974, and 8.1 percent from 1974 to 1981. of Dhaka City. The second section provides historical By virtue of being surrounded by the distributaries development, emphasizing physical growth of the city of several major rivers, the city has been subjected to and demographic features. The third section of the report periodic flooding since its early days. Major floods in highlights flood proneness of the city, including detailed the Greater Dhaka area have occurred in 1954, 1955, impact assessments from the 1988 and 1998 floods. The 1970, 1974, 1980, 1987, 1988, and 1998 due to spillover fourth section provides information on various measures from surrounding rivers. Among these, the 1988 and completed and ongoing to mitigate flood hazards. The 1998 floods were catastrophic. In the 1988 flood, it was concluding section of the report attempts to conduct 121 122 Building Safer Cities: The Future of Disaster Risk an analysis of integrated flood management plans for the basis for settlements. The population of Dhaka city Dhaka city, to protect it from natural disasters with min- at that time is unknown. imum environmental disruption. Mughal Period (1604­1764) Physical and Demographic Development The "new Dhaka" inaugurated by Islam Khan through the establishment of a fort, Chandnighat and the Chauk, The physical features, topography, and demographic fea- experienced growth under the subsequent Mughal Sub- tures of Dhaka City have always influenced its expan- ahdars until 1717, when the provincial capital was sion. Political importance and trade played significant officially shifted to Murshidabad. Dhaka enjoyed roles in the city's expansion during the Mughal and status as a provincial capital for slightly more than a British regime. This section briefly describes the growth century. During this period, administrative and defense and expansion of the city in the scale of time under five needs, coupled with flourishing commercial activities, major periods: pre-Mughal (before 1604), Mughal led to Dhaka's growth from a town to a metropolis. The (1604­1764),British(1764­1947),Pakistan(1947­1971), accounts left by foreign travelers, the extant of the Mughal and Bangladesh (after 1971). ruins, and the names of the localities that still survive show the extent of Mughal Dhaka (Karim 1964). It appears from various documents and Muslim sculp- Pre-Mughal Period (before 1604) tures that Mughal Dhaka encompassed the "old Dhaka" Growth and expansion of Dhaka city in the pre-Mughal within itself. The noteworthy feature of the city was its period is obscure. The near capital city of Vikrampur growth to the northern Phulbaria area. In this period, was in the limelight from the 10th to 13th centuries. the expansion to the west and the north was signifi- The Muslim occupation of southeastern Bengal can be cant. With the fort in the center, the expansion to the placed in the late 13th and early 14th centuries, when west followed the riverbank, and the city spread Sonargaon rose to prominence. This was first as a mint northward to Phulbaria on the fringe of the Ramna area. town and an administrative headquarters and subse- The peelkhana (elephant stable) was established at the quently for a short time as the capital under the early western end. Residential quarters for officials, govern- Iliyas Shahi Sultans. Sonargaon enjoyed the position of ment functionaries, and merchants grew in the area a metropolis in the region in the pre-Mughal period. between the fort and the peelkhana to the west and the It is evident from the various writings on Dhaka that fort and Phulbaria to the north. In this growth of Mughal the areas to the east, northeast, and southeast of Babur Dhaka, the general characteristics of a Mughal city were Bazar up to the Dulai river on the left bank (northern noticeable. The areas to the south and southwest of the bank) of the Buriganga formed the old town. The con- fort up to the riverbank grew mainly as commercial glomeration of Hindu-named localities in this part of areas, while to the north and northeast, residential areas the city bears testimony to the predominance of Hindu sprouted. craftsmen and professionals in the population of old The northern limit of the city extended to the gate- Dhaka city at that time. The Dulai River possibly formed way built by Mir Jumla (1660­63), near the present- the northeastern boundary of the old city, though it is day mausoleum of three leaders, at the southeastern difficult to determine the western limit of the pre-Mughal corner of the Suhrawardy Udyan. Mir Jumla's name is "old city." Considering testimony to the existence of a also associated with the construction of two roads con- mosque at that time, however, it can be assumed the city necting Dhaka with a network of forts built for the limits went beyond Babur Bazar on the western side defense of the capital city. A road headed north to a fort (figure 9.1). It is quite likely that, following the course of at Tongi-Jamalpur and another toward the east con- the Buriganga, settlements grew on the southern, west- nected Dhaka with Fatullah, where two other defen- ern, and northwestern parts of the city. These, of course, sive forts were constructed. These two roads influenced were sporadic growths with the riverbank determining the growth of the city in these directions. Flood Management and Vulnerability of Dhaka City 123 Figure 9.1 Demarcation between Pre-Mughal and Mughal Dhaka In the available early records of the East India Com- (Karim 1964). The physical size of Dhaka was about pany (1786 and 1800), the boundary of the city is men- 50 square kilometers; the population about 0.9 million. tioned as: Buriganga in the south, Tongi in the north, Jafarabad-Mirpur in the west, and Postogola in the British Period (1764­1947) east. The expansion of the city in the Mughal period was dictated by nature due to the highlands. After the acquisition of the Diwani in 1765 by the East With the establishment of the Mughal provincial cap- IndiaCompanyandtheshiftofBengal'scapitaltoCalcutta, ital at Dhaka, the city entered a glorious era and became Dhaka lost its political importance. Gradually the admin- the chief emporium for products from eastern Bengal. istrative and commercial importance of the city dwin- The commercial headquarters was established in the dled, and by 1828 the city was reduced to a mere district Mirpur area, Shah Bandar. Due to its commercial impor- headquarters, though it retained its position as a provin- tance, Dhaka attracted numerous traders--Portuguese, cial Circuit Court of Appeal. By 1840, this decline had Dutch, English, French, and Armenians--who estab- reached its nadir, and most of the former Mughal city lished trading houses in Dhaka in the 17th century. had been deserted or fallen victim to the encroaching Factories were also established in the Tejgaon area, which jungle (Ahmed 1986). The decline affected Dhaka seri- continued to enjoy commercial importance during the ously, and during this period Dhaka also suffered phys- next century (figure 9.2). The road built by Mir Jumla ical shrinkage. The jungle-beset city was shown in a formed an axis with European settlements on either topographical map prepared in 1859 as covering an area side, north of the Kawran Bazar and Amber Bridege only a little over three square miles (figure 9.2). 124 Building Safer Cities: The Future of Disaster Risk Figure 9.2 The buildup area of the Mughal capital Flood Management and Vulnerability of Dhaka City 125 The second half of the 19th century marked the begin- Railway crossing to the Purana Paltan was developed ning of the physical renewal of the city. In 1857, India as the open area of the city with the stadium forming came under the direct rule of the British crown and saw the nerve center of sporting activities. Jinnah Avenue some development of utility services. In 1905, Dhaka (now Bangabandhu Avenue) was laid to form the main became the capital of the new province of East Bengal thoroughfare along the western side of this expansive and Assam. Building the new town started beyond the open area. railroad in Ramna. The only locality developed as a fully To cater to the ever-increasing residential needs of the planned residential area was Wari. In 1885, Frederick new capital, the Dhanmondi area, adorned with paddy Wyer, the Collector of Dhaka, began developing the area fields in the early 1950s, was developed as a residential with "broad roads and proper drains." Wari became an area after 1955. The Mirpur Road formed an axis, and upper-middle class area considered "the sanatorium of the highlands on both sides of the road came to be Dacca." occupied, right up to Mohammadpur and Mirpur. In the The Siddheswari area to the northeast of the Race mid-1960s these two areas were developed by the gov- Course was cleared by the government in the early 20th ernment, mainly to accommodate the migrant Muslim century and the former jungle was developed as a res- population. The Tejgaon Airport and the Tejgaon Indus- idential area. At the same time, the wasteland around trial area came under governmental schemes in the early the Dhakeswari shrine was cleared by the local people 1950s. In the second half of the 1960s, the decision to who felt encouraged to occupy the areas around the have a capital for East Pakistan at Dhaka led to the devel- newly developed area of Ramna. Thus the "new Dhaka" opment of the area to the west of Tejgaon farm and the of the present century had its birth at the hands of the airport (now known as Sher-e-Bangla Nagar). British rulers. With the creation of the Dhaka Improvement Trust The impetus for growth created by the 1905 partition (DIT) in 1956 (transformed into the Rajdhani Unnayan of Bengal was seriously jolted by the annulment of the Kartripakkha in 1987), greater interest and care were partition in 1911, when Dhaka reverted back to the undertaken in road construction and city planning. The status of a district town. The establishment of the Uni- DIT developed the Gulshan Model Town in 1961, Banani versity of Dhaka, which came to occupy many of the in 1964, Uttara in 1965, and Baridhara in 1972 (though buildings of the Ramna area, was the only important first conceived in 1962). The Dilkusha Gardens adja- event in Dhaka's history until 1947, when Dhaka again cent to Motijheel were eventually engulfed by the ever- attained the status of the provincial capital of the east- growing commercial needs. ern part of Pakistan, initially called East Bengal and later In the mid-1960s the main railway line was shifted named East Pakistan. and directed eastward, after Tejgaon and before Kawran- bazar, before rejoining the old track near Swamibagh- Zatrabari cutting through Rajarbagh, Basabo and Pakistan (1947­1971) Kamalapur. The Dhaka Railway Station was moved from In 1947, India become independent of British rule and Phulbaria to Kamalapur. This eliminated the landmark Pakistan was created. Dhaka restarted its life as the that had long stood between the "old Dhaka" of the capital of East Pakistan. The needs of the officials engaged Mughals and the "new Dhaka" of the English. The rapid in administration, the business community, and the res- growth and development of the area between the old idents grew out of the sudden onrush of people to Dhaka. railway track and Kawranbazar necessitated this change. This contributed to the growth of the city in its new The loop through the heart of Ramna had to be aban- role as the provincial capital. doned. Since then, the old track has been developed In 1954, the Motijheel area, once desolate and lying into a broad road connecting Kawranbazar with Phul- on the fringe of marshes and swamps where the Nawabs baria through Plassey and Nilkhet to the northwest had built a garden house, was earmarked as a com- and Swamibagh-Zatrabari through Wari to the north mercial area. By that time, the area north of Nawabpur and Narinda to the southeast. 126 Building Safer Cities: The Future of Disaster Risk Table 9.1 Area and population of Dhaka City, 1600­2001 Approximate Period and year area (sq. km.) Source Population Source 1600 Pre-Mughal Period 1 Islam, 1974 Unknown 1700 Mughal Capital 50 Taylor, 1840 900,00 Taylor, 1840 1800 British Town 8 Islam, 1974 200,000 Taylor, 1840 1867 British Town 8 Islam, 1974 51,000 Census of Bengal, 1901 1911 British Town -- -- 125,733 Census of Bengal, 1911 1947 Capital of East Pakistan 12 Islam, 1974 250,000 Census of Pakistan, 1951 1951 Pakistan Period -- -- 335,928 Census of Pakistan, 1951 1961 Pakistan Period 28 Census of Pakistan 550,143 Census of Pakistan, 1961 1971 Capital of Bangladesh 40 Census of Bangladesh, 1974 1,500,000 Census of Bangladesh, 1974 1974 Capital of Bangladesh 40 Census of Bangladesh, 1974 1,600,000 Census of Bangladesh, 1974 1981 Dhaka Municipality 62.4 Census of Bangladesh, 1981 2,475,710 Census of Bangladesh, 1981 1981 Dhaka SMA 155.4 Census of Bangladesh, 1981 3,440,147 Census of Bangladesh, 1981 1991 Dhaka SMA -- 6,950,920 Census of Bangladesh, 1991 2001 Dhaka SMA -- 9,912,908 Census of Bangladesh, 2001 Bangladesh (1971 onward) own in a haphazard manner, and the topography of the area dictated the terms and direction of the growth. The creation of the independent state of Bangladesh in Since Dhaka became the capital of an independent coun- 1971 bestowed glory and prestige on Dhaka, now cap- try, the pressure on it has been enormous. The perma- ital of a sovereign country. This additional factor led to nent inhabitants of the city have registered a steady Dhaka's phenomenal growth since 1971. The low- growth. In addition to this growth, there was a very large lying areas on the eastern side, such as Jurain, Goran, floating population, the pressure of which has resulted Badda, Khilgaon, Rampura, and Kamrangir Char, Shya- in the growth of slums on any available vacant land. The mali, Kalayanapur on the western side came under occu- recently built high-rise buildings, both in the commercial pation. Dhaka's growth picked up at a tremendous and residential sectors, occupy the city's highlands and pace and private initiatives played the dominant role. demonstrate ever-increasing pressure on Dhaka as it The growth of the city followed the pattern set by the builds upward, an inevitable and common phenome- Mughal founders. The city was delineated the maxi- non in all modern cities facing population growth. Since mum limit to Tongi in the north and Mirpur in the north- the 1990s, Dhaka has been on the verge of change in its west. The southeastern limit reached Postogola. The urban character, with vertical growth replacing horizon- riverine surroundings with water-channels, marshes, tal expansion (Chowdhury and Faruqui 1989). Over the and lowlands form the western, southern, and eastern years, most of the low-lying areas of western Dhaka have boundaries of the city. With increased population been filled in to meet the city's residential and commercial pressure, the highlands spreading northward were occu- demands. The eastern side of the city is being filled in pied and built up. The intervening ditches, swamps, by private intervention. Table 9.1 provides built-up areas and marshes were filled in, not in any planned manner, and populations from the pre-Mughal period to 2001. but as exigencies arose and private initiatives dominated the process. Development under the aegis of the Dhaka Improvement Trust dictated nature, rather than allowing it to direct planned growth. In selecting the Major Floods in Dhaka sites for the Model Towns of Gulshan, Banani, Barid- hara, and Uttara, the method of selecting the high- A number of severe floods have struck Dhaka since its lands on the main Dhaka-Tongi axis road is clearly early days, and its vulnerability is reflected in the Buri- discernible. No serious effort at reclaiming land under ganga River's flood embankments, first built in 1864. a well-planned scheme to give the city homogenous and Severe floods in Greater Dhaka City are mainly caused cohesive growth is visible. Dhaka has grown on its by spillover from surrounding rivers flowing to and from Flood Management and Vulnerability of Dhaka City 127 the major rivers of the country, as well as internal water neither an operating policy nor person assigned to oper- logging. In recent history, Greater Dhaka city experienced ate the Rampura Regulator that controls the drainage major floods in 1954, 1955, 1970, 1974, 1980, 1987, of 40 percent of the protected area under Phase I of 1988, and 1998 due to the overflow of surrounding rivers. DIFPP. Among these, the 1988 and 1998 floods were catastrophic. Flooding due to rainfall is also a severe problem for certain city areas that may be inundated for several days, Balu River Flood in Dhaka East mainly due to drainage congestion. The water depth in In the 1998 flood, almost all of Dhaka East was flooded some areas may be as high as 40­60 centimeters, by spillover from the Balu River. The flood started around which creates large infrastructure problems for the July 22 and continued for some 65 days. The waters city, economic losses in production, and damage to exist- peaked on September 12. Residential areas such as ing property and goods. Impacts of the riverine floods Basabo, Mugdapara, Uttar Badda, and Joar Sahara are more severe and disrupt economic activities and were the worst affected. The ground floors of most build- livelihoods of people dependent upon urban activities. ings were inundated. This section provides causes and characteristics of floods with brief descriptions of their impact from the last two major floods, in 1998 and 1988. Balu River Flood in Dhaka West As it is protected by the flood control project, it was 1998 Flood expected that Dhaka West would remain flood-free, The main reason for the 1998 flood was excessive rain- even though the areas adjacent to Mymensingh Road, fall over the catchments area of the Ganges-Brahmaputra- Progati Sarani, DIT Road, and Biswa Road were sub- Meghna (GBM) river basin. Three different flood waves merged during the flood. Mahakhali, Gulshan, Banani, passed through the GBM river basin, and the last one Badda, Baridhara, etc., faced worse flooding problems. was synchronized with the peak flow of the Ganges The floodwaters remained in the western areas for about and the Brahmaputra. In addition, the impact of the 30 days. lunar cycle and its resulting high tide caused floodwa- The intrusion of floodwater from Balu River to Dhaka ters to recede slowly, prolonging flooding in the coun- East through unblocked culverts and open regulators try and the city for two months. was considered to be the general cause for river flooding The main causes of flooding inside the protected area in Dhaka West. It was revealed that three drainage struc- were hydraulic leakage, failure to operate the regula- tures, FS8 (Begunbari Khal Rgulator), FS5 (Shajadpur tors, and lack of timely pumping of accumulated water bridge on Pragati Sarani), and FS4 (Khilkhet pipe culvert upstream from the Rampura Regulator. Due to com- at Nijunja), served as a flood-carrying channel (figure 9.3). pletion of 75 percent of Phase I work of the Dhaka Structure FS4 remained open during the entire flood Integrated Flood Protection Project (DIFPP), it was and caused flooding in Nijunja adjacent to the airport assumed that the Gulshan, Banani, Baridhara, and and cantonment areas. Structure FS8 and FS5 were Tejgaon areas would not be flooded. However, near the closed after the intrusion of substantial floodwater from Cantonment Railway Station, there are 4 or more drainage the Balu River in Begunbari khal, Gulshan-Banani- pipes of about 4 feet in diameter that connect the Nikunga Mohakhali Khal, and areas adjacent to the DIT road area with the floodplain on the eastern side. An appar- were submerged. The channel at structure FS5 was ent lack of coordination between the Bangladesh closed by an earthen bund. Structures FS9, FS10, and Water Development Board (BWDB) and Dhaka Water FS11 on Biswa road were responsible for partial flood- Supply and Sewerage Authority (DWASA) to prevent ing in Rajarbag, Gopibag, and Fakirapul before they flooding was found. Although DWASA has responsi- were closed. Floodwaters in those areas were pumped bility for ensuring proper drainage, BWDB is in charge out by 30 pumps installed at FS12 on Segunbagicha of operating the regulators and gates. In fact, there was khal at a crossing with Biswa road. 128 Building Safer Cities: The Future of Disaster Risk Figure 9.3 Flood and drainage infrastructure of Dhaka GAZIPUR KHAL TONGI S2 S3 S FS1 UTTARA FS2 FS3 KHILKHET MIRPUR P FS4 REV S4 RI DHAKA WEST DHAKA EAST BALU S5 FS5 KALLYANPUR P FS6 FS7 S6 RAMPUR FS8 S7 KHILGAON FS9 S8 FS10 MOTIJHEEL SINGAIR FS11 S9 JATRABARI S10 P FS12 S11 FS13 DND BURIGANGA RIVER P N LEGEND Existing Flood Protection Embankement Rail Road Road River Sluicegate/Culvert Existing Pumping Station Under Construction Pumping Station Flood Management and Vulnerability of Dhaka City 129 Buriganga River Flood in Dhaka West flood displaced or dislocated 94 percent of families in the most severely affected areas, 52 percent of families The western part of the Dhaka Flood Protection Embank- in severely affected areas, and 50 percent in moder- ment ends in Lalbag Kellarmore at Swasan Ghat. It has ately affected areas. It was also estimated that the total a 2.2-kilometer embankment under construction from flood-affected population would likely be 4.55 million. Kellermore to the Buriganga Bridge, and it was not closed Table 9.2 provides details of affected households and during flooding, which allowed floodwater to flow populations. into the western part of the city. Flood fighting by local Effect on Water Supply and Production: It was people creating sandbag barriers saved Old Dhaka from found that 44 deep tube wells were affected by flood- flooding, especially in Lalbag Kellarmore. The differ- water; production of 13 of these was suspended. The ence in water levels between the outside and inside at estimated loss in water production due to suspension Kellamore point was about 1.5 meters. It was found that was 45 million liters per day. the flood protection work started immediately after The remaining tube wells were kept operational by the flood recession. adopting protection measures including the erection of a protection wall around the pump house and rais- Flooding Due to Excessive Rainfall ing housing pipes and electrical appliances above flood- water. It was also found that 5 of the 13 suspended During the 1998 flood, excessive rainfall in Dhaka caused tube wells were badly damaged and required replace- short-duration flooding in the areas of Shantinagar, ment. The estimated cost to rehabilitate the water supply Nayapaltan, Rajarbag, Dhanmodi, Azimpur, and system is about 127 million taka, detailed in table 9.3. Green Road. The runoff generated by rainfall could not flow to the surrounding rivers since the river stage was higher than the inside flow; therefore, the accumu- Table 9.2 Flood-affected people in Dhaka City by severity lated runoff in low-lying areas remained stagnant until of the flood, 1998 the river stage receded. Extensive water logging occurred Total affected in Dhaka West during the flood due to a higher river Affected Total affected population Severity of flood wards (no.) households (millions) water stage in the surrounding rivers. Most severe 22 203,000 1.20 Severe 15 150,000 0.90 Impacts Moderate 43 409,000 2.45 Total affected 80 762,000 4.55 It is evident from various studies that damage to infra- Source: Hye 1999. structure including roads, water supply, and housing was severe. It was estimated that 384 kilometers of paved roads went under floodwaters, of which Gulshan Thana Table 9.3 Cost of rehabilitation and replacement of accounted for a significant amount. Severe damage Dhaka Water Supply System (DWASA), March 1999 occurred in Sabujbag Thana, followed by Demra. Severe Rehabilitation and Estimated cost disruption of water supplies from deep tube wells (DTW) Item replacement needs (million taka) and suspended production occurred in the Cantonment 1 Repair and cleaning of 600 kilometer 15 Thana followed by Gulshan and Uttara. Considering the water distribution mains 2 Replacement of 5 deep tube wells 20 major impacts of floodwaters, it appears that Sabujbag 3 Rehabilitation of pump and electrical 42 and Gulshan Thanas were worst affected, followed by sub-station of 42 flood-affected DTWs Demra, Uttara, and Cantonment Thana. 4 Raising level of reconstruction of 10 walls of 42 pump houses Affected Population: Prof. S. A. Hye (1999) car- 5 Installation of 3 low-lift pumps 15 ried out a rapid appraisal of flood-affected people during 6 Reconstruction of 20 kilometer 25 the flood, dividing the flood-affected area into three cat- water mains egories: most severely affected area, severely affected 127 area, and moderately affected area. It was found that the Source: Ahmed 1999. 130 Building Safer Cities: The Future of Disaster Risk Housing Damage: The 1998 flood caused damage the 1988 flood was 14,000 and 400,000, respectively. to more than 262,000 shelter units, or 30 percent of the The damage was about Tk. 4 billion for residential build- 860,552 units in the Dhaka Metropolitan Area; the cost ings and more than Tk. 400 million for institutions. of damage was Tk. 2,311 million. Of these, 32 percent were permanent and semipermanent structures belong- Comparison of Flood Characteristics ing to wealthy or well-to-do households not dependent of 1988 and 1998 on assistance for repair and rehabilitation. About 36 percent of shelter units in the katcha-1 type, belonging to An analysis of water level hydrographs at gauging sta- lower-middle and poorer classes, suffered damage of Tk. tions in surrounding rivers and at Noonkhawa on the 283 million. Their owners had the ability to cope with Brahmaputra River revealed that in 1998, the first flood repairs but would face hardship. Nearly 32 percent of waves took approximately 6 days to reach Dhaka City units of Katch-2 and Jupri types, belonging to the poor from the India-Bangladesh border, while the flood peak and hard-core poor, suffered severe damage and required required approximately 4 days to reach Dhaka. Due to Tk. 195 million in repairs. The owners were too poor hydraulic reasons, the flood peak moves faster than the to mobilize funds on their own (Islam and Ali, 1999). flood trough. Figures 9.4 and 9.5 present hydrographs of surrounding rivers and rainfall in Dhaka City during 1988 and 1998. Table 9.4 presents flood characteristics 1988 Floods of the 1988 and 1998 floods in rivers surrounding Dhaka. In 1988, one of the most severe floods in recent history hit Dhaka and inundated 85 percent of the city. Floods Flood Mitigation Measures of this intensity hit the country approximately once in 70 years. Depths of inundation ranged from 0.3 to more The first flood protection embankment along the Buri- than 4.5 meters, and 60 percent of city dwellers were ganga River was constructed in 1864 to protect the river- affected ( JICA 1991, 1992). This unprecedented level bank from flooding and erosion and give a facelift to of flooding disrupted city life and air travel. Commu- the riverside. C.T. Buckland, the Commissioner of Dhaka, nication with Dhaka to the outside world was cut off launched a scheme to construct an embankment (known for about two weeks. Impacts and damage from the as the Buckland Bound), which was completed in 1988 flood were compiled and analyzed according to three phases in the 1880s. component 8 of the Flood Action Plan (FAP8). Plans for flood protection for Greater Dhaka have The entire eastern part of Dhaka and the entire been under study and consideration for many years, but low-lying area of the western part of Dhaka were the extreme flooding that occurred in 1987 and 1988 under floodwater. Parts of Mirpur, Tejgaon, Banani, Sher- brought into focus the urgent need to proceed with e-Banglanagar, Azimpur, and the Old Town were not immediate action. Subsequently, the Government of flooded. Bangladesh prepared an urgent flood protection and drainage plan, which included enclosing the greater Impacts Dhaka area with flood embankments, reinforced con- crete walls, and drainage/flood regulation structures Affected Population: According to available informa- such as sluices and pumping stations. tion on impacts of the 1988 flood on Dhaka, it was found that 2.2 million people were affected and the Construction of Flood Protection Embankment death toll was about 150. The figure for "affected pop- ulation" was found to be lower than that of the detailed Construction activities commenced with a "crash pro- analysis that was carried out by FAP8 based on data col- gram" in 1989, and most of the work defined as Phase lected through the flood damage sample survey (FAP8A). I has been completed. It provides flood protection facil- Affected Houses and Institutions: It is estimated ities to the western half of Dhaka and includes the that the number of institutions and houses affected in most highly urbanized areas, covering about 87 percent Flood Management and Vulnerability of Dhaka City 131 Figure 9.4 Water level hydrographs for Turag, Tongi, Buriganga, and Balu Rivers and rainfall in Dhaka during 1998 (a) 1998 Flood 9 0 8 7.97 80 7.54 datum 7.23 PWD 7 160 above (mm) meter in Rainfall 6 240 Level ater W 5 320 4 400 01/07 08/07 15/07 22/07 29/07 05/08 12/08 19/08 26/08 02/09 09/09 16/09 23/09 30/09 of the population. Important components of the flood for the eastern part of the city. This will be constructed protection measures are: under Phase II of the Dhaka Integrated Flood Protection · Approximately 30 kilometers of earthen embankment Project (DIFPP) to protect the area between Biswa Road along Tongi khal, Turag River, and Buriganga River. and the Balu River. The locations of regulators, sluice · Approximately 37 kilometers of raised road and gates, pump stations, embankments, and raised roads floodwall. are shown in figure 9.6. These flood control and drainage · A total of 11 regulators at the outfall of khals to the works have brought major changes in the flood regime of surrounding rivers along the embankment. Dhaka West, including major changes in land use. · One regulator and 12 sluice gates on the khals at the crossings with Biswa Road, DIT Road, Pragati Sarani, Construction of Storm Sewer and Pump Station Mymensingh Road, and railway line at Uttar Khan. · One pumping station at the outfall of Kallyanpur khal To alleviate the internal drainage problems of Dhaka, a to the Turag River, and another one at the outfall of storm-water drainage improvement plan was under- Dholai khal to the Buriganga River. These pump sta- taken by Dhaka WASA (JICA 1991). As a part of the plan, tions are for draining rainwater from parts of Dhaka many sections of the natural khals were replaced by con- West. crete box culverts. Converted khals include Dhanmondi · A special 10.53-kilometer embankment to sur- khal, Paribagh khal, Begunbari khal, Mahakhali khal, round the Zia International Airport. Segunbagicha khal, and Dholai khal. These khals are A rail-cum-road-embankment that will run along the no longer visible. The present storm-water drainage net- Balu River for a total length of 29 kilometers is proposed work under Dhaka WASA covers an area of approximately 132 Building Safer Cities: The Future of Disaster Risk Figure 9.5 Water level hydrographs for Turag, Tongi, Buriganga, and Balu Rivers and rainfall in Dhaka during 1988 (b) 1998 Flood 9 0 8 80 8.35 7.84 datum 7.58 7 7.09 160 above (mm) meter in 6 240 Rainfall Level ater W 5 320 4 400 01/07 08/07 15/07 22/07 29/07 05/08 12/08 19/08 26/08 02/09 09/09 16/09 23/09 30/09 Date Rainfall (mm) Buriganga (Millbarak) Turag (Mirpur) Tongi Khal (Tongi) Balu (Demra) 140 square kilometers. Important components of the also 65 small pumps with individual capacities of 0.142 drainage network are briefly summarized below. cubic meters per second, installed temporarily by Dhaka · 22 open canals with widths of 10 to 30 meters and WASA to drain storm water from various locations. a total length of approximately 65 kilometers. Moreover, DCC has constructed and maintains at · 185 kilometers of underground pipes with diame- least 130 kilometers of small-diameter underground ters ranging from 450 to 3000 millimeters. drains and approximately 1,200 kilometers of surface · 6.5 kilometers of box culvert with sizes ranging from drains that carry storm water to the main sewer lines. 2.5 meters by 3.4 meters to 6 meters by 4.1 meters. The Capital Development Authority (RAJUK) also · 2 storm-water pumping stations with capacities of constructs underground roadside drainage lines during 9.6 and 10 cubic meters per second at Narinda and the construction of new roads. Kallyanpur, respectively. · Recently, Dhaka City Corporation (DCC) constructed one storm-water pumping station with a capacity of Conclusions and Recommendations 22 cubic meters per second at the outfall of Dholai khal into River Buriganga. Dhaka WASA has taken Two severe floods hit Greater Dhaka City within a decade, over the operation and maintenance of the pump- causing enormous loss of life and livelihoods and damage ing stations. to property. Immediately after the 1988 flood, a number The Bangladesh Water Development Board is also con- of studies were carried out within the general frame- structing a pumping station at Goran Chadbari at the work of the Flood Action Plan (FAP) that specifically outfall of the Degun khal into the Turag River. There are addressed the issue of flooding in Dhaka City. The first Flood Management and Vulnerability of Dhaka City 133 Table 9.4 Flood characteristics of 1988 and 1998 floods in surrounding rivers of Dhaka City Parameters River Gauge station 1998 1988 Danger level in meters above PWD datum Buriganga Millbarak 6 6.1 Turag Mirpur 5.94 5.94 Tongi Khal Tongi 6.08 6.08 Balu Demra -- -- Date of crossing-danger-level at rising stage Buriganga Millbarak 26/07/98 29/08/88 Turag Mirpur 18/07/98 24/08/88 Tongi Khal Tongi 22/07/98 28/08/88 Balu Demra -- -- Number of days required by the flood front to arrive at Dhaka Buriganga Millbarak 7 7 City from India-Bangladesh border Turag Mirpur 6 7 Tongi Khal Tongi 6 6 Balu Demra -- 10 Number of days required by the highest flood peak to travel Buriganga Millbarak 4 6 from India-Bangladesh border to Dhaka City Turag Mirpur 4 6 Tongi Khal Tongi 5 6 Balu Demra -- 8 Height of peak flood level in meters above danger level Buriganga Millbarak 1.23 1.58 Turag Mirpur 2.03 2.41 Tongi Khal Tongi 1.46 1.76 Balu Demra -- -- Duration of flood in days above danger level Buriganga Millbarak 56 22 Turag Mirpur 69 30 Tongi Khal Tongi 65 25 Balu Demra -- -- phase of the Greater Dhaka Integrated Flood Protection taken place are Mugdapara, Manda, Basabo, Sabujbag, Project included embankments along the Turag and Khilgaon, Goran, and Rampura. Some of these areas the Buriganga Rivers to protect Dhaka West. Improve- were the worst affected during the 1998 flood. ments to the city's internal drainage system had also Evaluation of Phase I work revealed that the exist- been completed before the second severe flood in 1998. ing earthen embankment is unstable and cracking in Moreover, during the 1998 flood some protected areas large sections, though some parts have recently been went under water, indicating that current flood man- stabilized and other parts of the earthen embankment agement practices must be improved. have been converted to roads. Construction of embank- After implementation of the flood-control project in ments through low-lying areas without providing ade- Dhaka West, unplanned and uncontrolled expansion quate drainage facilities has caused internal flooding, of urban areas stretched rapidly toward the low-lying adversely affecting the residents in those areas. areas and floodplains adjacent to the flood-protection The eastern part of the city consists of low-lying flood- embankment and river. Residents of these lowlands plains that are submerged during the monsoon season. suffer from inundation due to accumulation of rain- They still remained unprotected. However, the growing water after heavy rainfall. Land development through population and land scarcity have forced people to settle land-filling processes in the low-lying areas is causing in these low-lying areas. Implementation of Phase II of a drastic reduction in water storage areas. Because of DIFPP will provide flood protection to this part of the the rapid population increase and scarcity of land in city. While designing and implementing the plan, envi- Dhaka West, unplanned expansion is also taking place ronmental impacts should be kept in mind and an ade- in Dhaka East at the same pace. It started on the eastern quate number of drainage facilities should be provided side of Biswa Road, DIT Road, Pragati Sarani, and for the proposed embankment. The hydrological data Mymensingh Road, and is gradually stretching toward should be analyzed during the design phase of the flood the Balu River. The areas where urbanization has already protection infrastructure. 134 Building Safer Cities: The Future of Disaster Risk Figure 9.6 Existing and proposed flood control and management infrastructure in Dhaka KHAL TONGI TONGI BRIDGE RIVER ZIA INTERNATIONAL GARUT AIR PORT DHAKA WEST DHAKA EAST ROAD. D.I.T GULSHAN MIRPUR BRIDGE ROAD A R. DEMRA MOTIJHEEL BISW YA HKAL DND ATIS LEGEND N Existing Embankement/Flood Wall Proposed Embankement Rail Road Existing Raised Road River Pump Station Sluicegate Flood Management and Vulnerability of Dhaka City 135 Dredging work in the Buriganga, Turag, and Balu Ahmed S.U. 1986. Dhaka­A Study in Urban History and Devel- Rivers should be regularly carried out to maintain nav- opment. Riverdale: Riverdale Company: 12. igability and reduce drainage congestion. In addition Chowdhury A.M. and S. Faruqui. 1991. "Physical Growth of to structural measures, nonstructural measures should Dhaka City." In S.U. Ahmed, ed., Dhaka Past Present Future. also be considered to reduce flood damage. These meas- Dhaka: Asiatic Society of Bangladesh. ures include flood zoning, flood forecasting and warn- Hye A. 1999. "Effect on Livelihood." In A. Nishat, M. Reazud- ing, flood proofing, flood insurance, and evacuation din, and others, eds., The 1998 Flood: Impact on Environment measures. Flood zoning could be one of the most of Dhaka City. Department of Environment in conjunction with effective measures to ensure that expensive investments IUCN Bangladesh. are not made in flood-vulnerable areas. It will also help Islam N. and K. Ali. 1999. "Housing Damage in Dhaka City to protect ecologically sensitive areas, natural drainage Due to the 1998 Flood." In A. Nishat, M. Reazuddin, and systems, and the surrounding wetlands that retain water. others, eds., The 1998 Flood: Impact on Environment of Dhaka Since the western part of Dhaka is already developed, City. Department of Environment in conjunction with IUCN little can be done in this area except protecting the lakes Bangladesh. and khals. The government has already issued a decree JICA (Japan International Cooperation Agency). 1987. "Study banning the filling in of any wetland for urban develop- on Storm Water Drainage System Improvement Project in ment. In exceptional cases, permission should be secured Dhaka City." Supporting Report, Local Government Division directly from the prime minister. The eastern part of the of Ministry of Local Government Rural Development and city is still largely a floodplain. Proper land development Cooperative, Dhaka. rules should be introduced without delay to minimize JICA (Japan International Cooperation Agency). 1991. "Master the loss from river flooding from the Balu River. Plan for Greater Dhaka Flood Protection Project." Flood Action It is anticipated that the city's future flood vulnerabil- Plan 8A, Main Report and Supporting Reports I and II, ity will be aggravated due to climate change. Experts are Flood Plan Coordination Organization (Presently WARPO), forecasting that floods with the magnitudes of those in Dhaka. 1988 and 1998 may occur more frequently. Thus, it is JICA (Japan International Cooperation Agency). 1992. imperative that a long-term flood-mitigation and climate- "Feasibility Study of Greater Dhaka Flood Protection change adaptation strategy be developed for the future Project." Flood Action Plan 8A, Interim and Main Reports, management of floods in and around Greater Dhaka City. Flood Plan Coordination Organization (Presently WARPO), Dhaka. Bibliography Karim A. 1964. The Mughal Capital. Dhaka: The Asiatic Society Ahmed F. 1999. "Impact on Water Supply Sanitation and Waste of Pakistan: 8­13. Management." In A. Nishat, M. Reazuddin, and others, eds., Nishat A., M. Reazuddin, and others, eds., The 1998 Flood: Impact The 1998 Flood: Impact on Environment of Dhaka City. Depart- on Environment of Dhaka City. Department of Environment in ment of Environment in conjunction with IUCN Bangladesh. conjunction with IUCN Bangladesh. Chapter 10 Flooding in the Pampean Region of Argentina: The Salado Basin Hilda Herzer This paper provides information to help in understanding by environmental artificialization and the cities' terri- the complexity of flooding processes. Three munici- torial expansion over time. This is worsened in some palities in the Pampean Region of Argentina's Río Salado cases by anthropic elements such as roads, railways, and basin were chosen for this analysis: Chascomús, Dolores, canals that hinder the already slow water drainage either and Junín. The study focuses on the impact of farming because of building defects or lack of maintenance; the and cattle-breeding practices and urban and social poli- construction of a series of canals and embankments cies in the development of disasters.1 A flooded area, (built according to periods of floods or droughts); the whether it is a city or a region, is the expression of rela- changes produced in the area's dominant type of pro- tions and conflicts among different areas and diverse duction; and changes in land use and land tenure. socioeconomic groups over time, within the context of From the standpoint that vulnerability plays an impor- natural, economic, social, and political processes. Under- tant role in social risk building, the quality and loca- standing of this complexity is sustained by the articu- tion of built property, land use, the state of infrastructure lation that correspondence between the "causality space" and services, lifestyles, and political authority and gov- (i.e., the place where causes of socio-natural hazards ernment (ethical, legal, and political frameworks within are produced) and the "impact space" (i.e., the places which a society develops) must be emphasized. where they generate an impact over population) is not always necessary.2 This idea has two implications: first is the impor- Why the Basin Floods tance of reevaluating the understanding of causality for populations and authorities; and second, the idea defies The Río Salado basin is part of the hydrographic system an institutional notion for environmental urban man- of the Río de la Plata basin.3 It covers 186,000 square agement that is territorially limited to the city itself. kilometers, more than half the surface of the Province Therefore, it demands the incorporation of the region of Buenos Aires. It includes 56 of its 134 municipali- or basin as the analysis, management, and action unit, ties, and, in socioeconomic terms, is one of Argentina's both from a research point of view and from its political- most important areas.4 Over centuries, the area was built administrative organization (Herzer and Gurevich 1996). and rebuilt by human activities that changed its origi- nal physical structure, among which we must mention hydraulic works that modified the basin's drainage. Examining the Process We will describe the Río Salado Basin's natural charac- Biophysical Characteristics of the Basin teristicsthatturnitintoaregionpronetoflooding.Among them, it is important to highlight the slope, the type of The basin comprises three main regions, each with its soil, and the shallowness of the river and its streams that own features: Salado-Vallimanca-Las Flores; Noroeste; make water drainage difficult during heavy rain. The Encadenadas del Oeste. The three regions are inter- characteristics of this natural system are superimposed connected by a large plain where drainage conditions 137 138 Building Safer Cities: The Future of Disaster Risk are deficient or absent.5 The Salado-Vallimanca-Las Table 10.1 Average rainfall for each region Flores region is the largest, covering 60 percent of the Average from basin, and the three municipalities analyzed are Region 1920­1985 1986 to 2001 located there. Salado Vallimanca 800 to 900 1000 to 1100 The Río Salado streams from the southwestern end Norte and Sur millimeters millimeters Noroeste 750 millimeters More than 1100 of the Province of Santa Fe at 75 meters above sea millimeters level. It crosses the Province of Buenos Aires from north- Lagunas 700 to 900 1100 millimeters west to southeast for 650 kilometers.6 Its regime is exclu- Encadenadas millimeters sively rainwater. The Río Salado springs from the del Oeste union of small lakes, and along its course many Source: Instituto Nacional de Tecnología Agropecuaria. lagoons and streams pour into it until it reaches its mouth in the Samborombón Bay. There, due to the influ- developed and disintegrated; therefore, the amount of ence of tides, flows invert when seawater enters. East- water that drains through the rivers is also small, less than ern winds also produce a flow stoppage, and water 10 percent.7 It is a low energy system and, consequently, accumulates in the outlet. In the northwestern region, the lack of capacity to evacuate water surpluses produces ancient dunes hinder water drainage to the east and generalized and prolonged floods. The ground water level north. Water accumulates on the surface during pro- increases during prolonged rainfall seasons, rising to longed rain or due to high phreatic levels. Higher rain- the surface and producing floods. Some of the basin's sec- fall during the last 20 years also affects the area. Since tors are affected by water salinity because the under- there is little alternative drainage, surplus water drains ground sheet of water is not deep enough and in some to the south, though there is a lot of standing water. cases it reaches ground level, affecting soil quality. During the last century, the runoff was modified in a large part of the basin due to the canalization of the The nature of the existing drainage pattern and the nature region's many watercourses. As a consequence, tribu- of rains show that the area suffers frequent, extended, and prolonged floods and inundation. For example, the 1980 taries were intercepted and deviated, while at the same and 1985 events lasted from four to five months all along time, the basin's drainage network was extended. Arti- the Río Salado and in 1993 the Río Salado Inferior was ficial canals were designed to evacuate surplus water affected during three to four months. The Vallimanca that accumulated upstream. However, the system's evac- stream also suffered prolonged floods lasting about three uation capacity is, at present, reduced or inefficient. months, although the streams in the southern region, with Since the basin is of an open type, its subsystems are sufficient slope, respond quickly. It is also important to consider the temporal development of flood events interrelated, and any change in the transport and accu- since, for example, the 1993 flood began in October but mulation conditions in one place necessarily produces all others were during autumn/winter, beginning in March changes elsewhere in the basin. This is important or April (Río Salado Basin Integral Master Plan--Plan when considering the effects of activities carried out Maestro Integral de la Cuenca del Río Salado). anywhere in the basin. It rains an average of 870 millimeters per year (period 1911­96) in the basin, though the rain regime is irreg- Territorial and Productive Expansion ular and rains are stronger in summer. Winters are dry due to strong western winds. The water balance shows In 1991, there were 1,300,000 inhabitants in the area, a surplus during the summer and a deficit during winter. 11 percent of the province's total population. According March is the dampest and rainiest month. Since the to the 2001 census, there had been a population increase 1980s, there has been a marked increase in mean rain- of only 50,000.8 This figure represents 10 percent of the falls, as shown in table 10.1, below. province's population. This increase has been very low, The Río Salado basin has a scant slope, the average approximately 0.5 percent annually from 1947 to 2001. being 1 centimeter per kilometer, thus limiting the The urban population represents about 80 percent drainage of surplus water. The drainage system is poorly of the total population; the majority live in the region's Flooding in the Pampean Region of Argentina: The Salado Basin 139 main cities. There are around 145 urban centers in the primary products and nonmanufactured foodstuffs, and region, which differ in size and importance, since their 57 percent to manufactured foodstuffs.9 populations vary from 500 to 92,000 inhabitants. The The province contributes largely to cereal and oleagi- most populated urban centers are Tandil, Azul, Olavarría, nous exports and plays a fundamental role as a cattle- and Junín, which, in 1991, had more than 50,000 inhab- breeding area.10 Cattle-breeding production decreased itants. Although these cities have not grown in numbers in real terms between 1980 and 1996, while agricul- of inhabitants, their territory has expanded. ture increased by 25 percent. The production of indus- The urban structure is characterized by intermedi- trial crops (mainly oleaginous) abruptly increased during ate settlements (20,000 to 200,000 inhabitants). They this period.11 Its participation in the farming and cattle- act as linking centers for less important settlements breeding gross product increased from 16 percent in located in their area of influence, within a distance of 1980 to 34 percent in 1996. Cereals increased from 28 150 kilometers. percent in 1980 to 31 percent in 1996. Milk produc- Many municipalities have lost population, mainly tion in 1980 was 18 percent higher than that of 1996, due to two factors: the flood and drought cycles that and poultry breeding increased more than 50 percent. strongly affect the productive sector and a productive structure that finds it difficult to generate new and stable labor sources in any economic sector. Economic Deregulation and its Effect on Flooding The whole area has a well-developed transport infra- structure, with an extended network of national and The transformation suffered from the development model provincial trunk roads, and rural roads. The national in Argentinean rural areas during the last decade, par- ones are important highways that join Buenos Aires with ticularly in the Pampean region, which ushered in a other provinces and only flood in extreme cases due to change of rules that became evident as a new farming inadequate transversal drainage or to their very low and cattle-breeding map. Cattle-breeding land was trans- level. Provincial roads may or may not be paved, and ferred to agricultural uses, and the farming and cattle- the latter suffer the biggest floods. The rural road net- breeding frontier expanded, incorporating semi-arid work, generally unpaved, floods most; its maintenance land through the use of new technologies, including is shared by the municipal government and farming and irrigation. The productive cycles were also modified, cattle-breeding producer associations. generating intensive tilling that brought about a total production increase. During the 1990s, the economic deregulation poli- Farming and Cattle Breeding Production cies in force changed Argentinean farmland (Teubal and Rodriguez 2001). One of the most important measures The provincial farming and cattle-breeding sector, par- was the derogation of export taxes. International ticularly in the Salado basin, fulfills a strategic role in repayments were tax-exempt and differential exchange the provincial and national economies. This is due to a rates that reduced prices received by producers were combination of several factors: large-scale, mechanized eliminated. Also, the export withholding tax was abol- farmland, generally adequate rainfall, a well-articulated ished, although it has recently been re-implemented.12 road network, proximity to agromanufacturing, and con- These policies enabled concentration in large firms at sumer and export centers. a detriment to medium and small producers. From the During the 1991­93 period, average agrofoodstuffs 1990s, "agribusiness" (production in large units with a exports were US$7.1 million. In 1998, exports reached strong investment in technology) began to develop. $15.2 million. This important increase was due to a price Implementation of the so-called convertibility plan, decrease in the main technological innovations (fertil- one of its pillars being the decree deregulating economic izers, agrochemicals, direct sowing, etc.) that led to their activities, implied for the farming and cattle-breeding generalized use and to an increase in yield per hectare. sector the disappearance of specific governmental organ- Out of the total increase, 43 percent corresponds to izations that controlled the operation of farming and 140 Building Safer Cities: The Future of Disaster Risk cattle-breeding markets: Junta Nacional de Granos Historical Evolution of Flooding (National Cereal Board), de Carnes (Meat Board), and others. Tilling, harvest, manufacturing, and commer- Toward the end of the 19th century and the beginning cialization quotas for certain products were elimi- of the 20th, a damp phase that produced floods in nated. From a taxation point of view, taxes on farming large sectors of the central Pampean region and the Río and cattle-breeding activities, mainly on exports, were Salado basin in the Province of Buenos Aires was reg- abolished. Road and railway networks, and port activ- istered. Floods in 1840, 1854, 1857, 1874, 1877, 1883, ities were privatized and deregulated. 1886, 1800, 1910, 1913, 1914, and 1919 are described The effects of these changes were the establishment by Moncaut and Posadas in their reports on flooding of agro-industrial complexes that favor capital concen- in the region. We must add the references made by Engi- tration, together with the advance of agriculture in areas neer Gandolfo about the flooding of Río Quinto, in that had historically supported cattle breeding. These which he described the course of the river's overflow in significant changes in land use, added to the absence 1915, 1919, and 1923, when it reached the borders of of any integral basin management, increased flooding the Province of Buenos Aires.14 risks for some sectors of society and certain cities. This Duringthedryphasethatdominatedtheperiod1930­ is more visible when an intense rain cycle occurs. 1957, the problems were the opposite. The semi-arid Cattle breeding did not suffer the same shock as agri- stripbecamemarginalforsummertilling.Reservoirsdried culture. This is due to price flattening in relation to up and Río Quinto reduced its reach to the La Amarga agriculture and to the absence of technological pack- swamps and did not surpass, between 1940 and the ages. On the other hand, the agricultural sector read- 1970s, the site of Justo Darat in the Province of San Luis. ily incorporated technological packages composed of Since 1970, a new damp phase has begun. Along a public and private technical assistance, seeds, fertiliz- large strip on the western side of the Province of Buenos ers, herbicides, pesticides, artificial irrigation, and satel- Aires, the annual average rainfall increased from 600­700 lite information. The new productive cycle extended millimeters to 1000­1100 millimeters. This increase, the productive surface area and increased planting. concentrated in the warm season, balanced the agro- From 1990 to 1999, there was a more intensified land nomic limitations for summer crops and began the use and in 1997, fodder areas were replaced with agri- expansion of the farming and cattle-breeding frontier culture, constituting the second expansion factor. During toward the west, gradually including traditionally semi- the last 12 years, agriculture has advanced consider- arid areas in eastern La Pampa, Córdoba, and San Luis. ably in this area due to macroeconomic changes, prices, During some years, the limit to that constant surface and export opportunities. and yield expansion was simply flooding produced by Before the 1990s modernization, the economic struc- large volumes of rainfall that accumulated and filtered ture of the Argentinean Pampean area was predomi- the soil, raising the underground sheets of water to levels nantly cattle breeding, followed by agriculture and milk near the surface.15 This favored the expansion of bodies production. Cattle breeding was carried out in the open of water that contributed to flooding during the last (with some winter reinforcement), and bovine cattle 23 years (the consequences were particularly severe in were predominant. Dairy farming was a secondary activ- 1980, 1991­93, and 2000­01).16 The harshness in each ity within cattle breeding, but it was important in areas area and in each case is related to the distribution and such as Chascomús. Presently, there is mixed land use intensity of rainfall and to the presence of works that in the region, where cereal production is combined with do not conform to a coherent system. cattle breeding. Cattle breeding occupies more than 70 percent of the total surface area in low-level areas with The Social Actors no agricultural use. In areas where the soil is appro- priate for agriculture, mixed systems are developed in A large number of social actors are involved in the basin's which both activities rotate with variable intensity accord- flooding problem. We will work with a simple typology: ing to environmental conditions and soil quality.13 economic, political, and community actors (Herzer and Flooding in the Pampean Region of Argentina: The Salado Basin 141 others 1993). It is important to understand that risk- governments have little capacity to solve problems caused sharing processes and disasters will always be the object by flooding, and their actions are mainly confined to of controversial interests. This is so because these emergency situations through relief distribution, build- processes are built from agreements and oppositions of ing defenses and pumping posts, and assisting evac- a variety of social actors and from different rationales, uees. Among the actions that are normally in the hands interests, and logic. of local governments, we may consider the implemen- Among the economic actors, we find farming and tation of public and private building regulatory instru- cattle-breeding producers, mainly represented at the ments. Unfortunately these norms do not exist in some Sociedad Rural (Rural Society), and the commercial and cases; in others, they are not enforced. In the majority industrial associations affected by flooding.17 The of cases, there is no control over their application. For main demand of the commercial and industrial associ- example, in Chascomús, the ordinance approved after ations, organized into local chambers, is to be included the 1985 flood forbidding construction on plots of land in the emergency benefits during floods. under 9.40 meters in elevation was never enforced. The farming and cattle-breeding organizations18 are The Civil Defense Municipal Board (Junta Municipal a permanent reference for local authorities; they con- de Defensa Civil) is in charge of coordinating emergency stitute the local farming and cattle-breeding emer- tasks and arranging emergency relief at all government gency committees19 and certify the producers' claims levels. It is integrated by voluntary organizations and to obtain the corresponding benefits, with more action municipal government offices qualified to act in emer- capability to channel demands and claims for solu- gencies. Actions at the provincial level are coordinated tions at various government levels.20 Their members, by the Civil Defense Provincial Board (Junta Provincial as landowners, participate in associations in charge of de Defensa Civil), which collaborates with municipal rural road upkeep and maintenance, necessary to offices and qualified provincial offices "as much as pos- move production out of the fields. sible" through the provision of required elements. In the cases studied, there was a fluid, although not Last, coordination at the national level is the charge alwaysagood,interactionwithlocalgovernments,depend- of the Emergencies Federal System (Sistema Federal de ingonthemunicipality'sabilitytorespondtotheirclaims. Emergencias) (SIFEM), in which the Emergencies Cab- These actors have their own strategies to face emergency inet (Gabinete de Emergencias) is constituted.23 This situations. During periods of flood risk, canals are often organization was created in 1999, within the jurisdic- built on private land without consultation with water tion of the President's Cabinet Chief. It recently began to authorities.21 These works are not so much to safeguard depend upon the Internal Affairs Minister. The armed thepopulation'swellbeingastodrainwaterfromtheland. forces, mainly the army, transport people and goods when The participation of communities is restricted to times they are summoned by civil defense. In this particular of flooding. In the municipalities analyzed, commu- case, they helped restore road services by setting up a nity actions varied from passiveness to mobilization bridge to replace one that was destroyed in the flood. such as road blockages. They also met with authorities From another point of view, we can mention the set of to demand answers to their requests and insist upon institutions linked to water resources that depend upon the execution of flood prevention works that did not the National Ministry for Public Works. The Regional fall under municipal jurisdiction.22 In all cases, it is Directorate for Provincial Hydraulics (Dirección Regional important to stress the assistance provided by neigh- de Hidráulica Provincial) should control all public and bors and their organizations during floods, even if it private works in the region. It should be noted here that only means providing shelter in neighborhood associ- there are 29 national-level and 19 provincial-level organ- ations and participating in the building and vigilance izations involved in the basin (with an exception made of defenses. Often, neighbors themselves build provi- for the basin committees), plus the municipal ones. A sional defenses to protect their homes. common trait that affects all levels is that their actions These actors relate with government authorities at are limited due to a lack of funding following the three levels: municipal, provincial, and national. Local recent economic crisis. 142 Building Safer Cities: The Future of Disaster Risk Local Government Management in Three Cases limit their suffering. On the other hand, they hope that the local government will solve the situation by build- Chascomús ing embankments. For years, every time there has been a risk of lake over- flow, stone and earth embankments are raised to pro- Dolores tect the city. Also, public pumping equipment that operates around the clock is installed. Pumping is alter- Water drains from the municipalities of Las Flores, Azul, nately done in different places, extracting water from part of Ayacucho, and part of Tandil toward Dolores, saturated underground sheets; this is the only avail- where it accumulates because of the system's scarce able resource to compress underground sheets. capacity. The building of embankments along National Although the lake's contour can be controlled with Road number 2 affected the producers on the western some success, the few rain drains that exist in the city side, which does not drain. The waterfront extends 25 center reverse their flow, flooding streets and homes. to 30 kilometers and must drain through a 1-kilometer This has happened in the past, and it still occurs. canal, a bottleneck that stops water from running through. The existing municipal regulations are not enforced.24 The same thing happens on the eastern fields, which meet The defense system is always transitory. Each emergency the sand dune dikes produced by the building of Provin- is unique. With each new risk, the municipality sum- cial Road number 11. In this case, important landowners mons the mass media to alert citizens about the pre- have not been allowed to open the dunes because this cautions to be taken in the event of floods or a great would jeopardize their economic interests.25 Likewise, storm, and it demands that provincial authorities carry since 1920, they have been against the direct drainage out the works included in the Salado Basin Integral of Canal A on a straight line toward the sea. Presently, Master Plan. this is done by a winding, slowly draining canal that Repeated flooding, which requires the consolidation affects 30,000 hectares. This situation explains some and expansion of the defense system and pumping plants, of the conflicts of interest in the area. has produced mistrust and uncertainty in the local The Dolores cattle breeders do not have great hope population. Since the process repeats itself continuously, that the Master Plan's works will be executed, or that there should be a permanent risk-management policy they will produce results. This is so for various rea- to repair the existing fracture, evident because of the sons; the first and most fundamental is that, with the impact of flooding, between the central western zone present economic situation, it is difficult to believe that and the promising eastern and southern zones. US$2 billion will be spent on these works. The second It also seems necessary to think about a series of meas- is because Dolores gathers water from the higher basin-- ures to adapt municipal regulations and control meas- its northern and central portion--and the only work ures to hinder the development of new urbanization in expected to be carried out in Dolores is the enlargement areas lacking infrastructure and, in particular, areas that of Canal A, increasing the drainage capacity of Primary modify the basin's drainage. Both matters are closely Canals 9 and 1, and improving urban defenses. Other linked. It is of little use to concentrate on the consoli- minor but necessary works are not being considered. dation of the city's defenses if the city continues to expand Third, and perhaps the most interesting, is the lack of into lower areas. political strength of the provincial Hydraulics Direc- The interviews conducted revealed the general opin- torate (Dirección de Hidráulica), since it does not con- ion that, because Chascomús is located on the Salado's trol the canalization works carried out by producers in lower basin, it is subject to the consequences of actions their fields and by neighbors in the city. These are ille- produced upstream (higher Salado). In this sense, the gal works inasmuch as it is forbidden to alter the water's citizens consider their town as the region's "sewer." This natural flow. Consequently, coordination is difficult perception produces a passive attitude in local society, when public interests are subordinated to private ones, inasmuch as only works carried out upstream will and what is done on one site affects other social groups. Flooding in the Pampean Region of Argentina: The Salado Basin 143 The Dolores case typifies the lack of general coordi- works was the previous secretary. The experience gath- nation in the basin both public and private, as well as ered during 19 years in office has enabled him to col- rural and urban, and the consequences that derive from laborate with the farming and cattle-breeding sectors, it. The handling of water seems to be nearly out of con- although this has been limited by a lack of resources. trol, and in this sense it brings to the forefront the absence Junín, as Dolores and Chascomús, shows a certain of the state. Within this panorama, the roles played by complexity regarding its environmental urban-rural the Provincial Hydraulics Directorate and the local gov- problems that puts forward the need to integrate rep- ernment in the siting of housing are not clear, since resentative public and private actors to address them. other municipal offices must bear unbudgeted expenses As in Dolores, water seems to be out of management when providing assistance during emergencies. Fur- and control and, in this sense, it once again highlights thermore, the lack of social organizations, other than the state's absence. the traditional ones--e.g., Rural Society and the Some of the flooding problems require coordination Rotary Club--makes it difficult for community inter- among national, provincial, and municipal actors to ests to be asserted. enable the channeling of water without the political and economic conflicts now at stake. The possible execu- tion of the master plan is no guarantee that the rural- Junín urban problems will be solved or that neighborhoods prone to flooding will be safe. In this last case, not only One of the outstanding questions in the Junín case is is a new urban code necessary, but also an organization the interest in concealing recurrent flooding. Two to control its application. years ago, the municipality's strategic plan only mar- ginally mentioned flooding in the environmental urban development subprogram. Nor does the prelim- Water Management: Elements of Organization inary study for the plan mention flooding. Various critical comments arise from the interviews In a basin with scant or no slope, water drainage will conducted. One of them refers to the Dirección Hidráulica impact other communities. In this sense, there are few Provincial's lack of management and control (it is located plans to coordinate public actions. Moreover, the com- in the 9 de Julio municipality) over the great number munities are in no position to coordinate private actions of private works that interfere with water drainage, alter- carried out in private fields. The whole set of canals built ing the water's course and affecting those who live down- on private property26 has never been counted or regis- stream. There is an urgent need to appoint an authority tered by cartographers. Nor has the task been undertaken to regulate community water use. During the last few by the state. years, the municipality also asked the province to There is no coordination within levels of government, study the possible regulation of lakes. The study was and there are many levels with shared and overlapping completed, and its findings were incorporated into a responsibilitiesinthebasin27.Theprovincialgovernment master plan for water management, but the plan was accusesthemunicipalauthoritiesofindiscriminateaction never executed. with no coordination within the province when building During emergencies, the municipality summons inter- public works and canals. Municipal governments com- mediateassociations--rural,commercialandindustrial-- plain that they lack support from provincial authorities and neighborhood associations to respond. For example, other than what is provided during an emergency. a multisectoral commission was recently formed to Between 1998 and 2000, the Salado Basin Integral address poverty and unemployment. But such integrated Master Plan was developed with IDB assistance, diag- planning has no executive continuity the rest of the time. nosingtheneedtodevelopprojects,buildaseriesofcanals, It is important to highlight that the Junín government and carry out a set of defense works--embankments-- has been in office since 1983, when the country returned to limit the flow of water into urban areas. Beyond shar- to democracy. The present undersecretary for public ingtheproposalsestablishedbytheMasterPlan,noaction 144 Building Safer Cities: The Future of Disaster Risk has been taken.28 Meanwhile, agricultural producers From the local management perspective, munici- proposed that the nation and the province implement palities lack the strength to establish regulations and theinfrastructureplan;municipalitiesshouldbeincharge guarantee enforcement, remaining subordinated to the of secondary canals and producers of tertiary canals. All role of spectators of speculative economic processes that parties also agreed that a foundation should be created to induce illegal use of urban land. When disaster arises, build other canals, but this has not yet been acted upon. they are only able to react to the crisis. Municipalities should not only produce urban reg- ulations to control risk areas, they should incorporate General Conclusions from the Three Cases participative management tools to enable a periodic monitoring of interventions in the territory. Neighbor- Response to disasters, together with prevention programs hood community actors could play this role. and their management, are marked with the indelible The main problems faced by the municipalities during seal of political and economic relations built up over times of flooding were: time. In this sense, these programs are not alien to a · Rivalries with municipalities in the higher basin. country's political economy; on the contrary, they tend · Lack of financial resources. to reproduce it. The political and economic situation at · "Dependent" relations with other government levels. the moment a disaster occurs or when emergency assis- · Lack of awareness about the need for permanent pre- tance is implemented is not independent of the economic vention. Work is only done during emergencies. and political interests of the different actors, the state, · Urban expansion patterns that contribute to an and the presence or absence of local social organizations. increase in the local population's vulnerability. In the cases analyzed, the producers' organizations, · Scarce stimulus for the development of policies agreed in particular the Rural Society, weigh most heavily in to with local society. decisionmaking. It is worth noting, in general terms, · Lack of power to discourage private works in urban the absence of urban social organizations, except for the and/or rural areas, particularly canals that produce new solidarity networks organized to confront the eco- negative impacts throughout the basin. nomic crisis and alleviate hunger through "barter net- · Lack of experience of the different actors in work- works" or food assistance. In certain cases, such as ing together constructively; atomization and a cer- Chascomús, Dolores, and Junín, recurrent flooding tain weakness in civil society organizations having has not been able to change the vulnerability that existed little management capability; a municipal state that prior to disasters and, since floods recur, the popula- leads urban management according to a traditional tion tends to be increasingly vulnerable. model, plus resistance to changing its structures. The weakness of local governments in the Province of Buenos Aires and their lack of true autonomy facing the provincial and national governments should be Conclusion noted. There is an economic and financial subordina- tion at each level that is clearly exemplified during emer- The author would like to point out several issues that, gencies. Municipalities do not have the necessary funds in her opinion, still need to be addressed: to face disasters; civil defense committees are organ- · The decantation of water courses proposed by the ized within them, but they depend upon volunteers. master plan could be highly dangerous without prior The municipality is highly dependent upon provincial interdisciplinaryenvironmentalimpactstudiestoguar- or national financial allotments for land acquisition and antee the benefits of works that will be carried out. evacuee assistance. The municipality also depends upon · Works should be compatible with the needs of the the provincial or federal government to declare a state various social actors--urban and rural, economic and of emergency that will allow for a temporary tax exemp- community, national and local--trying to guarantee tion. In summary, it is a complex web of political, admin- that productive systems will adapt to the specific con- istrative, and financial dependence. ditions of the environment, both urban and rural. Flooding in the Pampean Region of Argentina: The Salado Basin 145 · Discussion must be promoted among the region's Tandil (partially), Tapalqué, Tordillo, 25 de Mayo. Those of the producers about reorganizing their productive sys- Noroeste region are: Carlos Casares, Carlos Tejedor, Florentino tems, and urging them their systems to the existing Ameghino, Gral. Pinto, Gral. Villegas, Hipólito Yrigoyen, Lincoln, risk levels. The master plan's formulation does not Pehuajó, Pellegrini, Rivadavia, Salliquello, Trenque Lauquen, Tres call for discussions among the three different pro- Lomas, and 9 de Julio; and the Encadenadas del Oeste region ducers' associations. are: Adolfo Alsina, Cnel Suárez, Daireaux, Gral. La Madrid, Guam- · A participative process must be promoted in each iní, Laprida (partially), and Saavedra (partially). municipality, to involve citizens of different affected 5. Since the beginning of the 1990s, because of two hydraulic urban localities. works, they link directly through the Jauretche­Mercante­Italia · Multidisciplinary standpoints must be considered. If canal and the Alsina outlet canal, respectively. economic restrictions only allow for an engineering 6. For comparison, the Minnesota River runs 539 kilometers approach, an integral solution will not be reached across the states of Minnesota, Iowa, and South Dakota. and actions that up to now have proved to be insuf- 7. Río Salado Basin Integral Master Plan (Plan Maestro Integral ficient will be repeated over and over again. de la Cuenca del Río Salado), November 1999. 8. Data provided by the 2001 Census; they are preliminary and only correspond to total numbers. Notes 9. At present, Argentina contributes 6 percent of world exports 1. We perceive disaster as a process. Its understanding derives of wheat, from 10 to 20 percent of corn, from 8 to 13 percent from the creation of risk conditions in the course of time, as a of vegetable oils, and from 17 to 22 percent of edible oils. Meat result of the interaction between certain unchaining events exports still represent approximately 9 percent of world exports. (hazards) and society's vulnerability. 10. The basin significantly contributes to the country's farming 2. Some examples may be useful to clarify these concepts. Defor- and cattle-breeding production. From 1994­97, the production estation in a river's high basin may produce flooding in the low of corn, wheat, sunflowers, and soybeans represented 25 percent basin, due to an increase of the runoff or to the exhaustion of of the national total, and meat production was 30 percent. aquifers in the cities, and its impact on droughts in surround- 11. The edible oil industry is mainly located in the provinces of ing areas. Upstream refuse stoppage--clandestine refuse dumps-- Buenos Aires, Santa Fe, and Córdoba. At present, it is one of impact on flooding downstream. The social actors involved will Argentina's most competitive industries, with a manufacturing not only be different because of their differential capabilities at capacity of nearly 18 million annual tons. the political, economic, and social levels, but also because of 12. Given the high oleaginous production level, yield, and export, the different geographical levels of analysis, in which actors between January and March 2002, the government withheld develop their actions and the action's impacts. 20 percent for farming and cattle-breeding exports. 3. The surface of the Río Salado basin is equivalent to 50 percent 13. There are practically no recent or reliable data about the of the Nile basin in Egypt. It is similar in surface to the República number or type of cattle-breeding firms in the area. As an indi- Oriental del Uruguay (187,000 kilometers2). It is three-quarters cation, we can mention a study carried out by the Universidad the size of Greece (122,873 kilometers2). del Centro de la Provincia de Buenos Aires (UNC) that classi- 4. The territory of the Province of Buenos Aires is divided into fied producers in the district of Azul as follows: counties (partidos). Unlike other Argentinean provinces, in the ­ Firms or managers with other main income sources; usually Province of Buenos Aires, the counties' limits coincide with those absentee cattle producers who employ salaried workers: of the municipalities. These are local government units, and the 32 percent. mayor is the elected head of the executive. The counties that ­ Farming and cattle-breeding firms with their own machin- integrate the Salado-Vallimanca-Las Flores region are: Alberti, ery and salaried workers; some rent large extensions of land; Ayacucho, Azul,BenitoJuárez(partially),Bolívar,Bragado,Castelli, they are usually administered by the owner; they are mainly Chacabuco, Chascomús, Chivilcoy, Dolores, Gral. Alvear, Gral. mixed agriculture/cattle-breeding firms: 29 percent. Belgrano, Gral. Guido, Gral. Lavalle, Gral. Paz, Gral. Viamonte, ­ Family firms that suffer loss of capitalization; mainly small Junín, Las Flores, Lobos, Leonardo N. Alem, Maipú, Monte, cattle-breeding firms but they can rent farming land: half of Navarro, Olavarría, Pila, Rauch, Roque Pérez, Saladillo, Suipacha, the producers live on the premises: 25 percent. 146 Building Safer Cities: The Future of Disaster Risk ­ Independent family firms: mainly in arable land areas and 28. In Argentina, there is a long tradition of facing disasters caused they occasionally rent additional land: generally mixed pro- by flooding by means of great plans that imply costly infrastructure duction agriculture/cattle breeding: 12 percent. works. Beginning with Ameghino's well-known study, the Salado 14. The lack of official sources about flooding has led us to Basin has been the object of the most varied studies since the build our own database. 19th century. Several national institutions, with international 15. During dry years, the underground sheets of water were at financial assistance, have funded these projects; very few have 10 meters. been implemented, and only partially executed. Each one of these 16. See in adjoining online annex the flooding and drought mag- programs criticizes the previous one for lack of coordination and nitude index registered between 1978 and 2000. integration. The latest study is the Salado Basin Integral Manage- 17. A more detailed characterization would account for their dif- ment Plan, finished in 2000, proposing works for US$2 billion. ferent sizes, localizations in the basin, and positions in high or As an example only, it is worthwhile noticing that the protection low areas in the same microregion, since flooding has differen- projects for urban areas in Chascomús and Dolores cost US$6 tial impacts on them. million. Given the current recession, spending billions of dollars 18. The Rural Society (Sociedad Rural) gathers the largest num- for flood protection in the basin seems impossible. bers. Also, the Regional Consortiums for Agricultural Experi- mentation (Consorcios Regionales de Experimentación Agrícola) and the Agrarian Federation (Federación Agraria) are present. Bibliography 19. The declaration of a farming and cattle-breeding emergency Banco Provincia. 1995. Diagnostico Ambiental de la Provincia de must be approved by the provincial legislative power and it grants Buenos Aires, Tomo 1 Aspectos Básicos, Coord. Vega Leandro. reductions in gross income tax and stops the execution of Bianco, M. 1996. Junín expansión y usos de sus suelos. Aportes a banking credit debts with the Provincial and National Banks su geografía urbana. Own Edition. (Banco de la Provincia de Buenos Aires and Banco de la Nación). Bustos, E. 2001. "El asesinato de la Pampa Húmeda." Le Monde 20. The local Sociedades Rurales are grouped in a national level Diplomatique, año III, n°30 (December), Buenos Aires. organization. Censos Nacionales de Población y Vivienda. 1980, 1991. INDEC. 21. During the November 2001 flood in the Municipality of Junín, Comisión para el Desarrollo de la Zona Deprimida del Salado. "anonymoushands"openedabreachinLagunadeGómez(orchard 1984. Zona Deprimida del Salado, caracterización y propuesta area) that increased by 0.50 centimeters the Río Salado water de desarrollo, Buenos Aires. level in a very short period, worsening the situation of the bridge CONSEJO FEDERAL DE INVERSIONS. 1980. ASPECTOS GENERALES that is the only access to the city and nearby neighborhoods. INFORME FINAL (1RA. ETAPA), VOL. 1, BUENOS AIRES. 22. For example, in Junín, the access road to the city is a bridge Consejo Federal de Inversiones. 1980. Estudios Económicos over Río Salado, built on National Road Number 7, which col- Informe Final (1ra. Etapa), Vol. 3 parte F, Buenos Aires. lapsed during the November 2001 flood. Although the neigh- "Cuencas Hídricas." 1996. Chapter: Cuenca del río Salado. Insti- bors claimed at the municipality level, this is a national jurisdiction, tuto Provincial del Medio Ambiente de la Gobernación de la through the Road Concessions Control Organization (Organo Provincia de Buenos Aires: 35­55. de Control de Concesiones Viales) (OCCOVI), because it was Domínguez y Carballo, [Initials?]. 1985. "Inundaciones en la privatized and given in concession to New Roads (Nuevas Rutas). Provincia de Buenos Aires durante el período 1972­84." INTA. 23. It is presided over by the Cabinet Chief and it is integrated Halcrow, W. 1999. Plan Maestro Integral Cuenca del Río Salado, by the following ministers: Internal Affairs, Defense, Economy, Informe Final, Buenos Aires. Infrastructure and Housing, Health Care and Social Develop- Herzer, H. and R. Gurevich. 1996. "Degradación y Desastres. ment, and Environment. Parecidos y Diferentes. Tres casos argentinos para pensar y 24. It is forbidden to build under the 9.40-meters level. algunas dudas para plantear." In M.A. Fernández, ed., Ciu- 25. Their land is located on the other side of the dunes. dades en riesgo. Degradación Ambiental, Riesgos Urbanos y Desas- 26. Together, these canals can transport large volumes of water. tres. Red de estudios Sociales en Prevención de desastres en 27. The Basin's Master Plan identifies 29 national and 19 América latina. Perú. provincial organizations (9 of them are basin committees) involved Herzer, H. and others. 1993. The management of secondary cities in the basin. in Latin America. Nairobi: UNCHS-Habitat. Flooding in the Pampean Region of Argentina: The Salado Basin 147 Municipio de Junín. 2000. Plan Estratégico de Desarrollo. Interviews Natenzon, Claudia and others. 1993. "Inundaciones, produc- Department of Production, Dolores: Carlos Alvarez. ción agropecuaria y agentes sociales de la Cuenca del Río Department of Tourism, Chascomús. Salado--Provincia de Buenos Aires" Instituto de Geografía, Department of Tourism, Dolores Facultad de Filosofía y Letras, University of Buenos Aires. Department of Urban Planning, Dolores: Architect Conti PNUD (United Nations Development Programme). 2000. "El INCOBYP: José Cornejo (former manager Plan Maestro Cuenca mundo del trabajo: los recursos humanos, la inserción labo- del Salado) ral y la calidad de vida." Informe Final. Partido de Dolores. INCOBYP: Horacio Tavecchio Proyecto de Desarrollo Sustentable. Regional. Arg 96/015. INTECH Water Laboratory, Chascomús: José Fernando Tauber, Fernando. 1993. Chascomús. Pautas para una estrategia Bustingorry. de desarrollo. Municipality of Chascomús. Fundación S. Karaka- Provincial Civil Defense Directorate: Operations Director Alberto choff. Ripalna Tauber, Fernando. 1997. Reflexiones y Datos para una estrate- Red Solidaria of Chascomús: Alicia gia de desarrollo. Departamento de Publicaciones del Minis- Residents of Chascomús, Dolores, and Junín. terio de Economía de la Provincia de Buenos Aires. Rural Producer, Junín: Omar Girolami Teubal, M. and Rodríguez. 2001. "Ajuste, reestructuración y The Rural Society of Dolores: President Germán Galdos crisis del agro." Le Monde Diplomatique año III, n°30, (Decem- The Rural Society of Junín: President Alejandro Borges ber), Buenos Aires. Undersecretary of Public Works: Hugo Fusé Varley, Ann. 1994. Disasters. Development and Environment. UK: Undersecretary of Public Works, Delores: Engineer Petasi John Wiley. Vocational Historian: Juan Carlos Pirali Chapter 11 Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century Hans Günter Brauch This paper analyzes the relationship between urban- Rising Vulnerability to Disasters due ization--in relation to population growth and climate to Urbanization in the Mediterranean change--and natural disasters as outcomes of envi- ronmental stress for the Mediterranean space. It includes During the 21st century, urban vulnerability as a result trends of urbanization up to 2015 and 2030, popula- of population growth is expected to rise, as will extreme tion growth up to 2050, and models of climate change weather events due to regional climate change impacts, until 2100 that will increase both the vulnerability to leading to more frequent and intensive hydro- and impact of natural disasters. Based on a medium meteorological hazards both globally and regionally. The definition of the Mediterranean space (Brauch 2001; paper argues that a North-South cleavage in vulnera- Brauch and others 2003) this paper addresses the increas- bility to disasters may increase due to these factors: ing vulnerability of major cities to geophysical and · Vulnerability will increase in megacities if no major hydro-meteorological disasters due to rapid urbaniza- progress is achieved on poverty eradication, disas- tion combined with extreme poverty in most countries ter preparedness, and improved urban building of the Middle East, North Africa, and Turkey (Brauch standards. and others 2003). The paper states that in the 20th · The impact of extreme weather events will increase century an increase in reported natural disasters, fatal- in Mediterranean megacities. ities, and affected people could be observed in the · A major obstacle to an effective regional strategy of Mediterranean space; however, the fatalities diverged disaster reduction is that the Mediterranean space is between Southern European E.U.-member countries institutionally divided among three continents (Brauch and riparian countries in the Balkans, the Middle East, 2001; Baruch and others 2003). and North Africa. Geophysical and hydro-meteorological disasters share After outlining the theoretical and conceptual con- common features, but natural and human-induced text, trends in population growth, urbanization, and regional environmental challenges impact the Mediter- variations in growth patterns of megacities will be ana- ranean region differently. While climate change, deser- lyzed. Projected regional climate change impacts will tification, and the hydrological cycle have contributed also be noted, as well as a survey of reported disasters to environmental degradation in this region during for the Mediterranean. Two cases of recent urban dis- the 20th century (Brauch 2001, 2002b), human-induced asters, Izmit (earthquake of 1999) and Algiers (flash demand factors (population growth, urbanization, flood of 2001), will also be discussed as well as the agriculture/food demands) will increase pressure on the impacts of the projected sea-level rise for Alexandria. environment even more during the 21st century. These The paper concludes with suggestions for reducing vul- trends have impacted the vulnerability of urban cen- nerability and mitigating the impact of disasters in the ters to natural disasters in Europe, the Middle East, Mediterranean. and North Africa. 149 150 Building Safer Cities: The Future of Disaster Risk Theoretical Context: Human and Environmental 2002) pertain specifically to the environmental dimen- Security from a Grotian Security and an sion and the human level of security. In dealing with Equity-oriented Ecological Perspective environmental security policy issues (figure 11.2), the perspective of a Grotian pragmatist may best reflect the The reality that we observe is influenced by our intel- perspective of international financial institutions that lectual traditions and conceptual lenses (figure 11.1). is also preferred by this author (Brauch and others 2003). On international policy, three traditions may be distin- The security concept widened horizontally and ver- guished: a) the Hobbesian pessimist, where power is the tically (figure 11.2) during the 1990s (Buzan and others key category; b) the Kantian optimist, where interna- 1998). While military institutions and alliances focus tional law is crucial; and c) the Grotian pragmatist, where primarily on regional and national security through mil- cooperation is vital (Wight 1991; Bull 1977; Brauch itary means, international organizations have used con- and others 2003). On international environmental policy, cepts of environmental (NATO, OECD, OSCE) and human three positions may be distinguished (Gleditsch 2003; security (UNDP 1994). Environmental security refers to Brauch and others 2003): those of a): a Malthusian implications of environmental degradation, scarcity, and pessimist, who claims that resource scarcities will rise; stress due to disasters, migration, crises, and conflicts, and b) a Cornucopian optimist, for whom plenty of resources on the resolution, prevention, and avoidance of envi- will meet with all challenges (Lomborg 2001); and c) an ronmental damage (figure 11.3). The chairman of the equity-oriented pragmatist, who calls for multilateral IPCC, R.K. Pachauri (2000), defined environmental cooperation in international organizations and regimes. security as "the minimization of environmental damage While national measures of disaster reduction and pre- and the promotion of sustainable development, with a paredness in the Mediterranean are indispensable, close focus on transboundary dimensions." He pointed to sev- multilateral cooperation to improve international dis- eral linkages between poverty and natural resource stress aster response is needed. These efforts for disaster reduc- that also increase the vulnerability to and impact of nat- tion, preparedness, and risk management (UN/ISDR ural disasters. Pachauri's interpretation is of relevance Figure 11.1 Worldviews and environmental standpoints Worldviews/radiations Machiavelli, Grotius, Kant on security ( ) Hobbes, Morgen- pragmatist neo-liberal institu- thau, tionalist (optimist) (neo)realist Cooperation is needed International law pessimist and matters matters and prevails Standpoints on environ- Power matters (Democratic peace) mental issues ( ) I II III Neo-Malthusian Perspective of most Resource scarcity MENA states (pessimist) Equity-oriented reformist IV V VI Multilateral cooperation will/ U.N. system may solve challenges World Bank (pragmatist) (author's position) Cornucopian VII VIII IX Technological ingenuity will Wilsonian solve issues liberal optimism (neo-liberal optimist) Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 151 Figure 11.2 Horizontal and vertical security dimensions Security dimension Military Political Economic Environmental Societal Level of interaction Urbanization, Human (security) disasters Societal/community Middle East, North National Northern focus (NATO, E.U. countries) Africa focus International/regional Global/planetary Source: Brauch 2002a, 2003. Figure 11.3 Causes and outcomes of environmental stress Global environ- Effect Environmental Political Probable mental change (interaction) stress process outcomes Nature-induced Impact State Natural (supply side) extreme weather disaster events due to cli- prevention (environmental) mate change · climate change avoidance degradation · desertification · water supply decision environmental stress crisis Human-induced scarcity society economy (demand side) · population growth vulnerability adaptation and (due to poverty, mitigation deci- · urbanization population growth) sions · agriculture Migration conflict knowledge Mitigation and adaptation, reduce impact and vulnerability (Strategy and means) Source: Brauch 2002b. for North-South environmental security issues in the the linkages between urbanization and disasters. In the Mediterranean. Environmental security issues are often Mediterranean, there have been fundamental differences addressed from a humansecurity (Newman and Richmond in the conceptualization of security: while there has been 2001) perspective. For Kofi Annan (2001: xix) human a widening of the security concept in the U.N. family and security is a "people-centred concept" that "no country ... in most E.U. Mediterranean countries, in the Mashreq can achieve ... on its own, and none is exempt from (Selim in Brauch and others 2003), in Israel (Kam 2003), risks and costs if it chooses to do without the multilat- and Turkey (Aydin in Brauch and others 2003) the narrow, eral cooperation that can help us reach this goal." hard security focus on military and political factors has Thus, both the humansecurity perspective and the envi- prevailed. In the North and South, different concepts of ronmental security dimension are best suited for analyzing sovereignty and security prevail (Brauch 2001). 152 Building Safer Cities: The Future of Disaster Risk Model: Global Challenges, Environmental Stress major driver for the rapid expansion of megacities and and Outcomes for informal housing quarters that are highly vulnera- ble to disaster (UNEP 2002). Both past and projected This model distinguishes between causes and effects of population growth (table 11.1) are crucial for assess- environmental stress and outcomes: a) natural disasters; ing future vulnerabilities to disasters. b) environmentally-induced migration; c) crises; d) con- In the Mediterranean, the demographic data indicate flicts; and e) efforts of environmental conflict preven- two patterns (Zlotnik in Brauch and others 2003). Due tion and avoidance. In addition to urbanization as the to different stages of demographic transition (Lutz 1994, independent variable, the model includes population 1996, 2002), between 1850 and 2000, the population growth and climate change as two intervening variables. in the five Southern European EU-member countries Strategies for disaster reduction must address both vul- doubled, while that of the 12 EU dialogue partners (plus nerability and impact for the state, the society, and the Libya) increased 9-fold (table 11.1). From 2000 to 2050, economy by enhancing and exploiting available knowl- a declining population has been projected in the five edge (figure 11.3). Southern (-23.2 million) and Southeastern European Disasters (Hewitt 2002; Smith 2001; Abramovitz (-2.3 million) countries (except Albania), slight increases 2001) may be the result of purely natural processes as well in Cyprus, and major increases in North Africa (+96.6 as anthropogenic factors and of unintended technological million)andintheEasternMediterranean(+84.3million). malfunctioning or deliberate terrorist attack. Natural disas- In the 12 Middle East and North African countries, more tersmaybeacausefornature-inducedmigration(UNHCR/ people will be added until 2050 than presently live in the IOM/RPG 1996), for distress migration (Meze-Hausken five Southern European EU countries (177.3 million). 1998), and for environmental refugees (El-Hinnawi 1985; These different population growth patterns will affect Myers 1995; Brauch 2000/01); they may also be a cause vulnerability to disasters. for domestic or international crises that may escalate into violent conflicts, or that may be avoided by polit- ical efforts (Brauch 2002b; 2003). Natural disasters (e.g., Urbanization Trends in the Mediterranean Region Sahel, Bangladesh) have already become one among (1950­2030) several triggers for environmentally induced migration, which has led in some cases to domestic crises. Urbanization trends in the region have differed signif- For dealing with disaster reduction and risk man- icantly. While in Southern Europe, the urbanization rate agement, the political process is vital (UN/ISDR 2002). has been projected to increase to 75.2 percent by 2030 The state is responsible for the initiation of adaptation (from 44.2 percent in 1950), in North Africa, the urban- and mitigation measures. However, its resources are con- ization rate has been projected to increase even more strained by socioeconomic and political factors. Human rapidly--from 24.7 percent in 1950 to 63.3 percent by vulnerability and fatalities can be reduced by capacity 2030. In the Middle East and North Africa, urbaniza- building within society, by improved early warning sys- tion rates have differed (1950­2000), as have the pro- tems, better building standards, success in poverty erad- jections until 2030 (table 11.2). Between 1950 and ication, and good governance. 2000, the most rapid increase in the urbanization rate occurred in Libya, from 18.6 percent to 87.6, and in Lebanon, from 22.7 percent to 89.7 percent. Population Growth in the Mediterranean Region By 2030, about 94.6 percent of the population in (1850­2050) Israel will be urban, 93.9 percent in Lebanon, and 92 percent in Libya. This compares with 71.6 percent in Urbanization is caused by push- and pull-factors, the Greece, 76.1 percent in Italy, 81.6 percent in Portugal, structure of the economy, and the stage of economic 82.2 percent in France, and 84.5 percent in Spain. While development (Rakodi 1997; Lo 1998; Mitchell 1999a; in 1950, the urbanization rate in North Africa was the UNCHS 2001a, 2001b). Population growth has been a highestinEgypt,with31.9percent,by2030,atanexpected Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 153 Table 11.1 Population growth of Mediterranean countries, 1850­2050 Projection Real population development med. var. Changes 2000 2050 1950­2050 2000­2050 1850 1900 1950 1965 1980 (2000 rev.) (2000 rev.) (2000 rev.) Five southern European E.U. countries France 36.0 41.0 41.829 48.753 53.880 59.238 61.832 20.003 ­2.594 Greece 3.5 4.5 7.566 8.551 9.643 10.610 8.983 1.417 ­1.627 Italy 25.0 34.0 47.104 52.112 56.434 57.530 42.962 ­4.142 ­14.568 Portugal 3.5 5.5 8.405 9.129 9.766 10.016 9.006 .601 ­1.010 Spain 15.0 18.5 28.009 32.065 37.542 39.910 31.282 3.273 ­8.628 Total (5) 83.0 103.5 132.913 150.610 167.265 177.304 154.065 21.152 ­23.239 Two E.U. candidates and dialogue partner countries Cyprus 0.15 0.23 0.494 0.582 0.611 0.784 0.910 0.416 0.126 Malta 0.13 0.19 0.312 0.305 0.324 0.390 0.400 0.088 0.010 Total (Islands) 0.28 0.42 0.806 0.887 0.935 1.174 1.310 0.504 0.136 Yugoslavia and Albania Albania 0.5 0.8 1.215 3.134 3.905 2.690 0.771 Yugoslavia 7.25 9.5 16.345 23.205 20.088 3.743 ­3.117 - Bosnia and 2.661 3.977 3.458 ­0.519 Herzegovina. - Croatia 3.850 4.654 4.179 ­0.475 - Macedonia 1.230 2.034 1.894 ­0.140 - Slovenia 1.473 1.988 1.527 ­0.461 - FR Yugoslavia 7.131 10.552 9.030 ­1.522 Total 7.75 10.3 17.560 26.339 23.993 6.433 ­2.346 Ten Non-E.U.-Mediterranean dialogue partners (plus Libya) Algeria 3.0 5.0 8.753 11.823 18.740 30.291 51.180 42.427 20.889 Morocco 3.0 5.0 8.953 13.323 19.382 29.878 50.361 41.408 20.483 Tunisia 1.0 1.5 3.530 4.630 6.448 9.459 14.076 10.546 4.617 Libya 0.6 0.8 1.029 1.623 3.043 5.290 9.969 8.940 4.679 Egypt 5.5 10.0 21.834 31.563 43.749 67.884 113.840 92.006 45.956 Only North Africa 13.1 22.3 44.099 62.962 91.362 142.802 239.426 195.327 96.624 Jordan 0.25 0.3 1.237 1.962 2.923 4.913 11.709 10.472 6.796 Israel 1.258 2.563 3.879 6.040 10.065 8.807 4.025 Palestine Authority 0.35 0.5 1.005 ? ? 3.191 11.821 10.816 8.630 Lebanon 0.35 0.5 1.443 2.151 2.669 3.496 5.018 3.575 1.522 Syria 1.5 1.75 3.495 5.325 8.704 16.189 36.345 32.850 20.156 Turkey 10.0 13.0 20.809 31.151 44.438 55.668 98.818 78.009 43.150 Eastern Med. 12.45 16.05 29.247 43.152 62.613 89.497 173.776 144.529 84.279 10+1 dialogue c. 25.55 38.35 73.346 106.114 153.975 232.299 413.202 339.856 180.903 Total (12+1) 25.83 38.77 74.152 107.001 154.910 233.473 414.512 340.360 181.039 Sources: McEvedy/Jones 1978 for 1850, 1900; for projections to 2050: UN 2001. The data for 1960, 1980, and for the 1998 UN revisions are from Heilig 1998a, 1998b. 54.4 percent, it would be the lowest. This trend is also Marseille) have increased the least (1.1- to 1.8-fold) and reflected in the growth of major urban centers around will stabilize until 2015. In the Middle East and North the Mediterranean (table 11.3). Africa, Shubra el Kheima (Egypt) grew 25-fold; Tripoli, In Southern Europe from 1950 to 2000, the Mediter- Amman, and Rabat grew 10- to 15-fold; Istanbul, Aleppo, ranean coastal cities (Rome, Athens, Barcelona, Naples, Damascus, Beirut, Ankara, Casablanca, Tel-Aviv, and Izmir 154 Building Safer Cities: The Future of Disaster Risk Table 11.2 Changes in the urbanization rates of MENA countries (1950­2030) Ten Non-E.U.-Mediterranean dialogue partners (plus Libya) in percentage 1950 1960 1970 1980 1990 2000 2010 2020 2030 Algeria 22.3 30.4 39.5 43.5 51.4 57.1 62.2 67.5 71.7 Morocco 26.2 29.2 34.6 41.3 48.4 55.5 61.7 66.7 71.0 Tunisia 31.2 36.0 44.5 51.5 57.9 65.5 71.3 75.2 78.4 Libya 18.6 22.7 45.3 69.3 81.8 87.6 89.7 90.9 92.0 Egypt 31.9 37.9 42.2 43.8 43.6 42.7 44.0 48.2 54.4 Only North Africa 24.7 30.1 36.3 40.4 44.8 48.9 53.4 58.2 63.3 Jordan 35.9 50.9 56.0 60.2 72.2 78.7 80.1 82.2 84.4 Israel 64.6 77.0 84.2 88.6 90.3 91.6 93.0 93.9 94.6 Palestine Authority 37.3 44.0 54.3 61.1 64.0 66.8 70.0 73.5 76.9 Lebanon 22.7 39.6 59.4 73.7 84.2 89.7 92.1 93.1 93.9 Syria 30.6 36.8 43.3 46.7 48.9 51.4 55.4 60.6 65.6 Turkey 21.3 29.7 38.4 43.8 61.2 65.8 69.9 73.7 77.0 Western Asia 26.7 35.0 44.4 51.7 62.0 64.7 67.2 69.8 72.4 Urbanization rates for three continents and the world Africa 14.7 18.5 23.1 27.4 31.8 37.2 42.7 47.9 52.9 Asia 17.4 20.8 23.4 26.9 32.3 37.5 43.0 48.7 54.1 Europe 52.4 58.0 64.6 69.4 72.1 73.4 75.1 77.6 80.5 World 29.8 33.7 36.8 39.6 43.5 47.2 51.5 55.9 60.2 Source: UN 2002. Table 11.3 Growth of urban centers in the Mediterranean, 1950­2015 (millions) City 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Athens 1.8 2.0 2.2 2.4 2.5 2.7 3.0 3.0 3.0 3.1 3.1 3.1 3.1 3.1 Istanbul 1.08 1.37 1.74 2.20 2.79 3.60 4.40 5.41 6.54 7.66 8.96 9.95 10.72 11.36 Ankara 0.54 0.69 0.87 1.09 1.35 1.71 1.89 2.21 2.54 2.83 3.16 3.38 3.58 3.78 Izmir 0.48 0.56 0.66 0.77 0.89 1.05 1.22 1.47 1.74 1.97 2.21 2.39 2.55 2.70 Rome 1.57 1.91 2.33 2.64 2.91 3.00 3.02 2.93 2.81 2.65 2.65 2.65 2.65 2.65 Milan 3.63 4.05 4.50 4.99 5.53 5.53 5.33 4.98 4.60 4.25 4.25 4.25 4.25 4.25 Naples 2.75 2.96 3.19 3.39 3.59 3.62 3.59 3.42 3.21 3.01 3.01 3.01 3.01 3.01 Turin 0.88 1.05 1.25 1.42 1.62 1.64 1.60 1.50 1.39 1.29 1.29 1.29 1.29 1.29 Cairo 2.41 3.00 3.71 4.61 5.33 6.08 6.86 7.67 8.30 8.86 9.46 10.09 10.77 11.53 Alexandria 1.04 1.25 1.50 1.75 1.99 2.24 2.52 2.83 3.06 3.28 3.51 3.75 4.02 4.33 Shubra el Kheima 0.04 0.06 0.10 0.16 0.24 0.35 0.49 0.66 0.77 0.85 0.94 1.03 1.13 1.23 Tel-Aviv 0.42 0.56 0.74 0.88 1.03 1.21 1.42 1.62 1.80 1.90 2.00 2.13 2.27 2.40 Amman 0.09 0.14 0.22 0.30 0.39 0.50 0.64 0.78 0.94 0.99 1.15 1.31 1.48 1.65 Beirut 0.34 0.43 0.56 0.72 0.92 1.06 1.21 1.39 1.58 1.82 2.07 2.28 2.42 2.50 Damascus 0.37 0.46 0.58 0.73 0.91 1.12 1.38 1.56 1.73 1.92 2.14 2.43 2.78 3.17 Aleppo 0.32 0.39 0.48 0.59 0.72 0.88 1.07 1.29 1.55 1.87 2.23 2.62 3.05 3.49 Marseille 0.7 0.7 0.8 0.9 1.1 1.2 1.2 1.2 1.2 1.2 1.30 1.32 1.34 1.36 Algiers 0.50 0.62 0.81 1.07 1.28 1.57 1.62 1.67 1.91 2.30 2.76 3.27 3.74 4.14 Tunis 0.47 0.53 0.58 0.64 0.74 0.87 1.14 1.43 1.57 1.72 1.90 2.07 2.25 2.41 Tripoli 0.11 0.14 0.17 0.24 0.40 0.61 0.80 1.04 1.30 1.52 1.73 1.94 2.12 2.27 Barcelona 1.6 1.7 1.9 2.3 2.7 2.9 3.1 3.0 2.9 2.8 2.73 2.73 2.73 2.73 Casablanca 0.63 0.78 0.97 1.21 1.51 1.79 2.11 2.41 2.69 2.99 3.36 3.78 4.22 4.61 Rabat 0.15 0.18 0.23 0.34 0.49 0.64 0.81 0.98 1.16 1.37 1.61 1.88 2.13 2.34 Fes 0.17 0.22 0.28 0.32 0.37 0.43 0.51 0.59 0.68 0.79 0.90 1.04 1.18 1.30 Marrakech 0.21 0.23 0.24 0.28 0.32 0.37 0.42 0.49 0.58 0.69 0.82 0.96 1.10 1.21 Source: UN 2002. Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 155 5- to 10-fold; and Cairo, Tunis, Alexandria, and Algiers common damaging phenomena, followed by other risks 2- to 5-fold. Among the 30 largest urban agglomera- that trigger disasters. He noted major changes in megac- tions in 1950, two were in the Mediterranean: Milan at ity hazards but also in interactivity, risks, change in expo- number 14 with 3.633 million, and Cairo at number 25 sure,vulnerabilityandintheefficacyofhazardmanagement. with 2.410 million. In 2000, Cairo was number 20 with Not only will vulnerability rise, but also the impact of 9.462 million, and Istanbul number 22 with 8.953 extreme weather events.1 million. By 2015, Cairo will be number 18 with 11.531 million and Istanbul number 19 with 11.362 million (UN 2002). Climate Change: IPCC Projections of Extreme Mitchell (1999a: 29) estimated that Cairo's popula- Weather Events tion density was 37,726 inhabitants per square kilo- meter. Egyptian authors (Yousry 1998: 301) stated an According to the Third Assessment Report (TAR) of average population density for the Greater Cairo the Intergovernmental Panel on Climate Change region of 32,000, with the city of Cairo's population (IPCC 2001a: 13), between 1990 and 2100, "the glob- density at 78,300 (possibly reaching 87,000 in Dar ally averaged surface temperature is projected to increase Elsalam) and Giza's at 51,300 (in Imbaba and Mounira by 1.4 to 5.8°C and the mean sea level will rise by 0.09 even 84,000) inhabitants per square kilometer. As in to 0.88 meters." The TAR argued that global climate many megacities, social problems (inequality, unem- change increased the probability of some extreme weather ployment, crime) associated with the ever-increasing events during the 20th century and that, in the 21st rate of growth in Cairo surpassed the capacity of the century, "more intense precipitation events" and an government "to cope properly with ... and to manage "increase of the heat index" will become "very likely, it" (Yousry 1997: 134­137). This rapid urbanization over most areas" (IPCC 2001a: 3). made Cairo highly vulnerable to disasters. On October Due to regional climate differences, "expected cli- 12, 1992, an earthquake in Cairo caused 561 fatalities mate change will give rise to different exposures to cli- and economic damage totaling US$1,200 billion (Munich mate stimuli across regions." Less-developed regions Re 1998). are severely vulnerable and "in Europe, vulnerability is According to UNCHS (2001a: 13), in Arab coun- significantly greater in the south" (IPCC 2001b: 15). tries "urban growth rates will remain higher than total This has also been stressed in the IPCC assessments of population growth rates in the foreseeable future.... the climate scenarios for Europe pertaining to the changes Urban growth has been the result of rural-to-urban in temperature and precipitation during summer peri- migration as well as high fertility and declining rates of ods for the 2020s, 2050s, and 2080s. The projected mortality." The UNHCS notes that "many cities are trend for Southern Europe is obvious: the temperature now going through a critical phase of development, may increase most and precipitation may decline most marked by dwindling resources, increasing poverty, and in Mediterranean countries. Climate change produces serious environmental degradation." short-term and long-term impacts that can contribute Both international organizations and individual experts to disasters that vary according to the specific vulner- (Mitchell 1995; Ichikawa 1995) noted "the increasing abilities that may be reduced by both adaptation and disaster potential of megacities." IDNDR (1996) listed mitigation measures. among 17 cases of disasters in urban areas three from In response to human activities and the natural envi- the Mediterranean: a) an earthquake in 1980 in Naples, ronment, Europe and the Mediterranean are sensitive Potenza, Salerno, Avelino with 3,000 deaths; b) an earth- to extreme seasons (exceptionally hot, dry summers, mild quake in 1992 in Erzincan in Turkey that killed 547 winters), short-duration hazards (windstorm, heavy rain, and affected 230,000; and c) an earthquake in Cairo river-valley flooding), and slow, long-term change (coastal that destroyed 5,000 buildings and damaged 12,000. squeeze, sea-level rise) (IPCC 2001a: 647). Projected Mitchell (1999b: 475) noted that megacity hazards, such climate change would also be associated with increases as floods, earthquakes, and windstorms, are the most in heat waves. The rise in floods "will increase the risk 156 Building Safer Cities: The Future of Disaster Risk of drowning, diarrhoeal and respiratory diseases, and while numbers declined for Tunisia: 96,000 in 1990, in developing countries, hunger and malnutrition" (IPCC 30,000 in 1982, and 2,500 in 1986. 2001b: 12). Climate change will also impact on human Between 1975 and 2001, geophysical disasters caused settlements, and worsen existing trends further. "In such the most fatalities (in Turkey, Italy, and Algeria), while the areas, squatter and other informal urban settlements winter storms in France, and drought in Spain, Albania, with high population density, poor shelter, little or no Syria, and Morocco caused an increasing number of access to resources such as safe water and public health disaster-affected people. The 1999 earthquake in Turkey services, and low adaptive capacity are highly vulner- and the flood in Algeria in 2001, discussed later, demon- able" (IPCC 2001b: 13). The disaster impact is caused strated the vulnerability of urban centers. by hazards whose intensity is influenced by social, eco- nomic, physical, and environmental vulnerabilities. According to the ISDR (UN/ISDR 2002: 23) framework, Geophysical Disasters: Earthquakes any risk assessment relies on a vulnerability capability and Volcanoes analysis, a hazard analysis and monitoring, and knowl- edge development. Risk assessment contributes to prob- The many earthquakes in the Mediterranean are due to lem awareness, public commitment, the application of repeated collisions between the African Plate, the Eurasian risk reduction measures, and to early warning. Plate, and several small plates (Arabian, Adriatic, and Iberian). According to Wagner (2001: 213) during the 20th century, about 60 major earthquakes were recorded Analysis of the Trends in Disasters in the Mediterranean, some 20 in northwestern Turkey. in the Mediterranean During this period, earthquakes in the Mediterranean caused at least 250,000 fatalities (Wagner 2001: 213). Asurveyofnaturaldisastersfrom1975to2001(table11.4) in the Mediterranean indicates that more than half of all the natural disasters were reported for the five South European countries.2 During these 27 years, most natural Hydro-meteorological Disasters: Storms, Floods, disasters were reported in France (86), followed by and Drought Turkey (63), Italy (57), Spain (47), Greece (43), Algeria (36), and Morocco (23). With regard to fatalities, Turkey According to Munich Re (1998: 9) "in terms of frequency ranked first (27,375), followed by Italy (6,158) and of damage and total area affected, storms are, world- Algeria (4,124), Greece (1,573) and Egypt (1,386). But, wide, the most significant of all natural hazards... [O]ver for affected persons, Spain was in the lead (6,819,987), the period 1988 to 1997, two-thirds of the claim pay- followed by France (3,890,759), Albania (3,259,759), ments (US$130 billion) for natural catastrophes were Turkey (2,580,392), and Algeria (1,154,355). Around occasioned by storms." In the Mediterranean severe the Mediterranean, most disaster fatalities were the result winter storms and floods have increased in frequency of earthquakes (Turkey, Italy, Algeria), though drought and intensity and, in part due to rapid urbanization, the and famine affected more people. number of fatalities and economic losses have been rising From the 1980s to the 1990s disaster fatalities in the eastern and the southern regions. Due to popu- increased for Turkey, Egypt, Morocco, Italy, and France, lation growth, the impact of drought differed in the while the affected persons increased most for Spain but North and South (Mendizabal and Puig in Brauch and also for France (IFRC 2001: 186­197, 2002: 197­203). others 2003). In Albania the drought of 1991 affected about 3.2 In Europe, floods are the most common and the most million, while the earthquakes in Turkey in the 1990s costly natural disaster. From 1991 to 1995, the eco- affected more than 2 million. In North Africa, the number nomic cost of flood damage has been estimated at EUR of people affected by natural disasters increased from 99 billion. A study by the European Environment Agency the 1980s to the 1990s for Egypt, Morocco, and Algeria, (2001) stated that "the main areas prone to frequent Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 157 0 0 3 70 410 200 of fected -- 1,119 1,090 s 60,412 -- -- 1,090 -- -- -- -- -- 25,188 10,117 15,071 -- -- -- m Af 104,688 104,075 3,566,519 3,504,918 3,697,485 declaration a indstor W 84 76 4 0 0 0 -- -- -- -- -- -- -- 07 3 11 52 -- 13 96 4 15 41 4) Killed 469 239 111 -- -- 608 or . 2 9 2 0 2 0 1 1 8 1 3 1 3 6 2 3 1 Ev 60 34 13 -- -- -- -- -- -- -- -- -- 76 assistance; 0 -- 172 fected 10,150 67,622 47,220 -- -- -- 1,200 1,500 1,000 1,500 46,500 83,008 13,000 18,029 -- 92,157 -- Af 372,125 741,300 145,208 112,858 656,640 141,765 229,868 155,757 129,250 national 1,238,417 2,153,123 inter Floods 87 99 0 4 21 29 11 for 17 27 837 143 319 198 -- -- -- 108 -- -- -- 505 -- -- 450 673 -- 873 177 Killed 2,924 1,201 4374 call a 3) . 8 4 0 2 1 5 4 2 2 1 2 5 1 Ev 71 30 16 13 -- -- -- 12 -- 24 -- 17 38 17 11 4 145 fected; 0 0 af 20 fected 3,815 ,1153 700 -- ,0456 -- ,0002 -- -- -- -- -- -- -- 34,998 -- -- -- Af 930,925 834,765 414,405 406,360 eportedr 1,765,710 2,377,128 2,377,128 1,036,210 1,001,212 5,597,268 0 0 0 2 2 63 people 335 187 151 571 Earthquakes Killed 5,672 -- -- -- -- -- -- -- -- -- -- -- 3,452 2,881 -- -- -- 100 6,0077 26,087 26,087 35,735 2) thousands) . (in 0 Ev 33 71 51 0 1 2 1 1 4 1 6 -- 11 -- -- -- 23 -- -- -- -- -- 32 8 2 10 -- -- -- 79 killed; 2001 0 -- -- to eportedr fected -- -- -- -- -- -- -- -- -- 10,500 -- -- -- -- -- -- -- 31,400 Af 988,000 330,000 658,000 306,400 275,000 1975, 6,000,000 6,000,000 3,210,500 3,200,000 10,504,900 people e mor ought/famine 0 -- -- -- -- 0 0 -- -- -- 0 -- -- -- -- -- -- 0 -- -- -- -- -- -- 0 -- -- -- -- -- 0 or country Dr Killed 10 by . 1) 8 1 1 0 2 4 2 2 3 1 1 1 5 1 2 2 3 5 2 Ev -- -- -- -- -- -- -- -- 10 -- -- 28 apply: disasters ,751 700 ,400 943 1,893 2,029 fected 49,690 4,451 -- 11,500 83,959 -- criteria Af 944,035 921,154 419,420 348,956 105,575 662,165 280,342 442,973 160,650 3,890,759 6,819,987 3,779,928 3,259,756 3,700,060 2,580,392 2,038,320 1,154,355 12,625,625 22,148,384 these natural of by otalT 9 93 13 5 5 5 31 47 one 524 132 501 562 187 298 115 919 177 Killed 8,888 1,573 6,158 4,124 1,386 27,613 27,375 6,606 43,728 least killed at e 95 85 1-- 46 64 41 91 44 1-- 5 1-- 8 86 43 57 16 47 50 12 15 95 11 11 63 82 36 14 23 1992. Event 249 ---- 485 wher 1991). eportedr 1991, events in until . . (E.U.) (1974 People includes a a gency a ope c. Auth. T b a Mediter emer 11.4 Eur Her a n Africa Candidates of independence Maced. occo EM-DA ugoslaviaY eece oatia dan ableT South France Gr Italy Portugal Spain E.U. Cyprus Malta Slovenia Balkans Albania Bosnia Cr FYR Serbia/Mont. ugoslaviaY Easter Israel Jor Lebanon Palestinian Syria urkeyT North Algeria Egypt Libya Mor unisiaT otalT state since for Note: a a. b. 158 Building Safer Cities: The Future of Disaster Risk floods include the Mediterranean ... (where) a major The Earthquake in Western Turkey risk factor is the occupation of potential flood areas in August 1999 through uncontrolled building and inadequate land- use planning." 3 While the number of flood events and In Turkey, 23 of 63 reported events (1975­2001) were amount of economic damage have increased from 1992 earthquakes that caused 26,087 fatalities and affected to 1998, human fatalities have remained low in Europe. 2,377,128. In 1998 and 1999, three earthquakes and The most costly flash flood in November 1995 in Pied- one flood in Turkey were among the 100 major world- mont caused economic damage of US$13 billion but wide disasters (Munich Re 1999, 2000). The earthquake only 64 fatalities. These differences are due to the vul- with the highest economic losses (US$12 billion) and nerability of the victims and the disaster preparedness the second-highest fatalities in the 20th century occurred and response. on August 17, 1999, when 17,200 persons died. Accord- ing to the 2001 ISDR report, the continued high vul- nerability of Turkey is attributable to: a) population Vulnerabilities of Cities to Disasters: Cases of growth and urbanization; b) the failure to apply exist- Izmit and Algiers ing building regulations consistently; and c) the siting of industrial facilities wherever space is available, with From the comprehensive analysis of disasters in the no regard for environmental protection rules. In this Mediterranean during the 20th century (Brauch and earthquake, 321,000 people lost their jobs and about others 2003), the following different trends emerged 600,000 became homeless. On November 16, 1999, among its four subregions (table 11.4). the World Bank granted two loans totalling US$757.53 · From 1975­2001, the most natural disasters were million4 and on February 9, 2000, the European Invest- reported for Southern Europe (249), followed by the ment Bank provided a EUR 450 million facility.5 The Eastern Mediterranean (95), North Africa (82), and impact would be more severe if a future earthquake the Balkans (50). should strike Izmir, Istanbul, or Ankara due to the · The number of fatalities was highest in Turkey high population density and large areas with informal (27,375), followed by Italy (6,158), Algeria (4,124), housing. Geologists have predicted that in the next 30 Greece (1,573), and Egypt (1,386). Most fatalities years there is a 62 percent probability of a major earth- were from earthquakes. quake in this area with an intensity of 7 to 7.9 on the · Fatalities in the Middle East, North Africa, and Turkey Richter scale. At the same time, the populations of these amounted to 34,219 (79 percent of the total). cities are expected to increase (table 11.3).6 · About 82 percent of fatalities were from earth- quakes (35,737), half from two events in Turkey. · About 47.5 percent of all disaster-affected people The Flood in Algeria in November 2001 were affected by drought, though more than half of this was the result of a single event--a 1995 drought In North Africa, Algeria was hit most severely by dis- in Spain. The number of affected persons was high- asters (36 events, 4,124 fatalities, and 1,154,355 affected est for Southern Europe, followed by the Balkans, people), followed by Egypt (14 events, 1,286 fatalities, the Eastern Mediterranean and North Africa. and 289,342 affected people), and Morocco (23 events, · Of 38 floods with 2,924 fatalities that hit North Africa 919 deaths, and 442,973 affected people). Most fatali- from 1975­2001, 921 persons (31%) died in the ties in Algeria were due to earthquakes (2,881) and floods flood that struck Algeria in 2001. (1,201); in Egypt, there were 673 flood victims and 561 The 1999 earthquake in Turkey and the 2001 flood earthquake victims in Cairo in 1992. In Algeria, earth- in Algiers were among the most fatal natural events, quakes affected 1,001,212; in Morocco, drought affected where vulnerability was much higher than in Southern 275,000, and in Egypt, floods affected 229,868. The flash Europe. To illustrate this argument, these two recent floods that struck Algiers between November 9­13 cases are briefly reviewed below. caused 921 deaths (IFRC 2002), affected 50,423 people, Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 159 and caused economic losses in the range of US$300 impacts include: a) increased unemployment that million. According to UNICEF, "the floods have affected increases immigration pressure on European countries; over 10,000 families or 40,000 to 50,000 persons, two- and b) decreased water resources that increase friction thirds of whom are children."7 The factors contribut- among countries sharing the same water resources, lead- ing to the high vulnerability and high number of fatalities ing to political unrest. refer to high population density, poorly-built housing These studies have all projected severe consequences in flood-prone regions, severe administrative errors, and for Alexandria from sea-level rise. Due to high popula- lack of implementation of building standards in one of tion density, Alexandria and Cairo are highly vulnera- the poorest city sectors. ble to earthquakes and floods. Alexandria has been On August 8, 2002, the World Bank approved a US$89 projected to grow by 820,000 people until 2015. Food million loan to Algeria,8 and on July 24, 2002, the EIB self-sufficiency will decline and import needs will rise granted Algeria a loan of EUR 165 million for recon- significantly (Brauch 2002a, 2002b). structing flood-damaged infrastructure.9 According to urbanization projections (table 11.3), Algiers is projected to grow by 50 percent from 2000 to 2015 (2000: 2.76 Conclusions: Urbanization and Disaster million, 2015: 4.14 million). While population growth Preparedness and urbanization contribute to increased vulnerability, climate change may increase the impact of severe weather Drawing Lessons from the Case Studies events, especially in the Middle East and North Africa. The statistical analysis and the two case studies on rapid- onset disasters and projected slow-onset disaster in the southern and eastern Mediterranean have indicated that: The Projected Impact of the Sea-level Rise · Rapid urbanization has increased and will further for Alexandria increase the vulnerability to all types of disasters, especially for the poor living in informal housing and Egypt will be severely affected by the projected sea-level in flood-prone areas. rise in the Nile Delta and in its coastal regions. Sestini · The probability and intensity of hydro-meteorological (1992) analyzed its implications until 2020. In his view, disasters have been projected to increase due to climate "the socioeconomic structures of the Lower Nile Delta change impacts. Rapid urbanization may further probably will be affected more by population increase increase the vulnerability to all disasters and fatalities and urbanization, than by climatic changes" (Sestini and affected people in the years to come. 1992: 535­536). To counter the high fatalities in MENA countries, a Mohammed El-Raey (1991, 1993, 1994, 2000) dual effort is needed: a) a reduction in human vulner- produced a Vulnerability Assessment of the Coastal Zone ability; and b) a reduction in disaster impacts. of Egypt to the Impacts of Sea Level Rise, which concluded: "that a 0.5 meter sea level rise would cause migration of more than 2.0 million people, loss of more than 214,000 International Strategy of Urban Disaster jobs and a value loss of more than US$40.0 billion, Reduction mainly in Alexandria Governorate."10 The reports con- tained detailed vulnerability assessments for Alexandria The IPCC (2001a) observed significant changes in (El-Raey 1997), Roseta, and Port Said (El-Raey 1997a). extreme weather events and hydro-meteorological dis- El-Raey expects these impacts within the governorate asters globally. In response, international regimes and of Alexandria, where 40 percent of Egyptian industry epistemic communities (Haas 1993) dealing with dis- is located: a) increased vulnerability of slum areas to wind asters have emerged after the International Decade on and flood damage; and b) changes in the frequency, timing, Natural Disaster Reduction (IDNDR), especially the and duration of heat waves that will affect agricultural International Strategy on Disaster Reduction (ISDR) and yields, and increase insect pests. Secondary regional its Inter-Agency Task Force for Disaster Reduction (IATF), 160 Building Safer Cities: The Future of Disaster Risk the ProVention Consortium, the emergency and disaster- and Security (GMES)-initiative of the European Com- reduction efforts of UNEP and UNDP, the initiatives by mission and the European Space Agency could provide NATO's Euro-Atlantic Disaster Response Coordination data for a joint Euro-Mediterranean regional monitoring Center, and efforts by the EU in the area of civil protec- of the causes contributing to natural disasters (land-use tion of its member countries on behalf of the European changes, soil erosion, urbanization) and to the increase in Space Agency. vulnerability, especially of urban centers in the Middle East and North Africa (Sari in Brauch and others 2003). In the EMP context, experts from all Mediterranean Mediterranean Strategy of Urban Disaster countries should join training exercises to cope with Reduction disaster impacts. A Mediterranean strategy for disaster prevention Dealing with disasters in the Mediterranean as a common (MSDR) could be launched by the ISDR, and a Mediter- regional problem has been impeded because this space ranean Inter-Agency Task Force for Disaster Reduction is institutionally separated among three continents. (MIATF) could closely coordinate the regional efforts of Efforts at disaster reduction have been launched in the existing United Nations, European Union, Arab, and framework of the Euro-Mediterranean Partnership (EMP), other functional institutions to enhance cooperation, and on the scientific level, information networks (MEDIN) research, and training for disaster reduction. Humani- and functional organizations exist that involve some tarian organizations such as the International Federa- non­EU Mediterranean member countries in disaster tion of the Red Cross and Red Crescent Societies response, preparedness, and reduction activities. In (IFRC-RCS) and the industry could play a role in the Mediterranean these efforts are highly fragmented developing affordable insurance schemes for those coun- (Brauch and others 2003). Many EU-sponsored research tries that have become repeated victims of natural and projects, for example on climate, land-use changes, technological disasters. desertification, and urbanization have focused prima- rily on Southern Europe. Two different regional concepts could overcome exist- From Disaster Response to Disaster Reduction ing deficits: the geographic perspective of the Mediter- raneanspacecontainedintheBarcelonaConvention(1976), A disaster-impact assessment and disaster preparedness and in the Mediterranean Action Plan (MAP) that includes should become inherent goals of all development proj- all riparian countries, and the Euro-Mediterranean space ects by international financial institutions (e.g., in the of the Barcelona Declaration (1995) that includes all 15 Mediterranean Technical Assistance Programme: METAP), EU countries and 12 dialogue partners. In October 2001, by U.N. institutions, and in the EMP-framework. Risk- the 27 Euro-Mediterranean foreign ministers welcomed reduction measures should be included in regional efforts "the progress made within the Partnership with regard for environmental management, land-use planning, pro- to a system for preventing, alleviating the effects of and motion of improved building standards, and joint mon- managing disasters" and in the Valencia Plan of Action itoring of their proper implementation. of April 24, 2002, they recognized the "contribution and the experience accumulated by the pilot project on mitigation of natural or manmade disasters." In the Athens Reducing Vulnerability to and Impact of Disasters Declaration of July 10, 2002, the synergies between MAP and SMAP were stressed (Brauch and others 2003), The two basic strategic goals of a Mediterranean strat- but references to disaster reduction were missing. The egy of disaster reduction require: pilot project in the EMP context of the EU's Directorate- · A reduction in the vulnerability of human beings and General External Relations and of civil protection efforts economic losses to disasters; and of its Directorate-General Environment could be linked · A reduction in the impact of geophysical and hydro- moreclosely.ThejointGlobalMonitoringforEnvironment meteorological disasters. Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 161 Reducing the vulnerability to disasters of urban implemented, depends on the worldviews and mind- centers in the Mediterranean requires a deliberate sets that determine their political priorities (figure 11.1). strategy of poverty eradication and sustainable devel- Whether governments in the Middle East and North opment. The EMP (EU) and MAP (UNEP) and inter- Africa are willing to shift from a hard to a soft security national financial institutions in METAP offer existing agenda, with human security perspectives on environ- institutional frameworks. Disaster-reduction goals must mental security issues (including disasters), will depend be integrated into all national, regional, and urban devel- on the resolution of ongoing conflict. opment plans. This presupposes an enhanced knowl- edge of risk factors for the whole Mediterranean region for: a) social, economic, physical, and environmental Notes vulnerabilities; and b) all types of hazards and disas- 1. The "Urban Environment Information gateway" of UNEP, ters. Societal and political efforts, a higher public com- GRID, Arendal: "Cities Environment Reports on the Internet mitment, improved urban risk assessments, improved (CEROI)," at: does not offer any data pan-Mediterranean early warning systems, and improved on the Mediterranean area. Its extensive list of indicators does national and urban preparedness would permit a rapid not include any disaster-related indicators and could not be used and effective disaster response. In both case studies, for this paper. major legislative, administrative, and operational deficits 2. While both tables rely on the EM-DAT database, a major dif- were noted. Technical measures can contribute to better ference in both tables applies to data on Syria. disaster preparedness: a mapping of disaster-prone urban 3. See: EU Publication Office: "November 2001 EU topics and regions and the development of specific building codes new publications-Archive," at: : 5-6 of 10, accessed at 23 July 2002. urban areas are expected to continue to grow in the The full report can be downloaded at: . becomes an essential prerequisite to avoid negative 4. World Bank: "World Bank Provides Earthquake Assistance to impacts on humans and the environment. Reducing Turkey," News Release No. 2000/094/ECA, 16 November vulnerability requires addressing population growth 1999, at: . and reducing the impact of climate change and extreme 5. European Investment Bank: "EUR 450 million EIB support weather events, including the projected sea-level rise, for earthquake rehabilitation and reconstruction in Turkey," Press though remedial action requires a concerted global effort release 2000/003, 9 February 2000, at: . As long as policymakers ignore the linkage between 6. Hanna Ruitshauser and Kemal Gök. Zerstöre, töte die Stadt. the causes of environmental stress and effects and the Istanbul erwartet ein grosses Erdbeben in: Neue Zürcher Zeitung most likely outcomes in terms of disasters (figure 11.4), 17/18 August 2002: 61-63. no major progress can be achieved to implement the 7. See: UNICEF, 20 November 2001, at: . declared goals of sustainable development. The require- 8. World Bank: "Algeria: World Bank Loan To Reduce Impact ments of the 1998 Cardiff process to include environ- Of Natural Disaster On Urban Poor," News Release No: 2002/ mental concerns in all sectoral policies provides a political 055/MENA, 8 August 2002. at: , lever to promote the goals of disaster reduction in the search: "Algeria"; World Bank: "Shielding Algeria from Disaster. framework of the EMP and to realize the goals the 27 World Bank loan will reduce the impact of natural disasters on Foreign Ministers from EMP countries adopted in the urban poor", see also for detailed technical annex and environ- 2002 Valencia Plan of Action for "ensuring sustainable ment impact assessments. development with a high degree of environmental 9. European Investment Bank: "Algeria EIB lends EUR 165 mil- protection."11 Whether the dual challenges of both lion for post-flood reconstruction, the Greater Algiers urban road increasing vulnerability and impact are perceived by network and a section of the East-West motorway," Press the policymakers and the public, and timely and release 2002/061, 24 July 2002, at: . 162 Building Safer Cities: The Future of Disaster Risk 10. See: M. El-Raey: "Egypt: Coastal Zone Development and Cli- El-Raey, Mohammed. 1993. "Vulnerability of the Coastal Zones." mate Change Impact of Climate Change on Egypt," . 1993. 11. See the text at: < http://ue.eu.int/pressData/en/ec/71025.pdf>. El-Raey, Mohammed. 1994. Environmental Refugees Case Study: Egypt. Alexandria: University of Alexandria, Department of Environmental Studies. Bibliography El-Raey, Mohammed. 2000. "Egypt: Coastal Zone Development Abramovitz, Janet N. 2001. Unnatural Disasters. Worldwatch and Climate Change Impact of Climate Change on Egypt." Avail- Paper 158. Washington, D.C.: Worldwatch Institute. able from . Annan, Kofi A. 2001. "Foreword." In Rob Mc Rae, and D. Hubert, El-Raey, M., S. Ahmed, and E. Korany. 1997. "Remote Sensing eds., Human Security and the New Diplomacy. Protecting People, and GIS for Vulnerability Assessment of the Impact of Sea Promoting Peace. Montreal­Kingston­London­Ithaca: Mc Gill­ Level Rise over Alexandria City and Vicinity." Alexandria, Queen's University Press: xix. Egypt: Alexandria University, Department of Environmental Brauch, Hans Günter. 2000/2001. "Environmental Degrada- Studies, Institute of Graduate Studies and Research. tion as Root Causes of Migration: Desertification and Climate El-Raey, M., O. Frihy, S.M. Nasr, and K. Dewidar. 1997a. Change. Long-Term Causes of Migration from North Africa "Vulnerability Assessment of Sea Level Rise over Port-Said to Europe." In Peter Friedrich and J. Sakari, eds., Policies of Governorate, Egypt." Submitted to Journal of Environmental Regional Competition. Schriften zur öffentlichen Verwaltung Monitoring. und öffentlichen Wirtschaft, Vol. 161. Baden-Baden: Nomos: European Environment Agency. 2001. Sustainable Water Use in 102­138. Europe, Part 3: Extreme Hydrological Events: Floods and Droughts. Brauch, Hans Günter and Antonio Marquina, eds. 2001. The Environmental Issue Report, No. 21. Copenhagen: European Mediterranean Space and its Borders. Geography, Politics, Eco- Environment Agency. nomics and Environment. Collection Strademed 14. Madrid: Gleditsch, Nils Petter. 2003. "Environmental Conflict: Neoma- UNISCI - Mosbach: AFES-PRESS: 25­58, 109­144. lthusians vs. Cornucopians." In Hans Günter Brauch and Brauch, Hans Günter. 2002a. "A Survival Pact for the Mediter- others, eds., Security and Environment in the Mediterranean. ranean: Linking `virtual water' and `virtual sun.'" In TERI, eds., Conceptualising Security and Environmental Conflicts. Berlin- Global Partnerships for Sustainable Development: Harnessing Heidelberg: Springer, forthcoming. Action for the 21st Century. New Delhi: TERI, forthcoming. Haas, Peter. 1993. "Epistemic Communities and the Dynamics Brauch, Hans Günter. 2002b: "Climate Change, Environmental of International Environmental Cooperation." In V. Rittberger, Stress and Conflict." In BMU, ed., Climate Change and Con- ed., Regime Theory and International Relations. Oxford: flict. Berlin: Federal Ministry on the Environment, Nature Con- Clarendon Press: 168­201. servation and Nuclear Safety. Heilig, Gerhard A. 1998a. DemoGraphics '96. A population edu- Brauch, Hans Günter, M. Selim, P.H. Liotta, B. Chourou, and cation software developed for the United Nations Population Fund. P. Rogers, eds. 2003. Security and Environment in the Mediter- Vienna: Heilig. ranean. Conceptualising Security and Environmental Conflicts. Heilig, Gerhard A. 1998b. DemoTables '96. Vienna: Heilig. Berlin-Heidelberg: Springer, forthcoming. Hewitt, Kenneth. 2002. "Natural Hazards." In T. Munn, ed., Bull, Hedley. 1977. The Anarchical Society. A Study of Order in Encyclopedia of Global Environmental Change, vol. 3. Chichester: World Politics. New York: Columbia University Press. John Wiley: 479­493. Buzan, Barry, O. Wæver, and J. de Wilde. 1998: Security. A New Ichikawa, Atsushi. 1995. "Coping with Urban Disasters." In Framework For Análisis. Boulder - London: Lynne Rienner. OECD Observer 197: 15­16. El-Hinnawi, E. 1985. Environmental Refugees. Nairobi: UNEP. IDNDR (International Decade for Natural Disaster Reduction). El-Raey, Mohammed. 1991. "Responses to the Impacts of 1996. Cities at Risk. Making Cities Safer Before Disaster Strikes. Greenhouse-Induced Sea-Level Rise on Egypt." In J.G. Titus, Geneva: IDNDR. ed., Changing Climate and the Coast, vol. 2: West Africa, the IFRC (International Federation of the Red Cross). 2001. World Americas, the Mediterranean Basin, and the Rest of Europe. Disasters Report 2001. Focus on Recovery. Bloomfield, C.T.: Washington, D.C.: UNEP and USEPA. Kumarian­London: Europan: 186­197. Urbanization and Natural Disasters in the Mediterranean: Population Growth and Climate Change in the 21st Century 163 IFRC (International Federation of the Red Cross). 2002. World Munich Re. 2000. Topics­Annual Review of Natural Catastrophes Disasters Report 2002. Focus on Reducing Risk. Bloomfield, CT: 1999. Munich: Munich Re. Kumarian­London: Europan: 197­203. Myers, Norman. 1995. Environmental Exodus. An Emergent IPCC (Intergovernmental Panel on Climate Change). 2001a. Cli- Crisis in the Global Arena. Washington, D.C.: Climate Institute. mate Change 2001. The Scientific Basis. Cambridge­New York: Newman, Edward and O.P. Richmond, eds. 2001. The United Cambridge University Press. Nations and Human Security. Basingstoke­New York: Palgrave. IPCC (Intergovernmental Panel on Climate Change). 2001b. Cli- Pachauri, R.K. 2000. "Environmental Security: A Developing mate Change 2001. Impacts, Adaptation and Vulnerability. Miti- Country Perspective." Meeting report of the Wilson Center, gation. Cambridge­New York: Cambridge University Press. Environmental Change and Security Project, Current Events, Kam, Ephraim. 2003. "Conceptualising Security in Israel." In October 17, 2000. Available from: . ment in the Mediterranean. Conceptualising Security and Envi- Rakodi, Carole, ed. 1997. The Urban Challenge in Africa. Growth ronmental Conflicts. Berlin-Heidelberg: Springer, forthcoming. and Management of its Large Cities. Tokyo­New York­Paris: Lo, Fu-Chen and Y. Yeung, eds. 1998. Globalization and the World UN University Press. of Large Cities. Tokyo­New York­Paris: UN University Press. Sestini, G. 1992. "Implications of Climatic Changes for the Nile Lomborg, Bjørn. 2001. The Skeptical Environmentalist. Measur- Delta." In Lubomir Jeftic, J.D. Milliman, and G. Sestini, eds., ing the Real State of the World. Cambridge­New York: Cam- Climatic Change and the Mediterranean. Environmental and Soci- bridge University Press. etal Impacts of climate Change and Sea-level Rise in the Mediter- Lutz, Wolfgang, ed. 1994. The Future Population of the World. ranean Region. London­New York­Melbourne­Auckland: What Can We Assume Today? London: Earthscan. Edward Arnold: 535­601. Lutz, Wolfgang, ed. 1996. The Future Population of the World. Smith, Keith. 2001: Environmental Hazards. Assessing Risk and What Can We Assume Today? Revised and Updated Edition. Reducing Disaster, 3rd ed. London­New York: Routledge. London: Earthscan. UNCHS Habitat (United Nations Human Settlements Programme). Lutz, Wolfgang and A. Goujon. 2002. "Demographic transition." 2001a: Cities in a Globalising World. Global Report on Human In Ted Munn, ed., Encyclopedia of Global Environmental Change, Settlements 2001. Nairobi: UNCHS. Vol. 4: Peter Timmerman, ed., Social and economic dimensions UNCHS Habitat (United Nations Human Settlements Programme). of global environmental change. Chichester: Wiley: 211­214. 2001b. The State of the World's Cities 2001. Nairobi: UNCHS. Mc Evedy, Colin and R. Jones. 1978. Atlas of World Population UNDP (United Nations Development Program). 1994. United History. London: Allen Lane. Nations Human Development Report, 1994. New York: UN. Meze-Hausken, Elisabeth. 1998. "Mass-migration due to climate UNEP (United Nations Environment Program). 2002. Global change? A conceptual layout." Available from: . 270­296. Mitchell, James K. 1995. "Coping with natural hazards and dis- UNESCAP (United Nations Economic and Social Affairs Pro- asters in megacities. Perspectives on the twenty-first century." gram). 2001. World Population Prospects. The 2000 Revision. GeoJournal 37(3): 303­312. Highlights. Population Division, Department of Economic and Mitchell, James K. 1999a. "Introduction." In James K. Mitchell, Social Affairs, UN Secretariat, New York. ed., Crucibles of Hazard: Mega-Cities and Disasters in Transition. UNESCAP (United Nations Economic and Social Affairs Pro- Tokyo­New York­Paris: UN University Press: 1­14. gram). 2002. World Urbanization Prospects. The 2001 Revision. Mitchell; James K. 1999b. "Findings and Conclusions" In Data Tables and Highlights. UN Population Division, Depart- James K. Mitchell, ed., Crucibles of Hazard: Mega-Cities and ment of Economic and Social Affairs, UN Secretariat, New Disasters in Transition. Tokyo­New York­Paris: UN University York. Press: 473­502. UNHCR (United Nations High Commissioner for Refugees), Munich Re. 1998. World Map of Natural Hazards. Munich: IOM (International Organization for Migration) and RPG Munich Re. (Rassemblement du Peuple Guineen). 1996: Environmentally- Munich Re. 1999. Topics­Annual Review of Natural Catastrophes Induced Population Displacements and Environmental Impacts 1998. Munich: Munich Re. Resulting from Mass Migrations. Geneva: IOM. 164 Building Safer Cities: The Future of Disaster Risk UN/ISDR (United Nations International Strategy for Disaster World Bank and European Investment Bank. 1990. [What is Reduction). 2001. The Socio-Economic Consequences of the Earth- the publication title?]Washington, D.C.: World Bank; Lux- quake at Izmit, Turkey 17 August 1999. Geneva: ISDR. embourg: European Investment Bank. UN/ISDR (United Nations International Strategy for Disaster Yousry, Mahmoud and T. Atta. 1997. "The Challenge of urban Reduction). 2002. Living with Risk. A Global Review of Disas- growth in Cairo." In Carole Rakodi, ed., The Urban Challenge ter Reduction Initiatives. Preliminary Version. Geneva: ISDR. in Africa. Growth and Management of its Large Cities. Tokyo­New Wagner, Horst-Günter. 2001. Mittelmeerraum. Geographie, York­Paris: UN University Press. Geschichte, Wirtschaft, Politik. Darmstadt: Wissenschaftliche Yousry, Mahmoud, T. Abu-Zekry, and A. Yousry. 1998. "Cairo Buchgesellschaft. as a world city: The Impact of Cairo's orientation towards glob- Wight, Martin. 1991. International Theory: The Three Traditions. alization." In Fu-Chen Lo and Y. Yeung, eds., Globalization and In Gabriele Wight and Brian Porter, eds. Leicester: Leicester the World of Large Cities. Tokyo­New York­Paris: UN Univer- University Press. sity Press: 270­313. Chapter 12 Urban Land Markets and Disasters: Floods in Argentina's Cities Nora Clichevsky This paper provides a preliminary presentation of expe- potable water are already occupied. Flooding is central riences involving the relationship between urban land to these interrelationships. It occurs due to natural markets and flooding in Argentine cities within the wider events: a river overflows or there is intense rainfall in framework of urban vulnerability to disasters such as an environment not prepared to cope with it. flooding, landslides, and erosion. The analysis attempts Urban environment as a particular expression of a to reveal hidden aspects of two key issues: relationship established at a given time and space between · The role played by the state as a land market regu- society and nature has traditionally been considered, lator in urban areas with natural vulnerability to flood- in Latin America in general and in Argentina in partic- ing; as a producer of urban land and housing; and ular, in a partial, sectoral, and inarticulate way, both in as a builder of defenses against flooding. the academic world and in decisionmaking public organ- · Impacts on the land market of flooding and built izations. This subdivision, evident in studies and policy defenses, as well as access to urban land for differ- instruments as well as in the institutional forms adopted ent population sectors; land-value increases and by organizations dealing with urban-environmental decreases in urban areas according to flooding poten- issues, has defined this field as one of dispute more than tial; and whether flooding conditions are masked. of integration. In the few opportunities in which the · Impacts from the urban dynamics of population, pro- intention was to treat these matters in an articulate ductive activities, and the socio-occupational situa- manner, the proposals never went further than talk tions of inhabitants. (Clichevsky 2001b). This is an exploratory study: therefore, the recom- If environmental and urban problems in general have mendations present methodological considerations for hardly been dealt with, land market operations (with future work. The urban conglomerates of Santa Fe, their submarkets, legal and illegal), occupation by Resistencia, and Buenos Aires, and the city of Chas- poor urban population sectors with no access to sub- comús were chosen for this study because they are markets, and "disasters" that are generally such due to affected by recurrent flooding. These areas have built a a high number of affected people in urban centers, number of defenses with differential impacts, and each have been less so. Neither the impacts on land markets has its own urban dynamics. of the above-mentioned disasters, nor the improvements produced by structural and nonstructural measures and their impact on the market and sub-markets, have Disasters and Urban Land Markets: been studied. A Little-Studied Relation The localization of cities and the way in which they The Competition for Urban Space expand have brought about a series of environmental problems that arise from the relation between the nat- The ways in which a city is constructed are directly related ural environment and human populations. Even areas to the operation of land markets, legal and illegal. These prone to flooding, vulnerable to erosion, and with scarce markets and the related socioeconomic changes of the 165 166 Building Safer Cities: The Future of Disaster Risk last decades have resulted in forms of urban segrega- financial sector (Clichevsky 2001a); and it is condi- tion, different from previous ones, that have led to more tioned by other markets such as housing, industry, com- "ghettoized" cities. The land market is the place where merce, and services, generating sub-markets according private and public agents interact with different objec- to the specific demands of urban activities. tives in mind. These have changed during the last few The price of land is the deciding factor for a popu- decades and are affected by the stage in the land devel- lation's access to the portion of urban space where it opment process, such as production and commercial- (or its citizens) can settle: from a legal lot with all ization. At present, the main agents are: land owners; urban services to illegal lots with no services; direct the financial sector; developers (those who foresee busi- occupations (shanty towns, settlements, squatters, and ness and propose it to land owners, financial firms, etc.); squatted equipment) and various categories of slums, real estate firms (some of which are present from the hotels, and boarding houses. It is important to note that beginning of the "business" up to its commercializa- the price of land has little relation to the "production"7 tion); building enterprises; professional teams; the state, price plus the market's average profit. The price has acting at several levels; claimant users; and claimant more to do with rent and the extraordinary benefits that speculators. the various market agents are eager to obtain.8 The formal or legal land market, divided into sub- Historically, most Argentine cities incorporated markets according to types of product, coexists with the land under the logic of "production" using "skipped" informal or illegal market1 and, with direct land occu- subdivisions with the object of keeping interposed lots pation, constitutes different types of informal habitats.2 as a land value reserve9until lots further away were occu- The illegal market comprises clandestine and irreg- pied. The assumption was that demand would increase ular divisions of land into lots that have not been sub- and so would the value of the interposed lots. During mitted to authorities for approval and, therefore, will the past decades, a particular sub-market arose for never be legal unless special exceptions are later low-income population sectors, according to the type issued. Irregular lots3 are those that do not comply of land offered (size, services, and so forth), the price, with existing laws for production and commercializa- and the availability of housing financing, although this tion4 in force at the time of sale. has been disappearing (Clichevsky 2001a). The direct occupation of public and private land gen- At the same time, the new economic processes that erates "villas" (shanty towns) and settlements used as began during the 1990s brought new investment to residential areas that are completely unplanned from an the city through the construction of large residential urban-environmental view. This often includes land: and commercial equipment that covered more than prone to flooding; without infrastructure; with inade- 1,000 hectares. The residential sub-market changed at quate access to employment centers, primary education the same time that social polarization increased. The and primary health care services; with ground water response was new "merchandise": luxury houses for pollution and salinity; with proximity to clandestine high-income population sectors, country clubs, gated garbage dumps and brick kilns; and with housing con- communities, and even "private cities."10 These brought struction that does not meet existing regulations. Ille- environmental territorial changes--even modifying gal settlements have also taken over public areas such topography due to the removal of large quantities of as town squares. earth. As a result, new urban environmental problems The state hardly regulates the legal land market in arose: swamplands were blocked and natural drainage the production stage,5 and it is even less involved in was affected, leading to a greater possibility of flooding the commercialization stage. It operates according to than in the past. solvent supply and demand (according to profits expected There is competition among social population sec- by owners). It is imperfect, with monopolistic or oli- tors for urban space. For example, if there is interest in gopolistic characteristics according to the urban area; investing capital in land that happens to be illegally it is scarcely transparent;6 it is related to the interna- occupied by a low-income population group, efforts tional real estate market that is also connected to the will be made to evict these residents. The state may Urban Land Markets and Disasters: Floods in Argentina's Cities 167 also want to sell land--possibly during an adjustment The extent of risk, in this case for flooding, also depends and reform period--favoring economic groups who upon public and private measures that have been adopted wish to invest in the city rather than using that same to minimize a disaster's negative impact. For example, land to meet the residential needs of lower-income pop- "soft" credit will, in turn, lead to impacts on different pop- ulation groups. ulation sectors demanding land. Structural and nonstructural12 measures taken to minimize flooding risk and the possibility of their imple- Vulnerability, Land Markets, and Flooding mentation produce varying impacts on land prices that, in turn, produce impacts on different social popula- Vulnerability is the propensity to suffer damage due to tion sectors. While some people may benefit, in this the presence of a given force or a potentially destruc- case, middle population sectors who can afford a lot in tive energy. It is the inability to absorb through self- Colastiné, others, such as poor population sectors living adjustment the effects of environmental change due to in the same area who have been displaced and have to inflexibility or the inability to adapt to that change migrate, may be negatively impacted. (Gómez 2001). Vulnerability is a multidimensional concept, generally associated with a condition (one is vulnerable to) and it includes exposure, sensibility, The Selected Argentine Urban Areas and resilience (capability to resist or recover). In the social sciences, vulnerability is used to indicate falling Greater Santa Fe (GSF), including Santa Fe, Santo Tomé, beneath the poverty line; and in environmental stud- San José de Rincón, Recreo, and Sauce Viejo, has a ies, it is used to indicate vulnerability to flooding, among population of some 450,000, according to the 2001 other meanings. census. GSF is located on the right bank of the Paraná The concept of vulnerability implies the acceptance River on a vast plain characterized by difficult water of two fundamental matters: on the one hand, it estab- drainage and a corresponding formation of lakes, streams, lishes a dialectic relationship between the "outside," the and swamplands. It is, therefore, prone to recurrent elements that are external to the social group and may flooding caused by overflow and rainfall due to a lack turn into risk because they are outside the exposed of adequate drainage.13 group's control, and the "inside," the group's cultural GSF's traditional economy is based on government and socioeconomic characteristics (Clichevsky 2001b). services, a university, commerce, and private services, as There is a relationship between natural and social vul- well as port activities and a handful of food industries, nerability. A city is more socially vulnerable because it including breweries. According to the October 2001 is prone to flooding, has no services or equipment, no Permanent Household Survey (EPH), 22.8 percent of safety requirements (in relation to land market opera- the GSF population was unemployed and 14.9 percent tions), and has a degraded environment. The relation- was underemployed. It is important to note that, in May ship among the different scales is fundamental for the 1992, only 9.5 percent was unemployed (INDEC 2001). study of relations between the land market and flood- Income concentration is also important. In 1991, the ing because both have specific macro, micro, and even unmet basic needs (UBN) sector of the population was neighborhood-level determinations.11 16.5percent.Thisvariedfromareatoarea.IntheGuadalupe Flooding (as any other kind of "disaster") also affects neighborhood, the number was only 4 percent, while the land market as a whole (even outside the city center) in Alto Verde, the percentage of the population with in each of its sub-markets, on a larger or smaller scale unmet basic needs was 62 percent (similar to that of according to its dynamics and specific organization Colastiné). In 1980, the UBN population sector in San (i.e., participating agents who tend to be concentrated José del Rincón was 51.2 percent (INDEC 1982). and local). A total paralysis may occur in all or some In the early 1990s, several policies against flooding sub-markets, or there may be only a decrease for a spe- were implemented in Rincón-Colastiné and other neigh- cific period in the prices of more important property. borhoods. Provisional defenses were built through the 168 Building Safer Cities: The Future of Disaster Risk Flooding Emergency Rehabilitation Program (Programa Las Barrancas, El Burro, La Tablita, and Chascomús de Rehabilitación para la Emergencia de las Inundaciones lakes. These waters interconnect by streams or are nar- ­ PREI)14 and structural works financed by the Flood- rowly separated. They operate by quotas like a very slow ing Protection Program (Programa de Protección contra river tributary to the Salado River, gathering local water Inundaciones ­ PPI).15 surpluses and transferring them from lake to lake in a Greater Resistencia (GR) comprises the municipali- southwesterly direction. ties of Resistencia, Barranqueras, Fontana, and Vilelas Significant population growth has taken place since and has a population of nearly 360,000. The area includes 1960 and, according to the 2001 census, there are 30,170 the Río de la Plata Basin (Cuenca del Plata) and has inhabitants. The main economic activity in the area is historically been subject to flooding caused by Paraná tourism, followed by commerce, administration and River overflow. When the river overflows, its waters flow services mainly related to farming and cattle breeding, into the area via the Barranqueras stream and the and industry, though factories have been closing. The Negro River. Intense rainfall also impacts flooding, as population is mainly middle class and in 1991 UBN was does the ways in which the territory is occupied. only 10 percent. Until 1920, the main economic activity was timber, Although Chascomús suffers from recurrent flood- later replaced by cotton until 1960, when the world ing, 2001 saw the most severe flooding in 90 years, cotton crisis affected the region. Following the 1982­83 reaching a contour elevation reading of + 9.36 IGM. In floods, industries closed down and productive activi- 2002, lake overflow caused by upstream rainfall in the ties decreased; however, the population continued to lakes system caused flooding yet again. Although proj- increase. Most administrative, commercial, and service ects for defense works were prepared years ago, none activities are concentrated in Resistencia. At present, have been implemented. there is serious economic stagnation and structural The Buenos Aires Metropolitan Area (AMBA) is the unemployment. In October 2001, 15.9 percent of the result of processes that have operated for more than 400 population was unemployed and 14.4 percent was years, since the city's foundation and its expansion over underemployed (INDEC 2001). Income is strongly con- provincial territory. According to the 2001 census, there centrated, with 20 percent of the population receiving are more than 13.8 million inhabitants, unevenly spread 50 percent of the income, while the remaining 80 throughout a territory consisting of the City of Buenos percent receive the other 50 percent. According to the Aires and 32 municipalities in the Province of Buenos 1991 census, UBN throughout the territory represents Aires.AMBAislocatedonthePampeanplain,anareawith 22.7 percent of the population. little slope toward the Río de la Plata, and is cut through To protect this area from recurrent flooding, several by the important Reconquista and Matanza-Riachuelo defense plans were developed over the past decades. basins, the encased Maldonado, Medrano, White, Vega, During the 1982­83 flooding, a plan that partially andCildañezstreams,andseveralevenmoreminorstreams. protected the urban area16 was carried out since the As Argentina's capital city, Buenos Aires hosts a con- 1983 floodwaters were 8.60 meters over flood level, the centration of all economic activities: industrial, commer- highest for the entire twentieth century.17After the 1992 cial, financial, and administrative. Since the mid-1970s, floods, PREI constructed a series of defenses such as the AMBA as well as the rest of Argentina has gone through Northern Defense (Defensa Norte) and related works; important socioeconomic and political changes that pro- in 1997 when they were taken over by PPI. duced deindustrialization, unemployment, and impov- erishment of its population. In October 2001 (last EPH data; INDEC 2001) unemployment reached 19 percent, Chascomús while underemployment was 16.5 percent. It is impor- tant to note that, in October 1991, unemployment was Chascomús, in the Province of Buenos Aires, is located only 5.3 percent and in 1980, 2.3 percent. Income dis- in the Lagunas Encadenadas System on the Salado River tribution varies within AMBA and is related to poverty depression, integrated by the Adela, Chis-Chis, Vitel, indicators. In 1991, UBN population sectors in the City Urban Land Markets and Disasters: Floods in Argentina's Cities 169 of Buenos Aires represented 5.9 percent, while in some The State and Real Estate Agents neighborhoods they were only 2.6 percent (CEPA-INDEC Historically, land-market agents have felt that their rights 1994). In the municipalities of Berazategui and Merlo, were restricted when a state sector implemented stricter however, they reached 39.1 percent. regulations on land subdivision, as in AMBA during At present, floods due to rain and overflow are the the late 1940s. At present, they are against any legisla- result of urban activities in areas that have always been tion that might constrain their actions on areas prone prone to flooding and of changes in the water system to flooding. It is interesting to mention the case of regime produced by land occupation. Flooding recur- Resistencia: the Real Estate Chamber has accused APA rence and its severity have increased during the last of issuing an "expropriating" regulation about land use few years, affecting greater surface areas and people. compatible with natural aptitudes. From 1985 to March 1998, rainfall and sewage over- flows caused 26 flooding events in the City of Buenos Aires (Herzer and Clichevsky 2001). The Scarce Effectiveness of Public Policies on Land Market Regulation Considerations about the Processes Detected Results have been scarce in relation to the quality of urban land that is produced and occupied as a conse- The urban areas discussed account for multiple inter- quence of the previous aspect and of the state's own relations between space occupation, land markets, and objectives about the ways policies on the land market flooding. Many of the interrelations are masked and are elaborated, approved, and implemented. others are not transparent. Urban-metropolitan area The state as a land market regulator in conglomer- economic dynamics--a present and historically deter- ates with natural vulnerability to flooding has acted in mined situation--and a population's socioeconomic sit- the same way in all of the areas discussed: in Buenos uation will define the ways in which urban expansion Aires it allowed divisions into lots in areas prone to takes over a natural environment. Urban expansion flooding; in the city of Santa Fe, when a regulation was affects the organization of the land market and its rela- issued to stop divisions, these were ratified as excep- tionship with the state, space occupation by different tions. In Greater Resistencia, despite existing legal instru- social population sectors, and vulnerability resulting ments, municipal authorities have not been able to from settlements on land prone to flooding. comply with their role as land use regulators and instead have issued exceptions. The Resistencia City Council, the most important city in this conglomerate, constantly Complex Relations between the State votes for exceptions to its own regulations if they hinder and the Real Market Sector construction proposed by a firm or by a higher state body. Regardless, the exception is implicit (Clichevsky State policies toward land markets are subject to the 1994).18 The natural environment is the outstanding influence the real estate sector, including landowners absentee in regulations issued by the municipalities' and developers, exerts on the state. This is evident in planningareas(orwhenmunicipalitiesarenotautonomous legal market regulations defined by urban planning from their respective provinces, as in the case of the organizations, and all forms of exceptions to them, as Province of Buenos Aires). Legislation is issued on land well as in the definition of "structural" measures in flood- division, use, and occupation as if it were a flat surface, ing defense programs. It permeates all state actions with no geographical irregularities that might influ- regarding land use and occupation and has direct impli- ence the conformation of urban areas. cations on the impact that disasters produce by natural Although cities were founded historically at higher ele- effects--in this particular case, flooding--on popula- vations, later expansions did not take similar measures. tions. It is therefore an aspect that must be considered Thereareevenextremesituations,asinResistencia,where in disaster-management programs. the road network was designed over existing lakes and 170 Building Safer Cities: The Future of Disaster Risk the Planning Code even induced landfills. Furthermore, possibility of acting on land regulation from a compre- regulations forbidding divisions into lots to avoid occu- hensive flooding-risk standpoint. This is evident in the pation of areas prone to flooding have come too late, as case of Santa Fe, where the city defined flooding defense inthecaseofAMBAmunicipalitiessincethe1960s,because regulations for Colastiné and different regulations for urban land prone to flooding had already been approved the municipality of Rincón, though these areas are sup- for division before the regulations came into force. Also, posed to form an urban unit. A similar situation occurs precautionstakenregardingcontourelevationlevelshave in the City of Buenos Aires and AMBA municipalities not been sufficient to contain great floods.19 During the located in the Province of Buenos Aires, where the main last few years, these municipalities have changed regula- basins cross over several administrative jurisdictions. tions so that new investments may be located in areas prone to flooding, with the consequent problems caused The Role of the State as a Land and Housing Producer by changes in drainage, flooding previously safe areas. The state as a land and housing producer has also demon- Urban Planning Organizations Have a Poor strated how little it is concerned with financing hous- Understanding of Nonstructural Measures ing estates in areas prone to flooding, and it still sells in Flooding Prevention Programs land to its occupants through various regularization plans in those areas. This has happened in the Santa Even "nonstructural measures" designed to improve or Fe, Resistencia, and Buenos Aires conglomerates. minimize flooding risks are implemented by water serv- ices organizations (as is the case, for example, of the Provincial Water Administration in Chaco) and not by Land Markets, Direct Occupation, and Risk urban planning organizations. In the best case, these waterserviceorganizationssupportapprovedregulations. The legal land market has been able to operate with The elaboration of planning instruments--or changes very few regulations. Since prices are defined by qual- in the existing ones--to consider new situations regard- ity and other factors, areas prone to flooding are cheaper ing flooding and defense works are in some cases--as and can be purchased by lower-income population in the Resistencia or Santa Fe conglomerates--either sectors. Those who have no access to the legal market carried out by the organization in charge of programs resort to the informal market (with no control on land against flooding, SUPCE Chaco, or jointly between those suitability) or occupy public or private land in the worst organizations and the conglomerate's municipal plan- sectors, from an environmental point of view, those with ning area, SUPCE Santa Fe and the capital city, because a high flooding risk. This has happened in Buenos Aires, the other municipalities do not participate. In the case particularly the Matanza-Riachuelo and Reconquista of Buenos Aires, neither the Director Plan for Hydraulic basins; in Santa Fe, in Alto Verde and the western expan- Regulation (Plan Director de Ordenamiento Hidráulico) sion area; in Resistencia, along the course of the Negro nor the Environmental Urban Plan (Plan Urbano River; and in Chascomús, particularly in the Arroyo Los Ambiental), currently being considered by the legisla- Toldos basin. This increased the exposure of more socially ture, have been jointly--or even partially--carried out. vulnerable populations to the dangers of flooding. Changes introduced into the Planning Code in August Although it is true that in Resistencia and Buenos 2000 have not considered the Director Plan, either. More- Aires, middle population sectors have also settled in areas over, nonstructural measures either cannot be imple- prone to flooding, the consequences are different because mented or can be only partially implemented in they are less vulnerable. For example, in Resistencia, consolidated areas prone to flooding. people can evacuate to the homes of family and friends in higher neighborhoods. Poor population sectors must Absence of Metropolitan Management Organizations normally be evacuated to state or NGO shelters. Risks The lack of metropolitan management organizations in in terms of health care are also different for each social the conglomerates under analysis interferes with the sector. Urban Land Markets and Disasters: Floods in Argentina's Cities 171 Land Prices, Flooding, and its Masking this area believing that the defenses would prevent flood- ing. But defenses built against flooding from overflow Changes in Land Prices do not improve the situation of floods caused by rain- fall. This situation was not taken into account when the Regarding land value decreases for areas prone to flood- land was subdivided. People now feel that they have been ing, there is a difference in land destined for low-income cheated. The state is also responsible for this situation population sectors (areas with other environmental dete- due to lax regulations, as in the case of Rincón, or to rioration components or unfit to become urban) and exceptions granted, as in the case of the City of Santa Fe. that destined for middle-income sectors or productive Flooding processes are also masked in AMBA, par- activities. Land on the Salado Basin, in Santa Fe, or on ticularly in some Buenos Aires neighborhoods inhab- the Matanza-Riachuelo and Reconquista basins, in Buenos ited by middle- and high-income population sectors Aires, where a high percentage of UBN population sec- and where economic activities--mainly commercial tors are located and where prices have always been rel- ones for middle sectors--are located. A good example atively low, is different from land on the Vega or White of this situation is the case of the commercial and streams in the city of Buenos Aires, where middle- and partly residential area of Belgrano on Avenida Cabildo, middle-high-income population sectors and commer- and, especially, in the residential area of Belgrano and cial activities are located, even though the areas are prone Núñez, on Avenida del Libertador, periodically flooded to flooding. In these cases, the risk of flooding does not by the Vega and White streams (Herzer 2001). The flood- imply a decrease in land value. ing situation is masked and property does not suffer As a hypothesis, we could state that land prone to land value decrease (except during the time when flood- flooding, located in areas with better access and his- ing is severe, though once that passes, prices rise to torically valued by society (as the Buenos Aires North those found in areas that do not flood). This has been area), undergoes a smaller price decline than land in proven through surveys on the price of flats (since this areas that are socially stigmatized20 (as the Matanza- area is completely occupied and there are practically Riachuelo and Reconquista basins in Buenos Aires). This no vacant lots) on Avenidas del Libertador, Monroe, and relates directly to whether flooding is or is not masked, Cabildo, as well as on adjacent streets. as described further on. It is interesting to note the case of Chascomús, because A study on hedonic prices carried out on the Arroyo its situation is different from those of the other cities in Maldonado basin, also in Buenos Aires, where impov- this study. Chascomús has no permanent defenses. erished middle-income sectors and some poor popu- The municipality builds them on an ad hoc basis with lation sectors have settled, established that lots with two the help of neighbors and intermediate associations on years of recurrent flooding only devaluated by approx- the coastal avenue where a large number of commer- imately 30 percent. This may be due to the fact that cial, mainly tourist, activities are located. The owners this stream's area is at present an urban border between mask the existence of floods, withdrawing all defenses two zones and that middle-income population sectors as soon as the water recedes, in order to avoid a nega- prefer to settle in other city areas. tive image with tourists. Since part of the area's indus- trial activities have closed down, tourism is one of the Flooding Masking according to Different fundamental activities in Chascomús. The real estate Social Sectors market has also denied that country clubs--some of Flooding is masked by sectors involved in land mar- them already built, others only in blueprint--are in kets, particularly owners and real estate agents. In flood-prone areas (although some of them do not as Colastiné and Rincón, in the GSF, rain floods the area yet have proper authorization to be sold, they were despite defenses because the lack of a sewage system is offered and sold until the present economic crisis pro- added to atypical climatic phenomena.21 This situation duced stagnation in this sector). was masked by real estate agents when lots for middle Resistencia seems to be the only conglomerate in this population sectors were divided. People bought land in study where flooding is not masked. For example, during 172 Building Safer Cities: The Future of Disaster Risk flooding periods, advertisements in all of the local papers produce lots for poor people because, according to (at least through 2001) used expressions such as "high" regulations in force, they must build infrastructure works and "does not flood." This may have changed when the that cannot be afforded by certain large population Northern Defense was finished in 2001, but it is hard sectors (especially during these last few years of reces- to know for sure because of the short period of time sion and a high increase in unemployment rates). It that has elapsed and because of the economic crisis. seems, therefore, that the land value increase in the pro- To summarize, middle- and high-income population tected GR area (prices have gone up approximately four- sectors apparently mask flooding to avoid possible land fold since the defense was finished) benefits the original and property value decreases. They also want to avoid a owners and agents who bought land expecting a land- decrease in economic activities, especially in popular value increase--in the long term, because of present commercial zones like the Avenida Cabildo in the City day recession--after the construction of the defense. of Buenos Aires or in Chascomús in heavy tourist areas. In GSF, during an income concentration period when It is also true that after the flooding is over, activities return land was considered a lucrative business, the course of to normal, similar to areas not prone to flooding. We defenses in Colastiné-Rincón was "shifted" to increase could state, therefore, that areas destined for middle- and thenumberoflotsinsidetheenclosure.Therewasdemand high-income population sectors have a higher resilience. for weekend houses (even for permanent housing in an But, we must also bear in mind that urban processes are accessible location), as well as for individual lots and very complex--as are those that generate land prices-- country clubs. Defenses were even built to include land and it is therefore necessary to consider a large number prone to flooding that should have been left out. The of intervening factors, such as the already-mentioned land-value increase that resulted from the defense defined ones of stigma or social value in historical terms, envi- not only a change in land use, but also a change in the ronmental aspects, accessibility, and state investment. social sectors that might settle in the area. In Colastiné Norte, Bolivian horticulturists had to leave the area, and a gated community was built.22 Poor population sectors The State as Flood Defense Builder are being evicted from Colastiné Sur, and they will prob- ably be relocated on the western side of the city of Santa The state as a builder of flood defenses has promoted a Fe, on land prone to flooding. The same is happening in changeinsomeconglomerates'expansionareas,forexam- the neighborhood of La Boca, in Buenos Aires, due to ple, in Resistencia and Santa Fe. The Northern Defense price increases produced by the construction of defenses in Resistencia leaves a vast portion of land inside the to prevent flooding caused by southeastern winds enclosure,anditsoccupationwilltakealongtime,depend- ("sudestada") over the Río de la Plata in an area located ing on the kind of urban regulations for land markets that next to the city center and with a high tourist value. The will be implemented (at present there are none) and the neighbors--mainly low-income population sectors--are urban dynamics that develop. But this vast enclosure forced by real estate investors to leave their dwellings (in raises some questions: Who benefits from lots that do not many cases, slums). Although defenses were finished in flood? What happens to the poor population sectors in 1998, the general economic crisis and that in the real Resistencia? What will happen to the vast area enclosed estate sector in particular, which began a few years later by the Northern Defense that has no urban use? Will the and has become worse, has partly brought this process market react by dividing land into lots for that popula- to a standstill (Herzer and others 2001). tion sector? To date it has not, because in 2001, when the building of the Northern Defense was finished, poor population sectors were putting up settlements on land Vacant Land and Flooded Populations: A Paradox as their only chance to settle in the conglomerate. This means that the question does not focus on the One of the paradoxes of this subject is the vacancy of existence of land fit according to natural conditions, but a large amount of land suitable for residences and urban to socioeconomic conditions: land market agents cannot activities and the occupation of land prone to flooding. Urban Land Markets and Disasters: Floods in Argentina's Cities 173 This can only be explained by the operation of scarcely is said to consider metropolitan areas or conglom- regulated land markets. In Resistencia, for example, 850 erates involving more than one legal-administrative hectares are presented as vacant urbanized areas. If these area, it is important to stress the need to establish a were subdivided into conventional lots for individual metropolitan authority to deal with urban space mat- dwellings, the approximate density would reach 100 ters as a whole. inhabitants per hectare, with the possibility of settling · It is essential, particularly at a moment of economic 85,000 inhabitants. Considering an annual average crisis and public adjustment, that the state retrieve, growth rate of 1.64 percent, the city should limit all even partially, the land-value increase produced by its urban division extension projects. (Scornik 1998). In actions--through defense works and/or regulations-- Santa Fe, there are also large extensions of vacant land, that are seized by the real estate sector's private agents. and in Chascomús some divisions have not yet been This may be implemented through the payment of occupied. In AMBA there are enough vacant lots to settle a land-value increase tax, as is done in other Latin people in areas that are safe from floods (1.5 million in American countries such as Colombia.23 1997), but to achieve this, the state must take part in the In-depth studies must also be carried out about the land market with new instruments, and this requires polit- interrelationbetweenlandmarketsandflooding.Forthis, ical will (Clichevsky 2002). it is necessary to consider the time of the study: prior to thecommencementofflood-protectionprograms;during theirdevelopment;andafterwards;toobtainawiderpoint Recommendations of view about the range of possible land-value increases or decreases (for example, through the implementation The following recommendations arise from the con- of nonstructural measures that restrict land uses). siderations developed in this study: · The relationship between land markets and envi- ronmental changes must be further explored (disas- Aspects to be Studied: ters such as flooding are part of them) because it is necessary to identify the bidirectional impacts pro- · In what type of urban area is the program located: duced by both dynamic processes. growing or stagnant? · Present-day climatic changes must be considered · What is the type of flooding: overflow, rainfall, wind? because they imply unexpected and scarcely fore- Recurrence, aggressiveness (a peak that passes), or seeable impacts in most of the cities studied; also, permanency? the need to explore in depth the relationship between · How many people and which sectors does the pro- those changes and flooding; even more so when, in gram impact? spite of advances made during the last few years, there · What are the various state policies regarding land is little information as to how some of the areas and housing? prone to flooding operate. It is, therefore, necessary · How is the local land market organized? What are to produce new studies to clarify these questions. its relations with the regional, national, and inter- · Much is still unknown about the legal land market national ones (if any)? What are its relations with operation, its interrelations with the illegal market in agents? What are the sub-markets, the type of demand, each urban area, and the complex interrelations with and its dynamics? othereconomicsectors,particularlyfinancialones,nec- · What is the relationship between land markets and essary for urban policy decisionmaking. An in-depth other economic sectors? Relationships with macro- analysis is absolutely necessary, although it is known economic aspects? Is the real estate crisis linked to that there will always be a nontransparent area, prac- inflationary periods, to currency devaluation, etc.? tically impossible to unveil (for example, where does · What will the prices be before, during, and after the investment capital originate: from financial capital; program's implementation (both of structural and narcotics; the weapons market?). Although every study non-structural measures)? 174 Building Safer Cities: The Future of Disaster Risk · Whatwillbetheimpactofthoseprices(analyzedaccord- merely the opening of streets to including potable water, sew- ing to sub-markets) on different population sectors? erage, public and domiciliary lighting networks, and paving. · What are the growth and economic dynamism sce- 8. Starting with the original landowners; market operations narios for the conglomerate/city to be studied? become more complex with the incorporation of other actors, · What intervention instruments are proposed in land and the rent they produce benefits not only the owner, but also markets to strengthen flooding prevention; particu- the developer, the promoter, etc. Agents try to obtain maxi- larly, to implement measures for land markets so that mum benefits with minimum possible investment, and this the more vulnerable population sectors may improve defines the way in which they stand vis-a-vis the state: they obtain their situation? If these are not included in the plan- "favors" regarding compliance to regulations, or they urge, accord- ning, society's more privileged sectors will again ben- ing to their own objectives, the urbanization of certain plots of efit, reproducing and even extending the vulnerability land, or the granting of certain occupancy factors and uses in of already vulnerable social population sectors. areas forbidden by regulations. 9. Obviously, this is not the only explanation for "skipped" sub- Notes divisions; they also depend on who the landowners are and on 1. The legal and illegal markets are not dichotomous but com- each landowner's logic (Clichevsky 1975). plementary, even juxtaposed, and with scarcely defined limits. 10. Also, changes in the city's industrial, commercial, hotel, and The agents acting in one or the other may be different, or par- private recreation sectors have meant a different demand on land. tially or completely the same, according to the place and his- 11. The impact of different "natural disasters," such as landslides, torical moment. hurricanes, and flooding, is also different according to the eco- 2. Illegality/informality imply two different means of transgres- nomic and social dynamics of the urban center being analyzed. sion regarding ownership characteristics and urbanization The way in which disasters affect a city will be different than the processes. The former is based on a lack of property deeds (or impact on land markets, which will vary, depending on whether rental contracts); the latter on non-fulfillment of building the affected areas are high-, middle-, or low-income residential standards. areas (to simplify) or whether they are commercial or industrial 3. They could become legal if they were to comply with regu- ones. In this case, the use of the concept of social vulnerability is lations over a fairly long period of time. fundamental for the analysis of flooding impacts; the land market 4. Illegal commercialization may be established also by the type response to that impact produces, in turn, a chained impact. of selling transaction (multiple sales of the same lot, for exam- 12. The nonstructural measures constitute the environmentally ple). Also, rural land that cannot become urban because of its positive actions par excellence, because they aim at maintain- location has been sold as undivided plots. Cooperative prop- ing the natural regime of bodies of water. But, it is still difficult erty is built on lots that are smaller than the size established by to implement such measures (due to poor planning on the part law for land subdivision. of the planning specialists, and lack of institutional coordina- 5. The production process includes the necessary steps that must tion), and this is particularly impossible in already occupied be taken to turn rural land into urban land. According to state urban areas, where only structural measures can be adopted. legislation in force in each urban area and at each historic moment, The analyzed urban areas account for this situation. urban land may range from something that only has a subdivi- 13. The 1982­83 swelling presented, between December 1982 sion plan with no services or infrastructure to something that and August 1983, 5 successive peaks with a maximum of 7.33 m. has all urban services. in Puerto Santa Fe. Defenses in the neighborhoods of El Pozo 6. For example, it is difficult to know who the owners are because and La Guardia were surpassed and the latter was completely they use figureheads; the real price of transactions; the invest- flooded. Colastiné Norte was also flooded. The last important ments made to "urbanize" rural land. There is no transparency flooding before the present defenses were built occurred in 1992, in information; private and state owned land cadastres are only reaching 7.43 m. during 2 months. These defenses endured the partially updated, and some cities do not even carry registers. 1998 swelling, which reached 7.24 meters during 6 months. 7. The production price is the price paid for rural land plus all 14. PREI 3521 AR-BIRF. The components are defenses, sanita- subdivision costs (surveying plans) and incorporation of infra- tion, transport and energy, health and education, and housing, structure. According to legislation in force, it may range from as well as technical assistance. Urban Land Markets and Disasters: Floods in Argentina's Cities 175 15. PPI 4117 AR-BIRF-JEXIM. It has structural measures com- Clichevsky, Nora. 1975. El mercado de tierras en el área de expan- ponents (defenses, drainage, adaptation of bridges), housing and sión de Buenos Aires y su incidencia sobre los sectores pop- shelters; nonstructural measures and regional environmental ulares (1943­1973). Centro de Estudios Urbanos y Regionales programs. ­(CEUR), Instituto Torcuato Di Tella (ITDT), Buenos Aires. 16. The GR Paraná River hydro risk frequency was the result of Clichevsky, Nora. 1994. "Mercado de tierra y problemas urbanos a statistical study of hydrometric registers in Puerto Corrientes en ciudades intermedias: réplica del modelo de la gran ciudad? since 1901and in Puerto Barranqueras since 1906 through 1984; Los casos de La Rioja y Resistencia." Revista URBANA 14-15: it shows that, between 1904 and 1984, out of 30.347 readings 75­90. taken at the Puerto Barranqueras hydrometer, the swelling was Clichevsky, Nora. 1996. Política Social Urbana. Buenos Aires: over 6 meters a total of 1264 times. In 970 occasions, the water Espacio Editora. level was between +6 and +7 meters; in 233 events, between +7 Clichevsky, Nora. 2001a. El sector inmobiliario en el Area Met- and +8 meters, and 61 times, over 8 meters. ropolitana de Buenos Aires. Su dinámica en los años noventa. 17. After the 1982­83 floods, the Defense Plan approved in Copy. Buenos Aires. Resistencia intended to turn the city into a large enclosure, Clichevsky, Nora. 2001b. "Pobreza y Políticas Urbano-Ambien- foreseeing a population growth of up to 500,000 inhabitants, tales en Argentina," Doc. De Trabajo, CEPAL. but it was never carried out. Clichevsky, Nora, ed. 2002. Tierra vacante en ciudades lati- 18. The same happened in the City of Buenos Aires, where noamericanas. Cambridge: Lincoln Institute of Land Policy. more than 770 exceptions to the Planning Code were passed Dirección de Estadística y Censos, Provincia de Chaco. 2002. between the 1980s and 1991, even when they were forbidden Censo Nacional de Población, Hogares y Viviendas 2001, by national law (Clichevsky 1996). Unpublished. 19. It is possible that the basin's regime and the recurrence peri- Gómez, José J. 2001. "Vulnerabilidad y Medio Ambiente." Pre- ods were still unknown. sented at the International Seminar Las Diferentes Expresiones 20. Because, historically, they suffer other environmental prob- de la Vulnerabilidad Social en América Latina y el Caribe, lems (proximity to garbage dumps, polluting industries, etc.). CEPAL, Santiago de Chile, June 20­21, 2001. 21. Although there are 18 pumping stations in the enclosed area, Halcrow Consultora. 2001. "Plan Director de Ordenamiento problems arise because fewer than half of them are working, due Hidráulico y control de inundaciones de la ciudad de Buenos to the fact that parts of the systems were stolen from several units. Aires." Proyecto Ejecutivo para la Cuenca del Arroyo Mal- 22. The same happened in the AMBA, toward the end of the donado. Préstamo BIRF N° 4117­AR. Tomo III, Capítulo 8. 1990s, where gated communities were to be built in land occu- Herzer, Hilda. 2001. "La participación de las organizaciones pied by Bolivian horticulturists in the county of Escobar, although sociales." In Alcira Kreimer, David Kullock, and Juan Valdés, in this case, no defenses against flooding were built. eds., Inundaciones en el Area metropolitana de Buenos Aires, 23. This matter has been discussed, as a possible idea only, by Disaster Risk Management Series, N° 3. Washington, D.C.: officials working on the subject in Resistencia. In the City of The World Bank: 166­181. Buenos Aires, it figures in the Environmental Urban Plan Doc- Herzer, Hilda and Nora Clichevsky. 2001. "Perspectiva Histórica: ument that is, at present, being discussed at the legislature for las inundaciones en Buenos Aires." In Alcira Kreimer, David its approval. Kullock, and Juan Valdés, eds., Inundaciones en el Area met- ropolitana de Buenos Aires, Disaster Risk Management Series, N° 3. Washington, D.C.: The World Bank: 33­45. Bibliography Herzer, Hilda, Mercedes de Virgilio, Máximo Lanzetta, Adriana CEPA-INDEC (Comité Ejecutivo para el Estudio de la Pobreza Redondo, and Carla Rodríguez. 2001. "La renovación de La en Argentina). 1993. "Hogares con necesidades básicas insat- Boca ¿Quiénes pierden?" Vivienda Popular 47, September. isfechas­ (NBI) (1980­1991)." Documento de trabajo N.° 3, INDEC (Instituto Nacional de Estadísticas y Censos). 1982. Buenos Aires. Censo Nacional de Población, Viviendas y Familias, 1980. CEPA-INDEC (Comité Ejecutivo para el Estudio de la Pobreza Serie A, Resultados por Localidad, Buenos Aires. en Argentina). 1994. "Mapas de la Pobreza en la Argentina." INDEC.1991. Censo Nacional de Población y Vivienda 1991, Documento de trabajo n.° 4, Buenos Aires. Gran Buenos Aires Serie B No. 2.1, Buenos Aires. 176 Building Safer Cities: The Future of Disaster Risk INDEC. 1992. Censo Nacional de Población, Viviendas y Famil- IPEC (Instituto Provincial de Estadística y Censos). 2001. Informe ias, 1991. Serie A No. 2, Resultados por Localidad, Buenos Aires. Especial N° 100 Resultados Preliminares de la Encuesta Per- INDEC. 1993. Censo Nacional de Población y Vivienda 1991, manente de Hogares en el Aglomerado Santa Fe, Octubre de por Localidad Serie A No. 2, Buenos Aires. 2001, December 13. Gobierno de Santa Fe. INDEC. 1994. Censo Nacional Económico 1994. Resultados IPEC. 2002. Informe Especial N° 004, February 21, Gobierno Provisorios, Buenos Aires. de Santa Fe. INDEC. 1997. Encuesta Permanente de Hogares-EPH, Onda Pizarro, Roberto. 2001. La vulnerabilidad social y sus desafíos: Octubre, Buenos Aires. una mirada desde América Latina, Santiago de Chile, CEPAL INDEC. 1998. Encuesta Permanente de Hogares-EPH, Onda (LC/L,1490-P). Octubre, Buenos Aires. Plan Urbano Ambiental. 2000. Documento Final. Gobierno de INDEC. 2001. Encuesta Permanente de Hogares-EPH, Onda la Ciudad de Buenos Aires, Secretaría de Planeamiento Urbano. Octubre, Buenos Aires. Scornik, Carlos. 1998. Diagnóstico Urbano Expeditivo del Área INDEC. 2001.[CK auth]Censo Nacional de Población, Viviendas Metropolitana del Gran Resistencia (AMGR), PPI-SUPCE, y Familias, 2001, Provisional results. Provincia de Chaco. ch Chur Mitch, clean-up 1998. Hurricane Auxiliadora Community of after Honduras, PART III SOCIAL VULNERABILITY TO DISASTER IMPACTS Chapter 13 Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital Ben Wisner New Research on Old Questions Methods This paper reports for the first time on the results of a In each study city interviews were conducted with comparative study of urban social vulnerability in four citizen-based groups and municipal administrators at the of the world's megacities--Mexico City, Los Angeles, level of the constituent "city" level (e.g., the 23 central Manila, and Tokyo. Two other cities--Mumbai and "wards" of Tokyo, 18 "delegations" of Mexico City's Johannesburg--were also included, but their results federal district, and 26 municipalities of the State of will be analyzed in a separate paper.1 Mexico into which the metropolitan region sprawls to The goal of the study was to elicit and compare from the North, etc.). Descriptions of methodology are both municipal officials and representatives of civil soci- available in Uitto (1998), Velasquez and others (1999), ety definitions of social vulnerability to disasters, sources Takahashi (forthcoming), Wisner (2003a). In addition of information about groups defined as vulnerable, to the published papers and unpublished conference and programs that reduce their vulnerability. The gap papers by these team members cited below, the author between municipality and civil society concerning the has drawn from interim and final project reports.2 understanding of and approach to urban social vul- nerability was the major focus of the study. The main research findings document a paradox. Background: The Growth and Hazardousness Municipalities and civil society appear to have comple- of Megacities mentary strengths and weaknesses. Municipalities have the technical expertise and finances required to provide The second half of the 20th century witnessed the rapid social protection from natural hazards for socially vul- growth of very large cities. There have been primate nerable groups of people and to assist in increasing the cities and metropolises for centuries. However, these capacity of such people for self-protection. However, new urban regions--"megacities" with more than 10 municipalities lack detailed knowledge of these vulner- million inhabitants--are relatively recent. The average able groups and do not enjoy their trust. The situation of size of the world's largest 100 cities increased from 2.1 NGOs and other civil society organizations is the oppo- million in 1950 to 5.1 million in 1990. One-sixth of site. They possess detailed knowledge of vulnerable the world's population lived in cities with more than a groups and they have their trust. Rational action theory million inhabitants in 1990. In that year there were 12 (RAT) would suggest that partnership would therefore cities with more than 10 million people. Worldwide in be in the best interest of both municipalities and civil 1990 there were also 33 cities with five million or society. The paradox arises because the study revealed more inhabitants and 281 so-called "million cities" (Sat- many obstacles to that ideal collaboration. Removing terthwaite 1998). In developing countries, cities with these obstacles and providing additional incentives for more than 1 million people jumped six-fold between collaboration between municipalities and civil society 1950 and 1995. Worldwide in 2000, the number of should be a priority if the vulnerability to disasters of a cities larger than 5 million was estimated to be 41 (IDNDR growing number of people in cities is to be addressed. 1996; U.N. 1999). The United Nations believes this 181 182 Building Safer Cities: The Future of Disaster Risk number will rise to 59 by 2015, adding another Mumbai, Caracas, and Havana are following a world- 14 million people to the streets and homes in large cities wide trend. Of the 14 cities with more than 10 million (accounting for 21 percent of the world's urban growth) inhabitantsin1995,allbuttwohavecoastallocationsand (U.N. 1999). In the same year, there were 19 cities 17 of the largest 20 cities are located either on coasts or with more than 10 million residents, a number believed rivers(Mitchell1999b:29).Thesprawlinggrowthofsome likely to increase to 23 by 2015. megacities has meant that some are subject to wild fires The megacity poses special challenges for disaster on their peripheries (Sydney, Los Angeles, Cape Town). risk reduction for a number of reasons.3 First, there is Othersarelocatedinseismicallyactivezones(MexicoCity, their sheer scale and geographic complexity. They sprawl Los Angeles, Istanbul, Tokyo) and sometimes near active over large areas, and this alone makes the day-to-day volcanoes(Manila,MexicoCity,Jakarta).Table13.1reviews monitoring of hazards and vulnerability difficult and the hazardous locations of several megacities. adequate provision and protection of lifeline infra- structure problematic. These urban regions have grown Table 13.1 Megacities at Risk over and incorporated a variety of pre-existing villages (UNU Study Cities in Italics) and towns. Therefore, street patterns are complicated, Population and interconnectivity within the whole region is a phys- City/ 2000 Conurbation (millions) Hazards to which Exposed ical challenge for the transport system. The megacity is Tokyo 26 Earthquake, flood made up of a number of municipal jurisdictions, so Mexico City 18 Earthquake, flood, landslide coordinated administration in normal times or emer- Los Angeles 13 Earthquake, flood, wild fire, gencies cannot be assumed. drought Second, there is the environmental impact, or "foot- Lagos 13 Flood São Paulo 18 Landslide, flood print," of such large urban regions (Girardet 1992; 2000). Mumbai/Bombay 18 Flood, cyclone, earthquake These cities require large amounts of energy, water, food, Shanghai 13 Flood, cyclone, earthquake and other "inputs" and create large amounts of solid, Calcutta 13 Cyclone, flood liquid, and gaseous effluent and waste heat. Some- Jakarta 11 Earthquake, volcano Beijing 11 Earthquake, severe winter times these characteristics of the urban ecology of megac- Manila 11 Flood, cyclone ities (and other large cities) contribute directly to the Johannesburg- 8 Flood, tornado exacerbation of a natural hazard. An example is the 1995 Gauteng heat wave in Chicago described in the "social anatomy" Notes: "Population 2000": Population estimates for the six UNU project megacities refer to the "greater" metropolitan urban regions. The source is of this event authored by Kleinenberg (2002). Seven the World Urbanization Prospects (U.N. 1999). These numbers are gross hundred and thirty-one people (mostly poor and eld- estimates. As Mitchell notes (1995: 509), "It is impossible to be sure of erly) died. Another example is the mass movement of the exact population of the world's largest cities. Among others: defini- tions of cities vary; municipality boundaries vary; the existence, fre- a mountain of refuse, triggered by intense rainfall, that quency, and accuracy of urban censuses vary; and the rates of population killed 705 people at the Payatas solid waste site in the change vary." "Johannesburg-Gauteng": The new 2001 South African northeast of Manila in 2000 (Westfall 2001).4 However, census results are not available at the time of writing. The last firm esti- mate of the population of Gauteng Province (essentially identical with the such environmental aspects as the provisioning of megac- greater Johannesburg or Pretoria-Witwatersrand-Vereeninging (PWV) ities with water and fuel and the removal of waste urban industrial region was 7.4 million in 1996 (GPG 2001). Using the average African urban growth rate of 4.4 percent per year for 1950­2000 more commonly have an indirect rather than a direct calculated by U.N. (1999: 4), one can estimate between 8­8.5 million in effect on disaster risk. In an earthquake or flood, these the year 2000, and a total population approaching 9 million in 2002. Since the end of apartheid in 1994, there has been a series of spatial- lifelines can be broken. Secondary hazards such as nat- administrative reorganizations in and around South Africa's large cities, ural gas fires or spillage of hazardous waste can result. resulting in the case of greater Johannesburg of a consolidation of 51 Provision of safe water and even food for survivors can municipalities into 15 (GPG 2002). Therefore, comparable data are diffi- cult to obtain. Since the former PWV urban region (greater Johannes- also be made more difficult. burg) already had 7.5 million inhabitants in 1990 (Wisner 1995: 337), Third, many megacities are located in hazardous loca- use of the administrative area of Gauteng Province as a surrogate for the megacity may produce an underestimate. In short, it could well be that tions. These include coastal areas or river deltas, where the greater Johannesburg megacity reached 10 million in 2002. storms and floods are common. Cities such as Manila, Source: Adapted from United Nations 1999. Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 183 Finally, megacities are highly diverse socially. In them term that comes from the so-called "livelihood approach" livesomeoftherichestandpoorestpeopleonearth.There in development studies and development policy (Sander- is typically significant linguistic, ethnic, and religious son 2000). It is used here in two ways. First, more con- diversity, and there is usually a population of recent immi- ventionally, social capital refers to the access to resources grants, some foreign, and some undocumented or ille- and information that households have by virtue of gal. "Illegality" and "irregularity" also can extend to the their noneconomic social relations with other people. nonimmigrant population, where extensive squatter These may be based on kinship, friendship, common settlements have grown (Fernandes and Varley 1998). religious adherence, membership in cultural and sports clubs, shared hometowns, etc. Studies of the ways house- Social Capital and Urban Social Vulnerability holds cope with the stress of drought, flood, and other Reduction hazards have shown that socially mediated access and mutual aid play an important part in coping strategies It is this last characteristic of the megacity, its social and, in turn, the resilience of livelihood systems (Blaikie diversity, that presents both problem and opportunity. and others 1994: 67­68). This is true in urban as well as A four-year study (1997­2001) supported by the United rural circumstances. Nations University revealed a rich mosaic of social net- By extension, the term social capital is also used in a works, coping and support systems, and citizen-based second way. To the extent that institutions of civil society initiatives in six megacities. However, this study also such as NGOs and CBOs (citizen-based organizations) revealed that there was a large gap in understanding and can provide a bridge between the formal agencies of dis- approach to urban social vulnerability on the part of aster management in governments and urban dwellers, municipal officials and civil society groups. Often they these institutions themselves form "social capital." carried out parallel or even conflicting risk-reduction Despite some reservations, the term has been adopted activities. At the extreme there was sometimes animosity as a shorthand for a wide variety of socially facilitated and distrust between them. Certainly in no case were opportunitiesandcapabilities(seebelow)foreaseofcom- the possible synergisms between civil society resources munication with development policy experts. The phrase and those of the municipalities maximized. The study "social capital" has, indeed, been criticized as depoliti- showed that municipalities had technical knowledge cizing the essentially conflictual nature of development. and financial resources that could benefit socially vul- Critics believe the term confuses the real differences in nerable groups of people, but they lacked detailed knowl- political and economic power in most societies between edge of these groups and were not trusted by them. By a minority elite who control the means of production and contrast, nongovernmental organizations (NGOs) and financial capital and the majority who may enjoy other institutions and groups within civil society knew "other" capitals--"human," "social," etc. (Harriss 2001). a great deal more about these groups of vulnerables (e.g., There are similar problems surrounding the use of the homeless, the frail elderly, street children, the dis- the term human capital; however, again, for ease of com- abled, those in squatter settlements) and had their trust, munication with policymakers and for brevity, it refers but lacked the technical and financial resources. to local knowledge of the built and natural environment Cities need institutions of civil society to provide a and the skills, formally and informally acquired, that are "bridge" or "interface" between municipal administra- collectively called capacity in disaster-management cir- tions and the 600 million people (possibly closer to a cles (see below). billion) who live in unplanned, self-built, unserviced, Vulnerability is a term that occurs with a general mean- and legally unrecognized settlements (Toepfer n.d; Lee ing in livelihood literature and a more precise signifi- 1994: 396).5 The study discussed in this paper was an cance in disaster management. The more general meaning attempt to understand the obstacles in the way of improv- refers to susceptibility to harm or loss due to external ing that bridging function. shocks or stresses on the livelihood system (e.g., extreme Before entering into the detailed results of this study, price swings such as the current coffee price decline, several key terms require definition. Social capital is a political regime changes, extreme natural events such 184 Building Safer Cities: The Future of Disaster Risk as drought or flood, seasonality, etc.). In this paper, and striking similarities among these four large urban regions. in the UNU study, a more precise definition was adopted. First is their size. They all fall squarely into the strictest In words employed by the author and coauthors in 1994, definition of "megacity," with well over 10 million people "By `vulnerability,' we mean the characteristics of a person in their metropolitan areas. They are all extensive, but or group in terms of their capacity to anticipate, cope Los Angeles far exceeds the others in sprawl, and has with, resist, and recover from the impact of a natural the lowest average density. The other three are densely hazard" (Blaikie and others 1994: 9). populated, with Tokyo exceptionally so. Los Angeles is The author understands and defines vulnerability as the youngest city region among them. The others are the absence or blockage of capacity to anticipate, cope each twice as old, or nearly so, even dating Mexico with, resist, and recover from the impact of a natural City from the Spanish conquest and not from its Aztec hazard. Others have juxtaposed vulnerability and capac- origins (given in brackets). ity in this way (Anderson and Woodrow 1999; IFRC All four urban regions contain considerable flood- 1999). These capacities (in his work, Sen (1993) refers prone flatlands, even though Mexico City is the only to them as "capabilities") include an array of skills, one without a coastal location. Three have over many knowledge, and networks through which one gains years augmented their coasts with considerable land- access to information, knowledge, and material resources. fill that shares with the drained lakebed under the his- In short, vulnerability/ capacity is a function of the liveli- toric center of Mexico City soil conditions subject to hood system in its entirety, including social capital. subsidence and liquefaction. In all cases, there are hills In this approach, vulnerability makes up one of three adjacent to or intermixed with these flatter parts. There- components of disaster risk. In addition to vulnerabil- fore, despite their differences in climate, in all cases there ity, there are hazards (an event in nature) and mitiga- are times in the year, or particular climate events, during tion (measures taken by households or collectivities to which one can expect landslides. reduce the impact of the hazard).6 These four essential elements of disaster management and prevention are related in a systematic way. Risk (R) is a function of Social Vulnerability: Perception and Policy the frequency and magnitude of natural events, often Comparisons in Four Megacities called hazards (H), the vulnerability (or capacity) of people (V), and the ability of government agencies, other Despite these important similarities in the geography of groups and institutions, or households to prevent or hazardousness, the UNU study found different defini- mitigate, and prepare for, hazard events (M, as the short- tions or perceptions of who the highly vulnerable social hand for all these activities, is usually "mitigation"). groups are and what policies best address their needs. These relationships can be expressed schematically: Two-by-two comparisons of the megacities reveal R = [H × V] ­ M. both convergence and difference. Before overall pat- terns and their policy implications are discussed, two A final term that needs defining is civil society. This such sets of comparisons will be presented: Mexico City term is used to signify all institutions and organized versus Los Angeles and Manila versus Tokyo. activities in society that are nongovernmental and out- side direct control of the government, political parties, Mexico City versus Los Angeles and businesses. Nongovernmental organizations and citizen-based organizations are part of civil society. In Mexico City, the more detailed breakdown by age, gender, and socioeconomic status was generally thought Megacity Profiles by officials in the Delegations of the Federal District (DF) and the Municipalities of the Estado de Mexico to Comparing four of the UNU study megacities gives an be an academic luxury of a rich country (see table 13.3). impression of the range of geographic and socioeco- While a small number of respondents acknowledged that nomic conditions in them. Table 13.2 summarizes some some of these groups face additional risks and problems Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 185 Table 13.2 Comparison of four megacities Characteristics Greater Los Angeles Metro Manila Greater Mexico City Metro Tokyo Population 13 11 18 26 Size (000 km2) 87 15 22 14 Density (pop/ km2) 149 733 818 1,857 Age since foundation (years) 220 430 477 (666) 398 Situation Coastal and Coastal peninsulas Inland valley on Coastal, running inland valleys between bays plateau north and west into hills Topography Mix of flood plain, Coastal plain, river Center over ancient Flat in much of canyon, coastal cliff, flood plain, hilly lakebed, many ward (Ku) area, and estuary to East ravines to North, more relief in West, and South, Tama area to flatter to Northeast West Climate Semi-arid Tropical Semi-arid Temperate Political and economic Regional economic Nationally primate Nationally primate Nationally primate importance role, Pacific Rim and sub-regional in economic and in economic and and Latin America, economic role political terms, political terms, Regional economic in Asia regional economic world an regional and political role in U.S. role in the Americas economic center Percent poor 25 50 60 10 Percent in informal 5­10 30 40 2­3 settlement or illegal migrant Natural hazards Earthquake, fire, Earthquake, flood, Earthquake, flood, Earthquake, flood, flood, landslide landslide, typhoon landslide typhoon Last major disasters Northridge earthquake Payatas garbage dump Earthquake 1985 Earthquake and 1994; wildfires 1995 flood, fire, and landslide fire 1923 2000 Sources: Manila and Tokyo: Fuchs and others (1994), Lo and Yeung (1996), Tayag (1999), Velasquez and others (1999); Tokyo: Takahashi (1998; 1999; 2003), Tokyo Metropolitan Government (1995); Manila: Tayag (1999); Mexico City: Gilbert (1994; 1996), Townsend (2000), Puente (1999a and 1999b), Eibenschutz and Puente (1992); Los Angeles: Wisner (1999a; 1999b; 2003a), Bolin and Stanford (1991; 1998); General: U.N. (1999).7 Table 13.3 Groups perceived by disaster management professionals to be highly vulnerable to disasters (Percent officials) Mexico City Los Angeles Squatters (67 %), especially Elderly persons (100 %) Living in ravines Living over ancient mines Living near hazardous industries Children (23 %) Disabled persons (93 %) Children (93 %) Legal immigrants (16 %) Persons with special medical needs (86 %) Disabled persons (14 %) Mentally ill (54 %) Elderly (14 %) Illegal immigrants (29 %) Homeless (11 %) Foreigners/ foreign-born (29 %) Mentally ill (5 %) Homeless (21 %) Persons with special medical needs (5 %) Street children (14 %) Illegal street vendors (5 %) People living near oil refineries (7 %) Artisanal fireworks producers (5 %) People living near water pumping stations (4 %) Street children (2 %) People living in mobile homes (4 %) Notes: "Percent Officials": Percentage of 44 disaster-management officials interviewed in greater Mexico City and 28 interviewed in greater Los Angeles. "Legal immigrants": This includes people from the rural areas of the country where indigenous people live. "Foreigners/foreign-born": This was said to be mostly to do with lack of knowledge of English. 186 Building Safer Cities: The Future of Disaster Risk during disaster recovery, the consensus was different. Mexico City officials and their counterparts in greater Most believed that illegal or informal squatters, who most Los Angeles try to involve neighborhood groups and commonly live in ravines over ancient mines on some NGOs in the planning process, but more try to do so of the surrounding slopes, were generally vulnerable. in the Mexico City megacity (Carrasco and Garibay They thought that everyone in such a living situation 2000). This difference is due in large part to the history was vulnerable without finer distinctions. The excep- of social and political organization in the two urban tion to this concerned a more common belief that chil- regions. In Mexico there is a long history of political dren needed special protection. party patronage and clientalism that manifests itself in In greater Los Angeles, there was nearly universal the form of a variety of local associations and groups acknowledgement of the special vulnerability faced by that rely on such support. There is also a tradition of the elderly, disabled persons, children, and people with opposition and protest in Mexico that gives rise to other special, chronic medical needs (e.g., those on oxygen groups that do not enjoy party political support. or ventilators at home or those in need of frequent dial- It is striking, however, that despite claims of involve- ysis). The mentally ill or disabled were also recognized ment of citizens in the planning process, few municipal- in more than half of the interviews with disaster man- ities in greater Mexico City actually obtain information agement officials in greater Los Angeles. A smaller group aboutsociallyvulnerablegroupsfromneighborhoodgroups of municipalities believed that the legality of immigrant (where the fine-grained and detailed information exists). status, language ability of foreigners or the foreign At the level of the municipal jurisdiction, both sets born, and homelessness create situations in which people of officials, in Mexico City and Los Angeles, claim a high can suffer increased vulnerability to disaster. degree of intersectoral cooperation. This turns out in Despite differences in the way that social vulnera- practice to be a matter of legal formality--attending bility is defined and understood by municipal disaster the same planning meetings, signing off on the same managers, their approaches to planning and to the acqui- planning documents. However, more than half in both sition of information is similar. Table 13.4 summarizes cases claim to obtain information about socially vul- these data. nerable groups from other departments in the same Table 13.4 Knowledge of vulnerable groups and planning of programs to reduce vulnerability in Mexico City and Los Angeles (Percent officials) Mexico City Los Angeles Involve neighborhood groups in planning (71 %) Involve neighborhood groups in planning (50 %) Obtain information from neighborhood groups (9 %) Obtain information from neighborhood groups (21 %) Involve NGOs in planning (43 %) Involve NGOs in planning (21 %) Intersectoral coordination at municipal level (91 %) Intersectoral coordination at municipal level (100 %) Information from other government departments in Information from other government departments municipal government (68 %) in municipal government (61 %) Information from national agencies (30 %) Information from national agencies (14 %) Experience problems using social data (66 %) Experience problems using social data (71 %) Notes: "Involve neighborhood groups...": Many of these take the form of groups formed around someone who has taken the free 18-hour course called Citizen Emergency Response Training (CERT) made available to citizens. The inspiration for this kind of training came from the experience of sponta- neous citizen action after the 1985 earthquake in Mexico City, where L.A. Fire Department Chief Frank Borden had gone as an observer. The course included fire suppression, light search and rescue, first aid, transportation of the injured, communication, and team leadership. "Involve NGOs in plan- ning": One of the six municipalities that involves NGOs is the City of Los Angeles, where there is an active network of 70 NGOs with official status in the planning and emergency-response system called the Emergency Network Los Angeles (ENLA). There is a great contrast between a city like the City of Los Angeles and its relationship with NGOs through ENLA, and other, much smaller, municipalities that have no process for involving NGOs with the excep- tion of the two national, quasi-governmental bodies, the American Red Cross and the Salvation Army. "Intersectoral coordination...": Universal claims of coordination are explained by the legal requirement in California to follow what is known as the Standard Emergency Management System (SEMS), which mandates plans and exercises that involve multiple sectors and mutual aid contingency arrangements among cities and counties. "Information from national agencies": This was most commonly information from the National Center for Disaster Prevention (CENAPRED) or the National Institute of Statistics, Geography, and Information (INEGI). Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 187 municipal government. This sharing of information For example, in Los Angeles neither the Mothers of East takes coordination beyond mere formalism. L.A. nor the Community/Labor Strategy Committee-- A most striking result was how few municipalities two NGOs with long histories of activism on air qual- take advantage of the many publications and elec- ity problems--concerned themselves with the fact that tronic information sources made available by their respec- low income Blacks and Hispanics are more likely to tive national government agencies. In part, this is due live in flood-prone areas. to a lack of financial resources and time by understaffed, There were a few, predictable NGOs whose mandate small municipal offices. It is also a reflection of the back- specifically addresses aspects of risk communication or ground and lack of specific training in social sciences more general disaster management: these include the for those working on disaster management at the munic- Red Cross in both cities, or Emergency Network Los ipal level. In greater Mexico City, most have engineer- Angeles (ENLA) and the Salvation Army in Los Angeles ing backgrounds or come from the construction industry. and CARITAS in Mexico City. Ironically, however, these In greater Los Angeles, they either have had careers in NGOs have been so fully officialized and incorporated law enforcement or fire fighting. In neither case do man- into the metropolitan systems of disaster management agers find it easy to use social data. that they do not function as conduits to and from the These results provide support for the general find- poorest of the poor and other special-needs groups. ing mentioned earlier: municipalities generally have the In a similar way, there are some specialized disaster- technical resources to meet the needs of socially vul- oriented CBOs, such as the neighborhoods in central nerable groups, but they lack detailed information about Mexico City trained by the Association of Retired Fire them. Fighters8 and those in Los Angeles that have had Citi- zen Emergency Response Training (CERT). These suffer from narrowness of mission and, in the case of Los Ange- Social Capital and Civil Society in Mexico City les, a definite class bias. Most of the roughly 20,000 and Los Angeles CERT-trained individuals in the City of Los Angeles are The starting point of this research project was the hypoth- white and middle-class. These groups are not oriented esis that civil society (NGOs and CBOs) can provide a toward dialogue with marginal, socially vulnerable groups vital link or bridge between highly vulnerable popula- of people because their mission is defined as preparedness tions and municipal governments. In the ideal world, and response (not prevention and mitigation) and because such groups would have information about and trust of the socioeconomic class position of leaders and mem- relationships with marginalized people who the city bers of these groups. finds it difficult to understand or to approach. The UNU The second limiting factor concerns politics. In a research partially supported the hypothesis. However, number of interviews, municipal officials indicated that there are at least three factors that seem to limit the they believed that NGOs involved themselves in relief ability of civil society to function in this mediating, and post-disaster recovery work to further their own bridging manner. political ends. They were not trusted and collaboration The first limiting factor concerns the structure and suffered. From the NGO side, there was as often a history function of NGOs themselves. Most NGOs have their of antagonism with the government. Mistrust from the own fairly narrow and well-defined agendas and areas NGO side could have deep roots and center around larger of expertise and concern (Foreman 1998; Carrasco 2000; societal issues such as human rights and corruption-- Benson and others 2001). In part this is a natural result giving rise to such epochal changes as the electoral loss of how NGOs are formed and remain funded. They of Mexico City by the PRI political party. Mistrust could carve out niches in the urban ecology. Focused concerns also be focused on feelings of neglect and social exclu- include housing, legal empowerment, women's rights, sion by the communities served by the NGO, as was and sanitation. The problem observed is that such groups the case of the Pico Union Cluster near downtown Los see disaster management and the process of vulnerabil- Angeles. This is a low-income residential district popu- ityreductionthroughtheprismoftheirestablishedagenda. lated by Hispanic immigrants, especially from El Salvador 188 Building Safer Cities: The Future of Disaster Risk and Guatemala, many of them undocumented. The Mexico's many indigenous minorities. Fourth, the ques- housing stock comprises five- and six-story brick ten- tion of trust arises in the relations between groups of ements and poorly maintained, subdivided wood-frame people and NGOs/CBOs. While it was generally found Victorians that date from an earlier, more affluent period that specialized NGOs had reasonably frequent con- in this districts' settlement history. tact with their constituents (e.g., the frail; low-income Because damage from the Northridge earthquake in elderly; the homeless; street children; street youth; 1994 was not expected as far from the epicenter as the undocumented immigrants; etc.), a relationship of trust Pico Union district, little official assistance arrived for sev- cannot be assumed. This depends on the history and eral days, despite appeals by a consortium of 40 service credentials of the NGO. For example, in parts of greater NGOs called the Pico Union Westgate Cluster (PUWC). Los Angeles where the main citizen-based vehicle for PUWC provided improvised shelter and food for the disaster-risk management is the Neighborhood Watch-- displaced. Later, antagonism continued to build as the originally and primarily set up to prevent crime--youth Spanish-speaking field workers dispatched by FEMA gangs, a significant group of people, exclude themselves. turned out to be Puerto Rican and Miami Cuban, whose The third limiting factor concerns continuity and Spanish idiom, body language, and attitude toward capacity building. In numerous cases, NGOs that had the residents was considered disrespectful. Soon PUWC formed spontaneously in response to disasters such as also found itself lobbying to call back city engineers the 1985 earthquake in Mexico City or the 1994 North- whom residents believed had inadequately inspected ridge earthquake in Los Angeles did not persist beyond some structures in the neighborhood, a problem of per- the early stages of recovery. If the whole point of devel- ception and communication exacerbated by the well- oping a "culture of prevention" is to build networks at known reluctance of Central Americans to trust buildings the neighborhood level capable of ongoing hazard assess- that have suffered earthquake damage. ment and mitigation at the micro level, preparedness Four different kinds of trust (or lack thereof) are training, and the identification of vulnerable individu- evident in these examples.9 First, localities (neighbor- als, then the organizational base is weakened or even hoods) are often heterogeneous in their social compo- lost each time "emergent" NGOs rise and fall in response sition: socioeconomically, ethnically, in terms of age to specific events. ranges, stages of life, etc. This makes it difficult or impos- There were some exceptions in Mexico City. A few sible to treat localities in a unitary way as "communi- organizations stand out for effective response to the 1985 ties" (Morrow 1999; Buckle and others 2000; Wisner earthquake and continuous efforts since then without 2003b). One cannot assume that individuals and house- becoming dependent on political parties. CARITAS was holds trust one another. Second, there is trust or a lack directly involved in providing temporary housing and of it between institutions of civil society and formal insti- medical attention, as well as longer-term housing recon- tutions of governance. Moreover, it was observed that structionandemploymentgeneration.Theyactedasinter- CBOs might trust one agency within the state appara- mediary for housing grants, built 2,240 new houses, and tus but not another, one level of government (e.g., the bought 288 buildings and 1,308 plots of land for hous- local municipality) but not another (e.g., national). Third ing (Puente 1999a). They have stood by the families comes the issue of trust between individuals and house- involvedinpost-earthquakerecoverythroughasubsidiary holds and organs of formal governance. In Los Angeles, organization formed for the purpose, Apoyo a la Comu- undocumented immigrants and other marginal groups nidad.Hence,inthiscase,thereisstilltheconnectionwith of people (e.g., homeless youth, squatters) avoid contact the community and the potential for broader, continu- with officials because they fear deportation or eviction. ous capacity building. CARITAS is unusual because of its Frequently, there are also linguistic and cultural barriers. independence from the system of political patronage in In Mexico City there are similar obstacles to productive the city. This is because CARITAS is the local branch of a contact between government and such groups as street worldwide official Catholic charity. The political and vendors, street children, squatters, and illegal immigrants economicindependenceoftheCatholicChurchinMexico from Central America and from the rural homelands of dates from the Mexican Revolution in the 1920s. Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 189 In a similar way, the Tenants' Association of the gov- Manila were regarded by the majority of the officials ernment housing complex at Tlatelolco had had a long interviewed (87 percent) to be the social group at history of activism before the 1985 earthquake, and a highest risk (67 percent shared this view in Mexico City). strong sense of shared identity. One respondent proudly The elderly and disabled were universally seen as the noted that Tlatelolco was the location of the last stand of groups in Tokyo who are most vulnerable. Indeed, a AztecwarriorsagainsttheSpanishin1521.Morerecently, highly nuanced typology of the vulnerable elderly these center-city apartment dwellers witnessed the mas- emerged from interviews there in the 23 central wards. sacre of 400 university students by the Mexican army in These differences may well mirror the macroeco- 1968 in the Plaza of the Three Cultures. They protected nomic conditions prevailing in the two cities. As in many other students in their homes, and some of these Mexico City, national patterns of poverty give rise to bystanders were beaten by the army. Thus, after the 1985 migration to informal settlements in the countries' major temblorbroughtdownoneofTlatelolco's18towerblocks, economic centers. The resulting urban marginality there was not only the usual outpouring of solidarity creates a population of homeless children and youth. and mutual aid characteristic of Chilangos (citizens of These were the second most commonly mentioned in Mexico City). Rather unusually, the tenants remained Manila. organizedinordertofightgovernmentattemptstodemol- Reflecting its role as a destination for many legal for- ish the whole complex and re-house tenants on the out- eigners with little knowledge of Japanese, such people skirts of the urban region. Partly as the result of this were considered potentially at risk by 70 percent of dis- successful struggle, the tenants' association has remained aster planners in Tokyo. This has more to do with Tokyo's active in disaster-risk education and preparedness. They role as a global business and financial center and less hold workshops on preparedness for the residents and to do with illegal working-class immigration, although have collaborated with the Civil Protection authorities there are some of the latter who arrive from mainland oftheFederalDistrictindevelopingspecificdisasterplans. China, the Philippines, and North Korea. Table 13.5 pres- However, despite this collaboration with one arm of ents these data. government, the Tlatelolco residents remain critical of Despite very large economic and political differences others. For example, in 2001 they organized a success- between Manila and Tokyo, there is a similar approach ful campaign to stop plans by the Judicial Police to use to knowledge of disaster vulnerability and planning. a former community center building in the apartment Indeed, all four megacities show the same pattern. All complex as a holding prison. The tenants' association four attempt to involve neighborhood groups in plan- argued that this would put residents at risk. ning. All four also claim high degrees of intersectoral ContinuitywasalsoaprobleminLosAngeles.Thepre- coordination and connections with other jurisdictions mierNGOinvolvedinbuildinglow-costhousingandbro- within the megacities. However, none of the four cities kering assistance from the government for low-income, is particularly good at involving nongovernmental organ- ethnic minority people after the Northridge earthquake, izations in planning. Finally, disaster planners in all four Mano a Mano, was not able to maintain its funding base cities had difficulties using social data. after the initial recovery period and was supplanted by However, there are important differences within these the well-established NGO, Habitat for Humanity. Ironi- generally similar patterns. Manila and Mexico were better cally, Mano a Mano had acted as "cultural broker" and go- at the municipal level in acquiring information about between for Habitat in relations with Spanish-speaking vulnerable groups of people from neighborhood groups people during the emergency (Bolin and Stanford 1998). (65 percent and 57 percent of municipal level respon- dents saying they did). In Los Angeles, only one in five planners could count on this source of information, and Manila versus Tokyo a mere 9 percent in Mexico City. The contrast between Manila and Tokyo is similar to Another commonality between Manila and Tokyo is that between Mexico City and Los Angeles, only more the existence of legally established, strong, and well- extreme. Residents of urban informal settlements in financed metropolitan government structures that 190 Building Safer Cities: The Future of Disaster Risk Table 13.5 Groups perceived by disaster management NationalCenterforDisasterPreparedness--CENAPRED-- professionals to be highly vulnerable to disasters is located in Mexico City). Table 13.6 presents the data (Percent officials) on sources of knowledge and planning in Manila and Metro Manila Central Tokyo Tokyo. Squatters (87 %) Elderly persons (100 %) Bed-ridden elderly (61 %) Elderly living alone (48 %) Social Capital and Civil Society in Manila Elderly in general (26 %) Demented elderly (13 %) and Tokyo Street children (71 %) Disabled persons (100 %) Elderly (13 %) Legal foreigners (70 %) In all four cities, civil society has a double history as Disabled persons (7 %) Infants (39 %) Young children (7 %) Persons with special both critic and opponent of government and its part- medical needs (35 %) ner (Heijmans and Victoria 2001). In Manila this was Others (3 %) seen dramatically in the use of what was called "people's Note: Percentage of 31 disaster-management officials interviewed in metro power" to remove nonviolently both Presidents Marcos Manila (action officers responsible for disaster planning in the 16 admin- and Estrada. Less dramatic, day-to-day manifestations istrative subunits of metro Manila plus 5 at primary district (barangay) level, and 10 in a variety of government commissions; and 23 disaster- of this same tension exist in the work of the legal aid management officials in the 23 central wards in the case of Tokyo. "Street network in Manila that provides support for squatters. children": However, these respondents believed that they were not responsible for dealing with the vulnerability of street children, as their Recently they represented a group of fishermen and welfare falls under the Department of Social Welfare and Development. their families threatened by eviction from riverside homes Five officials believed that street children are too mobile and transient to in Malabon (part of Manila), where large-scale channel be the responsibility of any municipal jurisdiction. "Disabled persons": In the case of Tokyo, this included both physical and mental disability. modifications were scheduled in aid of flood control. "Legal foreigners": Officials mentioned only legal non-Japanese residents, Legal aid volunteers also represented the survivors of who may have difficulty understanding Japanese language warnings and instructions. While the growing presence of illegal immigrants, especially the Payatas garbage dump disaster mentioned earlier. among those doing casual labor, was recognized, no official believed that On the other hand, Manila also has a dense network of they were a group of vulnerable people for whom special disaster plan- non-profits and citizen-based organizations that work ning should be done. Likewise, the homeless in Tokyo subway stations, in parks, and along the Sumida River were acknowledged to exist, but in partnership with municipal governments. In contrast they "did not count" for planning purposes. See Wisner (1998). "Others": to the conflictual relocations of people with river and There were single mentions among the 31 officials (3 percent) of orphans, students living in boarding houses, women (battered, pregnant, or lactat- ocean-based livelihoods in the municipality of Malabon, ing), mentally disabled (due to drug use), persons in flood-prone areas. its counterpart municipality, Marikina, was able to achieve a negotiated relocation away from the flood plain with embrace a large portion of the megacity. In Mexico the financial assistance of the Philippine Commission City and Los Angeles there are no overarching metro- on the Urban Poor (PCUP) (Tayag 1999). politan coordinating structures. The existence of As noted above, 11 (65 percent) of the 17 Disaster metropolitan-wide institutions explains the high degree Action Officers interviewed in metro Manila acknowl- of intersectoral and inter-city coordination claimed by edged assistance from neighborhood associations and respondents. Coordination was only diminished in the NGOs in disaster management. However, as observed case of Tokyo by a countervailing process--the strict in Mexico City and Los Angeles, most of this involve- interpretation of privacy laws that prevents one depart- ment focuses on response and recovery (help with com- ment's sharing of social data with another, even at the munications, transport, evacuation, rescue, provision level of a single ward. of house repair grants and loans) and does not address Finally, municipalities in both Manila and Tokyo used vulnerability as such. information made available by national-level institutions Tokyo has a highly developed social infrastructure such as the Philippines' Presidential Commission on of citizens' organizations at the neighborhood level UrbanPovertyandJapan'sNationalLandAgency.Nosim- that serve various functions such as firefighting, first ilar widespread utilization of nationally generated infor- aid, hazard mitigation, and education about earthquakes. mation takes place in Mexico City (ironically, since the The volunteer fire corps (shobodan) dates from 1718. Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 191 Table 13.6 Knowledge of vulnerable groups and planning of programs to reduce vulnerability (Percent officials) Metro Manila Central Tokyo Involve neighborhood groups in planning (73 %) Involve neighborhood groups in planning (100 %) Obtain information from neighborhood groups (65 %) Obtain information from neighborhood groups (57 %) Involve NGOs in planning (18 %) Involve NGOs in planning (22 %) Intersectoral coordination at municipal level (100 %) Intersectoral coordination at municipal level (100 %) Obtain information from other departments in municipal Obtain information from other departments government (100 %) in municipal government (13 %) Obtain information from national agencies (100 %) Obtain information from national agencies (100 %) Experience problems using social data (71 %) Experience problems using social data (100 %) Notes: "Involve neighborhood groups in planning" (Manila). The 1992 Local Government Code specifies that local citizen groups must be represented in special bodies such as health boards, but not all municipalities have managed to involve neighborhoods in disaster planning. "Involve neighborhood groups in planning" (Tokyo): There is a centuries-long tradition of urban neighborhood groups based on ceremonial functions and other more practical activities such as fire fighting. Tokyo's neighborhood fire brigades date back to the 18th century. However, many groups are merely formal and not active. Fifty-seven per cent of officials expressed concern about the level of participation of citizens at the neighborhood level, and 83 percent characterized their ward's neighborhood groups as "formal." Only four wards (17 percent) said they had very active neighborhood groups. "Involve NGOs in planning" (Manila): Excluding the Philippine National Red Cross, which is present and active in all municipalities and treated as part of government for planning purposes. "Involve NGOs in planning" (Tokyo): The Japanese Red Cross was referred to in eight wards (35 percent), and Shakai Fukushi Kyougikai, an umbrella organization to coordinate social welfare organizations dealing with visual impairment, physical disabilities, and mental retardation in seven wards. However, only five wards claimed to have active involvement of NGOs in their plans. "Intersectoral coordination..." (Manila): Via the Metropolitan Manila Development Authority and its Metro Manila Disaster Coordinating Council. "Intersectoral coordination..." (Tokyo): Via the Tokyo Metropolitan Council. "Obtain information from other departments..." (Tokyo): Strict interpretation of privacy laws in all but three wards meant that there was very little sharing of information about vulnerable groups of people from one department (e.g., that dealing with the elderly, for example, or the disabled) and the department of disaster planning. "Obtain information from national agencies" (Manila): Particularly the Presidential Commission on Urban Poverty and the Department of Social Welfare and Development. "Obtain information from national agencies" (Tokyo): The Tokyo Metropoli- tan Council provides many maps and planning materials, as does the National Land Agency. Local groups that work to identify and mitigate haz- activism has not yet been channeled into the revital- ards in their neighborhoods (machizukuri kyo gikai) are ization of moribund neighborhood groups. Nor is there said to have numbered 56,000 in 1991, covering 37 crossover to concern with urban social vulnerability to percent of all households (Kaji 1991: 222). Likewise, disasters from the environmental activism in greater the Japanese Red Cross strongly advocates the impor- Tokyo that has seen years of protest over the expansion tance of local knowledge in disaster management of Narita airport or chemical and radioactive hazards. (Higashiura 1994). Indeed, the Tokyo Metropolitan The problem of formalism also affects metro Manila. Government (TMG) officially advocates "Reinforcing Since 1992, the Local Government Code has required citizens' disaster prevention groups" and "[e]nsuring the NGOs be represented in local health boards and the safety of especially vulnerable citizens" (TMG 1995: other specialized bodies in the 17 administrative juris- 78). However, despite the existence of this social cap- dictions that make up the megacity. However, by the ital, in practice, the UNU study found that neighbor- time of the UNU study, this requirement had been imple- hood groups find it hard to recruit young people, and mented in only four of metro Manila's constituent cities. their activities are narrow and formalistic. As noted Nevertheless, some good practice does exist as a basis earlier concerning CERT-trained neighborhood groups forfurtherprogress.Forexample,inthecityofMuntinlupa, in Los Angeles and the formal outreach by Civil Pro- in the extreme south of the megacity region, there is an tection authorities in Mexico City's Federal District, there arrangement similar to the relationship of NGOs to the is a middle-class bias also in Tokyo that excludes the City of Los Angeles Division of Emergency Prepared- homeless, mentally ill, and illegal immigrants. ness. In the Muntinlupa case, several emergency response Volunteerism has increased rapidly since tens of thou- NGOs10 are integrated into local government plans sands of young people converged on Kobe after the although they are do not sit on the local government earthquake there in 1995. However, this spontaneous council (Tayag 1999). 192 Building Safer Cities: The Future of Disaster Risk General Conclusions about Social Capital and · Limited or ritualistic use of community or neigh- Urban Disaster-Risk Reduction borhood groups · Political hostility toward NGOs Urbanization since 1950 can be understood as a neces- · Funding shortages and high turnover in NGO staff. saryresultofasuccessionofdevelopmentmodelsadopted The last three of these problems are very common by policy makers. So, too, the growing social abyss and block the effective use of social capital to reduce between rich and poor in these cities and between risk. Socially vulnerable and marginal groups in cities those labeled "illegal" and their "legal" neighbors has have needs but also capacities. Their local knowledge resulted from development policies. As a result, social and coping as well as their needs can be communi- and human capital are grossly underutilized in disaster- cated to government agencies through NGOs and CBOs. risk reduction activities. Ignorance and mistrust blocks City agencies have systems of risk reduction that may the emergence of the "hybrid" management of risk that benefit socially vulnerable groups. Here again, it is the would merge the science, technology, and engineering bridge, mediation, or interface11 provided by organs of available to municipalities with local knowledge and civil society that can provide access to official risk reduc- coping practice and social networks available among tion. The exponential growth of CBOs and NGOs during "illegal" urban dwellers and in squatter settlements. the decades of the 1980s and 1990s has expanded the In the words of Dr. Suvit Yodmani, Executive Direc- basis--with all the pros, cons, and difficulties men- tor of the Asian Disaster Preparedness Center in Bangkok tioned earlier--for much deeper and systematic rela- (Yodmani 2001): tions between cities and civil society. Viewed from this more optimistic perspective, the A paradigm shift in the development sector from income UNU study revealed at least in a few of the constituent poverty to human poverty has been paralleled in the dis- cities of the four megacity regions the following: aster management sector by a shift from seeing disasters · Innovative use of existing neighborhood groups for as extreme events created by natural forces, to viewing them as manifestations of unresolved development problems. preparedness or even for hazard and vulnerability mapping One such "unresolved development problem" is the · Some cases of excellent coordination between munic- lack of articulation and integration of the social insti- ipalities and NGOs tutions of civil society with those of national and sub- · Improvements in risk communication and increased national, formal governance. This is not a simple sensitivity on the part of some municipalities to the "technical" problem of decentralization, but a deeper needs of foreigners, both legal and illegal. problem of trust. It involves issues of accountability and The main conclusion, therefore, is that the social basis representation, and, ultimately, the question of legiti- for disaster-resilient cities must be continued capacity mation of the state. building across the whole of these heterogeneous pop- If the municipalities in this study of four megacities ulations (Eade 1997; Plummer 1999). Revitalized dem- are at all typical, then urban social vulnerability remains ocratic participation in the governance of cities, better a serious problem as yet insufficiently faced by munic- education systems, employment generation, broader ipal, metropolitan, or other higher orders of govern- inclusion of women, minorities, and youth all contribute ment. Among the specific problems documented are: to this (Gilbert and others 1996; Freire and Stren 2001). · Fragmented and uncoordinated responsibility for dif- ferent at-risk groups · Legal barriers to access to social data Notes · Staffing shortages and lack of training in the use of 1. The present paper utilizes the data from only four of the six available social data, resulting in little use of exist- study cities. This is for three reasons. First, the pair-wise com- ing sources parisons of Tokyo/ Manila and Los Angeles/ Mexico City pro- · Limited or sometimes no planning at the municipal vide sufficient insight into the common results and differences level for longer-term recovery issues encountered. Second, since the main surveys were conducted Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 193 in all six cities (1999­2000) there have been significant new significant improvement in the lives of at least 100 million slum events that have radically altered the relations among munici- dwellers by the year 2020" (UNDP 2002). Reducing vulnerability palities and civil society in Johannesburg and Mumbai. These to volcanic eruption and lahar flow, earthquake, flood, storm, and new conditions cannot be addressed in the space I have been landslide can be done in ways that also increase livelihood options allocated and deserve separate, extensive treatment. In the case and resilience among such people. The converse is also true. All of Johannesburg there has been a reorganization of the admin- efforts to implement this Millennium Target can be wiped away in istrative structure of the city as part of the dismantling of the an extreme natural event, as has occurred recently in the case of system of apartheid in South Africa (see: Beall and others 2002; hurricane Mitch (1998), flood and landslides in greater Caracas Bond 2000; Wisner 1995). In addition, there have been several (1999),earthquakesinGujarat,IndiaandElSalvador(2001),land- urban upgrading initiatives in parts of greater Johannesburg as slide in Algiers (2001), and volcanic eruption that destroyed the result of the visibility anticipated by authorities with much of Goma, Congo (a city of 500,000 people) in 2001. 60,000 delegates due to arrive for the World Summit on Sus- 6. Mitigation is sometimes also referred to as a combination of tainable Development (August­September 2002) (see Bond forth- personal and social protection (Cannon 2000; Wisner and others coming; Alexandra Renewal Project 2002; World Bank 2001; 2003). Greater Johannesburg Metropolitan Council 2000). "Upgrad- 7. "Age since Foundation": All the megacities were predated by ing" has in some cases involved forced evictions that may be seri- small settlements, some occupied for a long and undetermined ously affecting relationships between civil society organizations period. In the case of Mexico City, two ages are given. The first and the municipal state (Magardie 2002; AFROL 2001). In the uses the date of the establishment of the Aztec settlement of case of Mumbai, the reaction to the devastating earthquake to Tenochtitlan (1325 or 1345). The second uses the date of the the North in Gujarat (2001) stimulated the creation of an ambi- defeat of the Aztecs by Cortez (1524). The banks of the Pasig tious multi-sectoral disaster preparedness plan and the realign- River in present-day Manila were inhabited long before the Span- ment of relationships among municipal government, business, ish colonial period, but the age provided uses the date of Span- and civil society around it (see Rowell 2002; Sinha 2001). Thirdly, ish control of Manila (1571). "Percent Poor": Percent below the additional complexities in Johannesburg and Mumbai require locally defined poverty line. Clearly, one is dealing with relative further detailed follow-up, which is under way, in order to under- and not absolute poverty in these comparisons; however, in all stand how recent dramatic events in both cases may have mod- cases among the "poor," there is little or no surplus for finan- ified the patterns of municipal state-civil society relations shown cial investments in self-protection because of the relative expense in data collected earlier (1999­2000). of food, shelter, utilities, and transport. "Percent Informal Set- 2. The author would like to acknowledge the efforts of the team tlement or Illegal Immigrants": There is little squatter or infor- involved in the United Nations University research project coor- mal settlement in greater Los Angeles or greater Tokyo except dinated during 1997­2001. The team comprised: G. Berger for some semipermanent encampments of homeless individu- (video documentary consultant), R. Bolin (Los Angeles con- als and the illegal use of condemned buildings ("squats"). In sultant), M. Luna (Manila consultant), W. Pick (Johannesburg areal extent and numbers of inhabitants, these are not at all com- coordinator), S. Puente (Mexico City coordinator), S. Takahashi parable to the large informal settlements in greater Mexico City (Tokyo coordinator and Los Angeles consultant), J. Tayag (Manila and greater Manila. However, when one considers the numbers coordinator), J. Uitto (UNU academic officer in charge, 1997­ of illegal or undocumented persons in the urban population, a 1999), J. Velasquez (UNU academic officer in charge, 2000­2001), different but related percentage can be estimated for Los Ange- and B. Wisner (overall project coordinator and Los Angeles les and Tokyo. The commonality between the two measures is coordinator). "illegality" and the challenge for risk reduction that that pro- 3. On disaster risk in cities, see Hardoy and Satterthwaite (1989), duces. "Last Major Disaster": The fire bombing of Tokyo during Hardoy and others (1990), Hardoy and others (1992), Mustow the Second World War is not included, although more recent (1995), IDNDR (1996), Mitchell (1999a), Mansilla (2001). than the 1923 earthquake and fire, because it was not triggered 4. Two hundred and five bodies were recovered, and 500 were by an extreme natural event. While the catastrophic mass missing and presumed dead (Westfall 2001). movement of solid waste, and subsequent fire, that buried homes 5. Among these are those who are potentially the beneficiaries and people at Payatas, Manila, is partly the failure of a human of implementation of the Millennium Target 11, "to achieve a artifact (a solid waste dump site), the trigger was heavy rainfall. 194 Building Safer Cities: The Future of Disaster Risk 8. This Associacion de Bomberos en Retiro was unknown to munic- Bolin, R. with L. Stanford. 1998. The Northridge Earthquake: ipal authorities, and was unregistered as an NGO, prima facie Vulnerability and Disaster. London: Routledge. evidence that it was not the product of client-patron politics so Bond, P. 2000. Cities of Gold, Townships of Coal. Trenton, NJ: Africa often encountered among nominally independent NGOs in World Press. Mexico City. Bond, P. Forthcoming. Sustainable South Africa? Environment, 9. In the literature on environmental risks and technological haz- Development and the Post-Apartheid State (provisional title). ards, the concepts of trust and credibility have received much Brown, P. 1987. "Popular Epidemiology: Community Response attention (Douglas and Wildavsky 1982; Brown 1987; Wynne to Toxic Waste-Induced Disease in Woburn, Massachusetts," 1987; Slovic 1993; Bujra 2000; Slovic 2000; Barnes 2002). Treat- Science, Technology, and Human Values 12 (3 and 4): 78­85. ment of trust in this literature extends to discussions of it philo- Buckle, P., G. Marsh, and S. Smale. 2000. "New Approaches to sophical roots and implications for social theory (e.g., Beck 1992; Assessing Vulnerability and Resilience." Australian Journal of Beck and others 1994; Giddens 1992). Unfortunately, there has Emergency Management 15( 2): 8­14. not been a parallel exploration of these concepts in the context Bujra, J. 2000. "Risk and Trust: Unsafe Sex, Gender and IDS in of the study of natural hazards (with Kasperson and Pijawka Tanzania." In: P. Caplan, ed., Risk Revisited. London: Pluto (1985) as one of the rare exceptions). Press: 9­84. 10. These NGOs are: Brigada LAKAS, Lakas ng Kababaihan, Kaba- Cannon, T., 2000. "Vulnerability Analysis and Disasters." In D. likat Civiscom. J. Parker, ed., Floods, Vol. 1. London: Routledge: 45­55. 11. The term "interface" was used in this context by Rowell Carrasco, F. P. 2000. Participación Comunitaria y Cambio Social. (2002). Mexico City: Instituto de Investigaciones Sociales (UNAM) and Plaza y Valdes Editores. Bibliography Carrasco, F. P. and B.A. Garibay. 2000. Guia communitária para AFROL 2001. "Moving Alexandra Squatters Good for Environ- la prevención de desastres. Mexico City: Government of the City ment." Available from . University of Mexico (UNAM). Alexandra Renewal Project. 2002. Available from . on the Selection of Technical and Environmental Dangers. Berke- Anderson, M. and P. Woodrow. 1999. Rising from the Ashes: Devel- ley: University of California Press. opment Strategies in Times of Disaster. London: IT Publications. Eade, D. 1997. Capacity-Building. Oxford: Oxfam. Barnes, P. 2002. "Approaches to Community Safety: Risk Per- Eibenschutz, R. and S. Puente. 1992. "Environmental Degrada- ception and Social Meaning," Australian Journal of Emergency tion,Vulnerability,andUrbanDevelopmentinDevelopingCoun- Management 17 (1): 15­23. tries: A Conceptual and Operational Approach." In A. Kreimer Beall, J., O. Crankshaw, and S. Parnell. 2002. Uniting a Divided andM.Munasinghe,eds.,EnvironmentalManagementandUrban City: Governance and Social Exclusion in Johannesburg. London: Vulnerability. Washington, D.C.: The World Bank: 78­79. Earthscan. Fernandes, E. and A. Varley, eds. 1998. Illegal Cities: Law and Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Urban Changes in Developing Countries. London: Zed Press. Ritter. London: Sage. Fernandez, M., ed. 1999. Cities at Risk: Environmental Degrada- Beck, U., A. Giddens, and S. Lash. 1994. Reflexive Moderniza- tion, Urban Risk, and Disaster. Quito: A/H Editorial. tion: Politics, Tradition and Aesthetics in the Modern Social Foreman, C. H. 1998. The Promise and Peril of Environmental Order. Cambridge: Polity. Justice. Washington, D.C.: Brookings Institution Press. Benson, C., J. Twigg, and M. Myers. 2001. "NGO Initiatives in Freire, M. and R. Stren, eds. 2001. The Challenge of Urban Gov- Risk Reduction: An Overview." Disasters 25 (3): 199­215. ernment. Washington, D.C. and Toronto: The World Bank Blaikie, P., T. Cannon, I. Davis, and B. Wisner. 1994. At Risk: Institute and The Centre for Urban and Community Studies, Natural Hazards, People's Vulnerability, and Disasters. London: University of Toronto. Routledge. Fuchs, R.J. 1994. "Introduction." In Fuchs, R.J., E. Brennan, J. Bolin, R. and L. Stanford. 1991. "Shelter, Housing and Recovery: Chamie, F. Lo, J.I. Uitto, eds., Mega-city Growth and the Future. A Comparison of U.S. Disasters." Disasters 15 (1): 24­34. Tokyo: United Nations University Press: 1­13. Disaster Risk Reduction in Megacities: Making the Most of Human and Social Capital 195 Fuchs, R.J., E. Brennan, J. Chamie, F. Lo, J.I. Uitto, eds. 1994. International Federation of Red Cross and Red Crescent Societies Mega-city Growth and the Future. Tokyo: United Nations Uni- (IFRC). 1999. Vulnerability/ Capacity Analysis. Geneva: IFRC. versity Press. Kaji, H. 1991. "Promotion of People's Participation in Disaster Giddens, A. 1992. The Transformation of Intimacy. Cambridge: Management: The Japanese Experience." In United Nations Polity Press. Centre for Regional Development (UNCRD), Regional Disas- Gilbert, A. 1994. The Latin American City. London: The Latin ter Management in the Latin American Region. Report and America Bureau. Summary of Proceedings of the UNCRD-CISMID Interna- Gilbert, A., ed. 1996. The Mega-city in Latin America. Tokyo: tional Workshop on Regional Disaster Management, Piura, United Nations University Press. Peru. Nagoya: UNCRD: 23­26, 222. Gilbert, R., D. Stevenson, H. Girardet, and R. Stren. 1996. Making Kleinenberg, E. 2002. Heat Wave: A Social Autopsy of Disaster in Cities Work: The Role of Local Authorities in the Urban Environ- Chicago. Chicago: University of Chicago Press. ment. London: Earthscan. Lee, Y. 1994. "Myths of Environmental Management and the Girardet, H. 2000. "Cities and the Culture of Sustainability." In Urban Poor." In R. Fuchs, E. Brennan, J. Chamie, F. Lo, and F. Dodds, ed., Earth Summit 2002. London: Earthscan: 202­211. J. Uitto, eds., Mega-city Growth and the Future. Tokyo: United Girardet, H. 1992. The Gaia Atlas of Cities: New Directions for Sus- Nations University Press: 390­411. tainable Urban Living. New York: Anchor Books. Lo, F. and Y. Yeung, eds. 1996. Emerging World Cities in Pacific GPG (Gauteng Provincial Government). 2001. "Stand Up and Be Asia. Tokyo: United Nations University Press. Counted." Gauteng News: Newsletter of Gauteng Provincial Gov- Mansilla, E. 2001. Riesgo y Ciudad. Mexico City: LA RED. ernment, July­August. Available from . Magardie, K. 2002. "Home Improvements in Alexandra." BBC GPG (Gauteng Provincial Government). 2002. Department of News. Available from . 2000/2001. Available from . Mitchell, J.K. ed., 1999a. Crucibles of Hazard: Mega-cities and Greater Johannesburg Metropolitan Council. 2000. "Report on Disasters in Transition. Tokyo: United Nations University the Interactive Planning Workshop for Johannesburg." In World Press. Bank, Upgrading Urban Communities: A Resource Framework. Mitchell, J.K. 1999b. "Natural Disasters in the Context of Washington, D.C.: World Bank. Available from . cities and Disasters in Transition. Tokyo: United Nations Uni- Hardoy, J. and D. Satterthwaite. 1989. Squatter Citizen: Life in versity Press: 15­55. the Urban Third World. London: Earthscan. Morrow, B. H. 1999. "Identifying and Mapping Community Vul- Hardoy, J., S. Cairncross, and D. Satterthwaite. 1990. The Poor nerability." Disasters 23 (1): 1­18. Die Young: Housing and Health in Third World Cities. London: Mustow, S., ed. 1995. Megacities: Reducing Vulnerability to Nat- Earthscan. ural Disasters. London: Thomas Telford. Hardoy, J., D. Mitlin, and D. Satterthwaite. 1992. Environmental Plummer, J. 1999. Municipalities and Community Participation: A Problems in Third World Cities. London: Earthscan. Sourcebook for Capacity Building. London: Earthscan. Harriss, J. 2001. Social Capital and the De-Politicization of Devel- Puente, S. 1999a. "Natural Disaster Risk Management and the opment. London: Anthem. Geography of Urban Vulnerability: the Case of Mexico City." Heijmans,A.andL.Victoria.2001.Citizen-BasedandDevelopment- Annual Meeting of the Association of American Geogra- Oriented Disaster Response. Quezon City: Center for Disaster phers, Honolulu, Hawaii. Preparedness. Puente, S. 1999b. "Social Vulnerability to Disasters in Mexico Higashiura, H. 1994. "Community Vulnerability and Local Knowl- City: An Assessment Method." In J. Mitchell, ed., Crucibles of edge." The World Conference on Natural Disaster Reduction, Hazard: Mega-cities and Disasters in Transition. Tokyo: United Yokohama, May 23­27, 1994, Technical Session A: Vulnera- Nations University Press: 295­334. ble Communities. Rowell, J. 2002. Mainstreaming Risk Management in the Slums of International Decade for Natural Disaster Reduction (IDNDR). Mumbai: A Livelihood Perspective. Final Dissertation, MSc in 1996. Cities at Risk: Making Cities Safer ... Before Disaster Strikes. Disaster Management. Cranfield Disaster Management Centre, Geneva: IDNDR. Cranfield University. 196 Building Safer Cities: The Future of Disaster Risk Sanderson, D. 2000. "Cities, Disasters and Livelihoods." Envi- Management."InT.Inoguchi,E.Newman,andG.Paoletto,eds., ronment and Urbanization 12(2): 93­102. Cities and the Environment: New Approaches for Eco-Societies. Satterthwaite, D. 1998. "Why Half the World's Population does Tokyo: United Nations University Press. not Live in a City." In IFRC, World Disasters 1998. Oxford: Westfall, M. 2001. "On-Site Integrated Urban Upgrading for Vul- Oxford University Press: 11 (box 1.2). nerable Slum Communities of Payatas." Project Overview. Sen, A. 1993. "Capability and Well-Being." In M. Nussbaum and Manila: Asian Development Bank. A. Sen, eds., The Quality of Life. Oxford: Clarendon: 30­54. Wisner, B. 1995. "Bridging `Expert' and `Local' Knowledge for Sinha, R. 2001. "Report on a Meeting of Government, Business, Counter-Disaster Planning in Urban South Africa." GeoJour- and NGOs to Consider the Mumbai Disaster Preparedness nal 37( 3): 335­348. Plan." Available from . Wisner, B. 1998. "Marginality and Vulnerability: Why the Home- Slovic, P. 1993. "Perceived Risk, Trust and Democracy." Risk less in Tokyo Don't `Count' in Disaster Preparations." Applied Analysis. 13 (6): 675­682. Geography 18(1): 25­33. Slovic, P. 2000. The Perception of Risk. London: Earthscan. Wisner, B. 1999a. "Urban Social Vulnerability in Greater Los Takahashi, S. 1998. Social Geography and Disaster Vulnerabil- Angeles." Annual Meeting of the Association of American Geo- ity in Tokyo. Applied Geography 18 (1): 17­24. graphers, Honolulu, Hawaii. Takahashi, S. 1999. "Urban Social Vulnerability to Hazards in Wisner, B. 1999b. "There Are Worse Things than Earthquakes: Tokyo." Annual Meeting of the Association of American Geo- Hazard Vulnerability and Mitigation Capacity in Greater Los graphers, Honolulu, Hawaii. Angeles." In J.K. Mitchell, ed., Crucibles of Hazards: Mega-cities Takahashi, S. Forthcoming. "Coping with Urban Social Vul- and Disasters in Transition. Tokyo: United Nations University nerability to Hazards in Tokyo: Current Status of Disaster Press: 375­427. Mitigation Plans and its Implications." In S. Sassen, ed., Wisner, B. 2003a. "Urban Social Vulnerability to Disaster inGreater Sustainable Cities. Encyclopedia of Life Support Systems. Paris: Los Angeles." In S. Sassen, ed., Sustainable Cities. Encyclope- UNESCO. dia of Life Support Systems. Paris: UNESCO, forthcoming. Tayag, J. 1999. "Urban Social Vulnerability to Hazards in Metro Wisner, B. 2003b. "The Communities Do Science! Proactive and Manila." Annual Meeting of the Association of American Contextual Assessment of Capability and Vulnerability in the Geographers, Honolulu, Hawaii. Face of Hazards." In: G. Bankoff, G. Frerks, and T. Hilhorst Toepfer, K. n.d. "The Urban Revolution." U.N. Center for Human eds., Vulnerability: Disasters, Development and People. London: Settlement (UNCHS/Habitat). Available from . Wisner, B., P. Blaikie, I. Davis, and T. Cannon. 2003. At Risk: Townsend, R. 2000. The Aztecs. London: Thames. Natural Hazards, People's Vulnerability, and Disasters. 2nd edi- Tokyo Metropolitan Government (TMG). 1995. Tokyo and Earth- tion. London: Routledge. quakes. Tokyo: TMG. World Bank. 2001. "The Johannesburg Story." In Upgrading Low- Uitto,J.I.1998."TheGeographyofDisasterVulnerabilityinMegac- income Settlements: A Regional Roundtable for Africa. October ities:ATheoreticalFramework."AppliedGeography 18(1):7­16. 3­5, 2000, Johannesburg, South Africa. Available from United Nations. 1999. "Key Findings." World Urbanization . Prospects: The 1999 Revision. New York: United Nations. Wynne, B. 1987. Risk Management and Hazardous Waste: Imple- United Nations. 2002. "Key Outcomes of the Summit." Johan- mentation and the Dialectics of Credibility. Berlin: Springer-Verlag. nesburg Summit 2002: World Summit on Sustainable Develop- Yodmani, S. 2001. "Disaster Risk Management and Vulnerabil- ment. Johannesburg, South Africa, August 26­ September 4, ity Reduction: Protecting the Poor." Paper presented to the 2002. Available from . Asian Development Bank's Asia and Pacific Forum on Poverty. Velasquez, G., J. Uitto, B. Wisner, S. Takahashi. 1999. "A New Available from Approach to Disaster Mitigation and Planning in Megacities: . Chapter 14 Living with Risk: Toward Effective Disaster Management Training in Africa Prvoslav Marjanovic and Krisno Nimpuno In recent years, the world has witnessed a succession The African Ministerial Statement to the World Summit of disasters--floods, wildfires, storms, earthquakes, vol- on Sustainable Development states that the increased canic eruptions, and landslides. These claimed many incidence of natural disasters in Africa poses a major thousands of lives, caused material losses in the bil- obstacle to the African continent's efforts to achieve sus- lions of dollars, and inflicted a terrible toll on devel- tainable development, especially in view of the region's oping countries in particular, where disasters divert insufficient capacities to predict, monitor, handle, and attention and resources from development needed mitigate natural disasters. Reducing the vulnerability of desperately to escape poverty. the African people to natural disasters and environmental It is a well-known fact that today's disasters are risksismentionedasarequirementtoachievethepoverty- often generated by, or at least extended by, human activ- reduction goals of the Millennium Declaration alongside ities. At the most dramatic level, human activities are other basic requirements, including economic growth, changing the natural balance of the planet and inter- access to sources of energy, and basic health services. fering with the atmosphere, oceans, polar ice caps, forest Extreme weather events such as floods and droughts cover, and the pillars that make our world what it is. induced by climate change are singled out. Population growth and associated pressures cause Following the meeting of the Council of Ministers more people to live in flood plains or in areas prone to of SADC (South African Development Community), landslides. Inadequate land-use planning, poor envi- held in August 2000 in Windhoek, it became apparent ronmental management, and a lack of appropriate insti- that the SADC region is not well prepared to deal with tutional and legislative arrangements increase the risk disasters and that a major drive toward an effective train- and multiply the effects of disasters. ing program in disaster management and vulnerability Living with risk is the order of the day, and we must reduction is necessary. The University of the Witwa- learn to reduce these risks through appropriate meas- tersrand has responded to this need and, in partnership ures focused on planning, forecasting, and mitigation. with Disaster and Emergency Reference Centre from the We need to build a world of resilient people, commu- Netherlands (DERC) and the National Disaster Man- nities, and nations. agement Centre of South Africa (NDMC), has initiated In recent years, there has been a major conceptual activities to support the needs of the region and the shift in how people seek to cope with disasters that arise soon-to-be-established Disaster Management Facility from natural hazards. While humanitarian response within the SADC structure. capacities are vital and need continued attention, the The creation of the disaster-management structure focus on addressing risk underlines the recognition that withinSADCisprimarilydrivenbytheeventsthatunfolded human intervention designed to reduce the vulnera- in February 2000 with the advent of Cyclone Eline. In bility of communities and assets can reduce the impact theinitialphaseofthefloodingdisaster,SADCwasreduced of disasters. Gradually, environmental and development to a spectator while the international aid providers took stakeholders are becoming more involved in the man- center stage. Even when SADC managed to intervene, agement of risk and vulnerability reduction due to their the thrust was not as focused as it could have been. Such close interaction with natural resources management. alevelofregionalunpreparednessisuntenableandshould 197 198 Building Safer Cities: The Future of Disaster Risk not be repeated in the future. The general view emerged the National Disaster Management Centre to address that SADC ought to have a disaster-management struc- the training and capacity building needs that have ture backed by vibrant and sustainable support from been identified. which to draw resources to assist stricken member states. The events in February 2000 also demonstrated that the focus on relief will not be sufficient to provide long-term The Status of Disaster Management and sustainable means for dealing with disasters. The in SADC Countries preparedness measures need to focus on vulnerability reduction; relief should not be viewed as a panacea to all Before one can assess the training and capacity-building problems associated with major disasters. needs for disaster management, it is necessary to review While there is always a crucial role to be played by the current status of disaster management in some SADC the international community in its humanitarian assis- countries. tance following major disasters, it should be stressed that SADC countries do not have intimate experience Zimbabwe in dealing with aid agencies. Obstructive "red tape," with officials clinging to rules and regulations that · Major hazards include drought, flooding, epidemics, frustrate the efforts of the aid provider, is common, public transportation accidents, industrial accidents, and failure to realize the magnitude of the disaster skills-- forest fires, and environmental degradation. whicharedifferentfromthe"businessasusual"approach-- · The central government initiates disaster-prepared- thatareneededisoftenatthecoreofeffectivemanagement ness programs through relevant sector ministries with of disasters. Equally, the skills for preparing disaster local administrations taking the responsibility for management plans are just as lacking, despite a legal implementing and maintaining their effectiveness. requirement for such plans to be prepared in many of · The Ministry of Local Government, Public Con- the SADC countries. The assessment of risk and damage struction and National Housing is charged with the is yet another area in which there is a lack of capacity coordination role, as stated under the Civil Protec- within SADC. tion Act No. 5 of 1989. There is therefore a profound need for an effective · Civil Protection Act No. 5 also established a National and accelerated training and education program to ensure Civil Protection Fund, which receives money from adequate preparedness for natural disasters, which, sta- the government and the public to be applied to tistically, are on the increase. South Africa has pre- enhancing civil protection measures through research, pared a Disaster Management Bill, which was training, and the acquisition of materiel, among other promulgated and has become a national Disaster Man- applications. This act is under review and will address agement Act. The principal agency for the implemen- the legislative gaps in the areas of fire and ambulance tation of the Act is the National Disaster Management services, and the enforcement of sectoral prepared- Centre. As a part of this initiative, a Disaster Manager ness planning. has been appointed in each of the local authorities in · A National Civil Protection Coordination Commit- South Africa, with the task of preparing a Disaster tee is responsible for civil protection functions and Management Plan for his area. It quickly became evi- is comprised of senior officers selected from gov- dent that the capacity to meet the requirements of the ernment ministries and departments, parastatals, and new policy and legislation is lacking and particularly NGOs. Similar multisectoral representation is main- so in areas most vulnerable to disasters. It became evi- tained at the provincial and district levels. All three dent that training and capacity building in disaster man- levels have functional subcommittees with respon- agement are a prerequisite to effective implementation sibilities according to specialty. of the new policies and legislation. · Mission Statement: "to provide for and ensure optimal This paper presents the initial results of a joint ini- emergency preparedness and disaster prevention at tiative between the University of the Witwatersrand and the individual, community, sectoral, local authority, Living with Risk: Toward Effective Disaster Management Training in Africa 199 and national level through regulatory mechanisms reforest lands, improve storage and rainwater harvest- and coordinated strategic planning for emergencies." ing, ensure sound management of the environment, · Constraints and needs: equipment, including rescue, and discourage people from living in flood plains. communications, and early warning; inadequate fund- · The government has established strategic grain ing; legislative gaps; and staff development. reserves. · With UNDP assistance, Tanzania has prepared a draft disaster-management policy and enabling legislation; Zambia carried out a training needs assessment and risk assess- · Types of disasters include drought, floods, epidemics, ment/vulnerability analysis; developed a framework refugees, fires, pest infestations, internally dis- for a disaster-management plan; created an MIS; placed persons, persons with HIV/AIDS, transportation and started a public awareness program. accidents, water hyacinth, industrial accidents, and · Further requirements for disaster management in mining accidents. Tanzania include strengthening the disaster-man- · Institutional framework allows for a national disas- agement department through training; preparing ter committee, technical committees, specialized sub- sectoral-, regional- and district-level disaster man- committees, provincial disaster committees, and agement plans; inventorying resources; mapping district disaster committees. temporary shelter areas and sources of relief goods; · DMMU consists of a secretariat and three regional increasing public awareness; and establishing national offices with the functions of coordination, vulnera- and regional emergency committees. bility/risk assessment, training/capacity building, advocacy, stockpiling, resource mobilization, needs South Africa assessment, impact assessment, and research. · TheDisasterManagementTrainingCentreoffersathree- · Events/disasters occurring in South Africa include week disaster-management course at the SADC-level. droughts, floods, wildfires, fires in informal settle- · Funding: annual budget allocation of approxi- ments, industrial accidents, displacement, high winds, mately US$200 million. and disease outbreaks and epidemics. · Stakeholders include the government, United Nations · In January 2000, South Africa launched a "White system, NGOs, community, and the private sector. Paper on Disaster Management," which presented an approach that prioritizes prevention rather than focus- ing primarily on relief and recovery efforts. The White Tanzania Paper emphasizes the importance of preventing · Major hazards faced are drought, floods, landslides, human, economic, and property losses, and avoid- epidemics, pest infestation, earthquake, accidents, ing environmental degradation. The White Paper also fire, and civil strife, which create flows of refugees placed special emphasis on pursuing international into Tanzania. and regional cooperation with vigor. · Tanzania has legislation covering disaster manage- · The Disaster Management Bill was published for ment; however, it needs an amendment in order to public comment on 21 January 2000, while the focus on the management of disaster activities final draft of the bill was introduced into parliament rather than on relief coordination. in early 2001. · Other government approaches to preparedness include · The National Disaster Management Act was adopted crop and food security monitoring by the Ministry by parliament in 2002. of Agriculture and the Department of Meteorology; · The cabinet established an Inter-Ministerial Commit- a national policy to conserve land from overgrazing, tee on Disaster Management in 1997, which consists soil compaction, and erosion; a national environ- of nine cabinet ministers and their deputies. An Inter- mental policy to control degradation and raise public departmental Disaster Management Committee was awareness; and policies to control population growth, also established, comprising representatives of national 200 Building Safer Cities: The Future of Disaster Risk and provincial government departments, national · Mozambique has suffered from a wide range of nat- associations and institutions, the private sector, and ural and man-made disasters. Since its independence NGOs. Similar structures exist on provincial and local from Portugal in 1975, Mozambique has been victim government levels. to drought, floods, cyclones, massive war-provoked · The National Disaster Management Centre was estab- population displacements, coastal oil spills, erosion lished in 1999 and continues to have responsibility and landslides, wildfires, pests, epidemics (cholera, for coordinating actions during all phases of a dis- bubonic plague, meningitis), forest fires, and large aster, with a focus on information management. transportation accidents. The three hazards men- · Local authorities are responsible for disaster pre- tioned first--droughts, floods, and cyclones--are a paredness, but may seek assistance at the provincial priority for disaster reduction. level if resources are not sufficient at the local level. · After successive floods in the Limpopo Valley · Shortcomings and/or needs include stockpiling equip- (1976­77), the Buzi and Pungue Valleys (1978), and ment; training and capacity building; simulation exer- the Zambezi Valley (1979), it became evident that cises; vulnerability assessments; and communication Mozambique needed to develop an integrated Dis- and early-warning systems. aster Management Policy that combined preven- tion, mitigation, preparedness, and response in the context of post-war reconstruction and development. Seychelles Mozambique is prone to continuing and recurrent · TheNationalDisasterCommitteeischairedbythePrin- natural disaster threats. Lack of planning and prepa- cipal Secretary of Environment and falls under the ration in anticipation of such hazards increases the aegis of the Cabinet of Ministers, which is lead by the loss of life and exacerbates the vulnerability of the vice president. The committee consists of senior mem- population when disaster strikes. Mozambique cur- bersofvariousministries,NGOs,andtheprivatesector. rently lacks an integrated disaster-management policy · Seychelles has a national disaster plan, which out- and structure. lines the roles and responsibilities of each relevant · The now-established institutional framework provides organization. Additionally, several ministries have for an interministerial Disaster Management Council their own response plan, i.e., the health sector. chaired by the prime minister. A free-standing National · Weaknesses: not much emphasis is placed on pre- Institute for Disaster Management, INGC, is the paredness, and the concept of disaster management secretariat for the council and the principal body for is not well established; public awareness of disaster disaster management at a national level. Provincial management is low. and local-level structures will be established. INGC · Positive aspects: good communication systems; small replaces the former council, but still reports to the population; some personnel have been trained in dis- Ministry of Foreign Affairs and Cooperation. aster management; the risk of disasters is minimal. · INGC is responsible for developing a Disaster Man- · Seychelles would like the following to be addressed agement Plan that includes prevention, mitigation, on a regional scale: organize disaster-management preparedness, and response. In the absence of a Dis- training; evaluate disaster-management capabilities aster Management Plan, the roles of various partners in each country and carry out vulnerability assess- in civil protection have so far not been defined. Con- ments; serve as focal point for information collection sequently, there is little coordination between them and dissemination of knowledge related to disaster in mitigation matters. The coordination mandate of management; organize simulation exercises. INGC is now seen as taking responsibility for control of all bilateral- and multilateral-supported disasters. · Although INGC keeps these institutions informed of Mozambique current disaster issues, there is little tangible activity · Mozambique is undergoing a process of rapid sociopo- that coordinates disaster-prevention programs. The litical, economic, and institutional transformation. various ministries continue their individual mitigation Living with Risk: Toward Effective Disaster Management Training in Africa 201 efforts without the benefit of a civil-protection frame- Malawi work or a disaster-response plan. · The Department for Coordination and Disaster Man- · The Ministry of Public Works tries to follow and pro- agement, Relief, and Rehabilitation has the respon- mote the existing building and planning regulations, sibility for disaster-management issues in Malawi; andtheMinistryofHealthstrengthensitsgeneralpublic the most frequent hazards are flooding, drought, epi- healthservices,whichcurrently,underdonorpressure, demics, and refugee flows. focus above all on AIDS prevention information. · Malawi has national legislation establishing the insti- · The structure of disaster responsibilities in Mozam- tutional framework for disaster management. The bique leaves much to be desired. INGC, the inter- legislation also provides for a National Disaster Pre- governmental committee, is the coordinator for sector paredness and Relief Committee (NDPRC), which ministries, U.N. agencies, provinces, and the pri- includes relevant ministries and departments, as well vate sector, and reports to the prime minister through as some NGOs. This committee provides policy guid- the Ministry of Foreign Affairs. ance to the Commissioner for Disaster Preparedness, · Seven task force subcommittees exist and have well- Relief, and Rehabilitation. defined roles and responsibilities. · District Development Committees are the local gov- · Contingency planning, from the national to the dis- ernment equivalent of the NDPRC; the chief execu- trict levels, is in place; the degree of its effectiveness tive of each district assembly is responsible for the is not known. local disaster management program, assisted by NGOs. · With the help of UNOCHA, the Government of Malawi Mauritius has prepared a draft national disaster-management plan; an accompanying manual is planned when · The dominant threat, by far, is cyclones--few miss funds are available to complete the planning process. the island nation. Consequently, the government has · The government is now turning its attention to dis- developed sophisticated planning, preparedness, early aster mitigation, prevention, and reduction. This warning, and responder responsibility systems. Despite approach recognizes that a disaster can wipe out the frequency and severity of cyclones in Mauritius, decades of development efforts. no lives have been lost in many years. · Malawi's primary short-term needs are in the area of · Mauritius uses a standardized system of cyclone ter- communications: radios, communications vehicles, minology and classes of cyclone warnings, familiar and satellite phones. to all citizens. · The government convenes a planning session of rel- Namibia evant ministries, departments, and essential services, which comprise the Central Cyclone Committee. This · National disaster preparedness in Namibia is in an annual meeting is two weeks long. Local cyclone emerging stage, with much left to be done. committees also meet annually. · Since drought is endemic to Namibia, the country has · The warning system for the population is a simple a National Drought Policy. Some strategies being imple- system of red or blue (termination) flags, displayed mented under that policy include pest, drought, and on public buildings, police stations, fisheries posts, flood management. The government has established and the like. a National Drought Fund for implementation. · Relief operations are conducted by the government · The Ministry of Health and Social Services runs a through a standing cabinet committee. Disaster Surveillance and Epidemic Management Sec- · Other hazards include torrential rain and flooding, tion; it also conducts public awareness campaigns as well as landslides. on epidemics such as HIV/AIDS. · Mauritius has a thriving NGO community and many · Environmental disaster mitigation activities are being U.N. specialized agencies; therefore, it has a congen- carried out by the Ministry of Environment and ial climate for strengthening disaster preparedness. Tourism; some national policies are in place. 202 Building Safer Cities: The Future of Disaster Risk · The Directorate of Maritime Affairs is implementing · The disaster management structure for Botswana is a national oil-spill contingency plan. at three levels. · A search and rescue center is being established, less- · Theneedsidentifiedare:afinalizeddisasterprofileand ening Namibia's dependence on South Africa for National Disaster Management Plan, which will form airborne operations. the basis for a legal framework; capacity building at · Twenty-three regional trainers have been trained in the national and local levels; training; development disaster preparedness and management; other of a strategic plan and budget; and decentralization training has also been carried out. of basic stockpile items. · Namibia's needs in disaster management include train- ing; equipment, including boats and helicopters; and Angola medical equipment. · Namibia feels that SADC member nations would · Very little is in place for effective disaster manage- benefit from a common ground (terminology, etc.) for ment in Angola. An ad-hoc technical group has training. been created to assist in preparing a disaster man- agement plan. · In 1999, Angola suffered damage from heavy rains in Botswana Benguela, and desertification and drought affected the · The body responsible for Disaster Management in provinces of Luanda, Bye, Moxico, Kuando-Kubango, Botswana is the National Disaster Preparedness and Namibe, as well as other parts of the country. The Committee, which is chaired by the Deputy Perma- overuse of land causes landslides in Moxico and the nent Secretary. Representatives include the Deputy northern and southern sides of Luanda. Permanent Secretaries from the line ministries, the · The government has made available some financial Botswana Defense Force, the Botswana Police Ser- resources to fund small projects, but the amount of vice, and the Botswana Red Cross. The National funding is insufficient. Disaster Management Office was established in 1998, · Angola supports a regional coordinating mecha- and is the secretariat to the NCDP. Its mandate is to nism for disaster management. coordinate all disaster management activities. Dis- It is also of interest to summarize the recommenda- trict Disaster Management Committees chaired by tions of the SADC Disaster Management Steering district commissioners are also in place. Committee to the Council of Ministers made in 2002. · Botswana is characterized by an inadequate level of The Terms of Reference for the SADC Disaster Manage- preparedness; inadequate search and rescue capa- ment Mechanism should be as follows: bility; poor information management systems; and · Act as an information and communications hub link- bureaucratic red tape. ing individual national centers, regional sector struc- · The most common disaster in Botswana is drought. tures, related NGOs, training and research institutions, Botswana is also reported to have the highest rate of and other existing capacities in a two-way commu- HIV/AIDS infection in the world, which makes man- nication system agement of the pandemic a government priority. In · Facilitate regional collaboration for all aspects of February 2000, the cyclones originating in the Mozam- disaster management bique channels resulted in the heaviest floods ever · Coordinate any regional response to disasters as recorded in Botswana's history. requested by member states · Other disaster include floods, veldt fires, epidemics, · Promote and, where necessary, implement and manage animal diseases, vectors such as malaria-carrying disaster-management programs within the region in mosquitoes and tsetse flies, pest infestations, cyclones, collaboration with member states strong winds, earthquakes, transport accidents, · Mobilize resources, facilitate appeals for assistance, refugee influx, industrial accidents, and chemical and undertake fundraising for disaster management spills. and disaster response Living with Risk: Toward Effective Disaster Management Training in Africa 203 · SADC should develop a protocol on disaster man- or program can meet all needs. It is thus evident that a agement in the region cooperative network approach is essential to ensure that · SADC should develop disaster-management standard effective training and capacity building in disaster man- operating procedures agementcanbedevelopedandimplemented.Inaddition, · The Council of Ministers should ensure that each it is also evident that any training and capacity-building member nation clearly indicates one focal point for effort has to be established bearing in mind two main disaster management in each country factors: · Member states should promote and encourage the · The urgent short-term need to provide basic train- active participation of civil society and the private ing for all those involved in the first line of disaster- sector in all aspects of disaster management managementactivitiesandtodosoasquicklyaspossible · Member states should be encouraged to promote · The need to establish a stable and sustainable train- in-country seminars and workshops to capacitate ing and capacity-building program to respond to local communities and stakeholders the long-term needs of the region. · Member states should be proactive and should sen- The first of the above factors requires an extensive sitize government officials and communities to the series of short courses to address short-term needs and importance of disaster prevention or mitigation and would focus on fundamentals of disaster management preparedness from a practical point of view, while the second factor · The SADC region should work collectively to under- requires the establishment of formal regional training take staff development, to include: programs in disaster management, which will lead to º Creating a coordinated staff training program in formal qualifications at undergraduate and postgradu- disaster management, taking appropriate meas- ate levels. The short-term needs can only be met through ures to develop and retain staff in order to ensure "mobilization" of all available resources and will require institutional memory and providing the resources international cooperation and regional effort. and facilities that allow disaster management The long-term needs, on the other hand, require thor- personnel to maximize their performance ough planning and development and are best suited to º The region should work collectively to create a a network approach similar to the one taken by Water- common language and procedures for disaster Net (a regional initiative for a regional Masters Program management in Water Resources Management) to develop a regional · SADC should commit to assisting member states in educational program in Disaster Management and Vul- undertaking disaster vulnerability assessments and nerability Reduction. risk mapping The long-term training and capacity-building objec- · SADC member states should ensure that there is close tives should be: collaboration between and among SADC institutions · Reducing the incidence and impact of crisis and dis- and other organizations involved in similar disaster aster occurrences management activities in the region · Eliminating risks and vulnerability to such events · Member states should promote the appropriate, timely, · Promoting effective national and regional strategies and effective dissemination of accurate information in crisis and disaster prevention, preparedness, mit- to all stakeholders, particularly disaster managers igation, response, and recovery and high-risk communities. · Efficiently coordinating and collaborating all phases of crisis and disaster management, between and among Training and Capacity-Building Needs for national and international partners Disaster Management in SADC · Making effective and efficient institutional and leg- islative arrangements From the above summary, it is evident that the training · Developing appropriate planning and intervention and capacity building needs are extensive, diverse, and strategies focused on vulnerability reduction rather multidisciplinary, and that no single training institution than relief 204 Building Safer Cities: The Future of Disaster Risk · Fostering international cooperation and sharing of The Potchefstroom University Initiative resources and experiences The African Centre for Disaster Studies (ACDS) was · Building effective information bases and other resource established in January 2002 at the Potchefstroom Uni- management versity for Christian Higher Education within the School · Utilizing regional resources and focusing efforts on for Social and Government Studies. The explicit aim of developing sustainable regional capacity to deal with the ACDS is to address the need for education and disaster management. research in disaster-related activities within Southern Africa and the wider African context. At the time of writ- ing, the ACDS center is planning to offer: Disaster Management Training Initiatives in South Africa Certificate Course in Disaster Studies (to start in 2003) The objective of the certificate course in Disaster Stud- In South Africa there is a widely recognized need for ies is to provide the student with skills and competency disaster management training and capacity building. in the management of hazards, risks, vulnerability, dis- Three main initiatives are worth mentioning: the Uni- asters, and their associated secondary effects. The main versity of Free State Initiative, the Potchefstroom Uni- emphasis of the course is on disaster-risk reduction versity Initiative, and the University of the Witwatersrand through the use of vulnerability-reduction techniques Initiative. and hazard assessment and mitigation, within the con- text of sustainable development and sustainable liveli- The University of Free State Initiative hoods. The program lasts one year and consists of the following modules: Disasters: A Theoretical Perspective; The University of Free State has introduced formal edu- Disaster Risk Reduction; Disaster Planning; Disaster cation (Magister and an Advanced Diploma in Disas- Recovery; Disasters and Sustainable Development; Dis- ter Management) as well as informal training and an aster Information Management and Communication. education program (short courses). The formal pro- gram has three formal contact sessions each year of five Short Courses in All Aspects of Disaster Studies (to days each. During the first contact, session students start in 2003) The short courses currently planned are: are oriented and receive all course material for the Basic Course in Disaster Management; Integrated Devel- first year. The course consists of eight compulsory opment Planning and the Municipal Disaster Plan; courses, two electives, and a research project reported Disaster Preparedness Planning. in a mini-dissertation format. The course can be taken over a minimum period of two years (full-time). How- An Undergraduate Degree in Disaster Management (to ever, students will be allowed to take the course over start in 2004) Details of the program are not known at a three-year period (part-time). Compulsory courses this time. are: Research Methodology, Hazards of Disaster Man- agement, Strategic Management, Advanced and Spe- The University of the Witwatersrand Initiative cific Disaster Management Principles and Practices, Advanced Disaster Risk Management, Information Tech- The University of the Witwatersrand has responded to nology in Disaster Management, Public Health, and the urgent needs for training and capacity building in Management of Disasters (Natural and Man-made). disaster management by partnering with DERC and the Elective courses include: Trauma Management, Polit- NDMC to offer a series of short courses and develop a ical Strategic Planning, Information Management, Ethnic new postgraduate and research program in Disaster and Cultural Conduct, Media Liaisons, Environmental Management. Degradation, Disaster Vulnerability, and Risk Assess- TheDisasterPreparednessTrainingProgram(short ment. Admission is in accord with the university's admis- courses) has been developed as a basic training program sions policies. for municipal and district disaster managers. The first Living with Risk: Toward Effective Disaster Management Training in Africa 205 training cycle of six modules, developed for the munic- Module 2. Hazards and Vulnerabilities ipal disaster managers of Limpopo province, was imple- mented in 2002, January through October. The next cycle The module reviews hazard profiles and possible effects of short courses will be offered during 2003 and regu- and impacts of hazards. It shows that hazards are not larly thereafter. The short courses have been certified by the cause of disasters, and establishes the causative links the NDMC and are currently in the process of verifica- between disasters and vulnerabilities. These vulnera- tion by the South African Qualifications Authority. bilities prove to be the causes of natural disasters, tech- The six training modules aim at developing a good nological disasters, and civil strife. De facto disasters understanding of the principles of disaster reduction are failures in development planning. They are events and impart skills in disaster mitigation as well as in that can be avoided, which is the essence of disaster disaster response. Three modules address mitigation reduction. planning and three modules focus on emergency For each hazard a profile is presented, related to loca- response. The program is also suitable for senior NGO tion, frequency, and intensity, and the forces causing staff. The program consists of six modules of one week these hazards are explained. Threat recognition elements each, which are conducted at two months intervals. A are studied. Social, economic, physical, and managerial short description of each of the modules is given below: vulnerabilities to various hazards are reviewed, with a variety of case studies. The course reviews administra- tive, technological, legal, and cultural aspects of disaster Module 1. Introduction to Disaster Management reduction and emphasizes the importance of community participation in all public efforts in this field. Vulnerabil- This module places disaster management in the frame- ity mapping is an important activity for guiding disaster- work of development planning and identifies strategy reduction efforts, and remote sensing as well as field options to reduce disasters. Introductions to the main monitoring are presented as options in this respect. natural hazards, to conflict and displacement-related emergencies, and technological hazards are presented. Introductions to natural disasters cover the underlying Module 3. Mitigation Planning phenomena of meteorology, hydrology, and seismol- ogy that pose hazards. An overview is given of the legal The principles of disaster mitigation, in particular those and administrative measures that societies take to ensure related to prevention and preparedness, are studied as safe and sustainable development. Public perceptions the main tools in disaster reduction. The focus of these of emergencies and existing disaster myths and their efforts is to address vulnerabilities. The historical process influence on emergency responses are studied. Human of devising mitigation strategies is illustrated by vari- behavior, both individually and collectively, form a major ous case studies and demonstrates that disasters can be factor in the development of disasters. The critical fac- prevented. Hazard-specific preparedness and mitiga- tors affecting human health during emergencies are pre- tion measures are reviewed for relevant locations. Such sented and mitigation options are reviewed. Several case mechanisms relate to international monitoring (e.g., studies of major emergencies are presented. by remote sensing), by national agencies (e.g., volcano The various phases of disasters--prevention, mitiga- observatories), or by community organizations (e.g., tion, preparedness, emergency response, relief, recov- river level monitoring or pre-cyclone shuttering of build- ery, and rehabilitation--are reviewed and the roles and ings). Specific impacts of hazards are countered by actors in disaster interventions are identified. The various protective measures, while monitoring mecha- multi-disciplinary nature of disaster management, in nisms are developed for threat recognition and warning. particular,earthsciences,physicalplanning,publicadmin- Thepurposeandeffectofmitigationmeasuresarereviewed. istration, sociology, and health, is discussed. The common Choices and priorities in mitigation are identified and options for disaster reduction are presented. The factors elements of mitigation planning are reviewed in rela- that increase and decrease disasters' impacts are analyzed. tion to the community in question. The training needs 206 Building Safer Cities: The Future of Disaster Risk of such actors are assessed and plans for appropriate are studied. Much attention is paid to the environmental materials are developed, while the role of mobilization impact of emergency settlements. Site selection, self-help and public information is stressed. A collective memory housing, phased infrastructure, and central services and hazard consciousness are needed at all levels. are covered, with particular attention to site selection, Mitigation planning is presented as a variety of safety availability of building materials, and water supply. The measures, which depend on a culture of safety and UNCHS manual for resettlement forms the basis for this participation. section. Camp management issues, logistics, security, and social services are addressed, and relief programs of NGOs in food supply, fuel, water and sanitation, edu- Module 4. Emergency Response cation, and health services are covered. Preparedness to take action when hazard threats occur is the prime instrument in reducing impacts. Develop- Module 6. Developing Disaster Plans ment of response capacities and the availability of logis- tics and supply resources are fundamental. Planning of Disaster plans are administrative instruments for the intervention alternatives, schemes for management of effective deployment of the necessary response resources, public behavior, and the determination of various alert through command and control measures. They clarify phases allow organized interventions under condi- mandates and clear lines of command. The plans iden- tions of potential chaos. Coordination between secu- tify which resources are needed, where they may be rity, rescue, and logistics services, and public initiatives found, and how they can operate smoothly in a coor- and responses are to be vested in an agreed decision- dinated response. They confirm collaboration agree- making process. Crisis management emerges therefore ments that in times of emergency make all necessary as an important technique to maximize the effect of the resources available for deployment and stipulate req- allocated resources that can bring order into potential uisition procedures. Both government and nongovern- conditions of chaos. The importance of establishing ment agencies are included, while the necessary responses Emergency Operation Centers is stressed. The role of from the affected populations are channeled and con- relief, especially of international NGOs, is analyzed, trolled. Communication systems are of critical impor- with particular attention to its disruptive effects on the tance and allow a rapid decision-making process that regular development process. The differences between facilitates smooth operations. Public regulations and a local and international relief are studied and mecha- legal basis for the plan are essential, but smooth imple- nisms to reduce negative effects are discussed. The mobi- mentation of the plan depends above all on good lization of local response capacities is a central theme. cooperation and trust. Mitigation plans are incremental development plans for decreasing, and possibly eliminating, vulner- Module 5. Displacement and (Re-) Settlement abilities. They form a key element in development plan- ning, but require constant improvement and monitoring Refugee influxes and various types of internal displace- by safety specialists. The appointment of municipal and ment have become major issues in Africa. The causes district mitigation officers is presented as an important of displacement due to war, civil strife, environmental element in disaster reduction. degradation, and economic changes are reviewed. The In addition to the above, a number of short courses social consequences and the issues of settlement, shelter, have been delivered outside of South Africa, including: health, and economic survival are reviewed. The present · Kenya: Introductory Disaster Management training policies and planning methods in emergency settlement courses were conducted in Kenya, for leading NGOs are studied, while options for sustainable settlements are in the field of relief, which expressed a desire to presented. Early site selection is a key activity in proactive develop proactive disaster reduction programs (e.g., planning. Innovative "Self Help" building technologies CARE Kenya and the Anglican Church). Living with Risk: Toward Effective Disaster Management Training in Africa 207 · Tanzania: Several training workshops were conducted and an in-depth analysis included in the curricula. The in two national educational institutes, with partici- program will be based on extensive materials developed pation from neighboring countries. A draft disaster by DERC, materials currently in use by the U.N. agencies plan was developed for the Ministry of Health. and universities in the United States, Asia, and Europe, · Mozambique: A training program is being developed, with appropriate modifications to suit South African and as part of an ongoing U.N. project on flood mitiga- regional needs. tion, addressing community needs in flood-prone In accordance with the requirements of the National areas. Surveys of vulnerability factors in informal Qualifications Framework and the South African urban settlements have been concluded. Qualifications Authority, the programs will be outcomes- · Ethiopia: In collaboration with U.N. Habitat, a train- based and will undergo normal verification and qual- ing course on earthquake-resistant adobe housing ity assurance processes. Entry into the program will be was conducted at the University of Addis Ababa. in accordance with the university's admissions poli- · Egypt: An introductory workshop on vulnerabilities cies, which also include recognition of prior knowledge. in informal settlements was conducted at the Center Master's degree program students will be required to of Planning and Architectural Studies, a nongovern- take additional courses and submit a research thesis mental human-settlements planning institute. equivalent to six months of full-time engagement. The postgraduate program in Disaster Management In addition to the three initiatives discussed, a number is focused on two main areas of activity: of technikons in South Africa are currently developing · A research program with a focus on understanding appropriate disaster-management programs, some of disasters in Africa which will be offered through distance learning. · A graduate diploma and a Master's Program in Dis- The University of Cape Town also has an active pro- aster Management. gram in disaster management but details about it were The research program was initiated during 2002 with not available at the time of writing. support from the Ford Foundation. Two research proj- ects have begun: droughts in Southern Africa and wild- fires in Southern Africa. Two Masters of Science theses International Network on Disaster Management are expected to be completed by the end of 2002 and Training in Africa--DIMITRA one doctoral thesis is to be submitted in 2003. The research program is expected to grow in other areas of The International Network on Disaster Management interest and will include but not be limited to prob- Training in Africa--DIMITRA--is an initiative aimed at lems such as information management and dissemina- mobilizing the limited resources in Africa and extend- tion, remote sensing and disaster management, and ing training and capacity-building initiatives to the con- institutional aspects of disaster management. tinent. It is envisaged as a Pan African initiative open The proposed programs leading to a graduate diploma to all with a focus on vulnerability reduction and and/or Master of Science in Disaster Management are learning to live with risk. The initiative was launched currently going through the University Academic Struc- with support from the World Bank's Disaster Manage- tures for approval. A blended learning approach, com- ment Facility and the ProVention Consortium at an Inter- bining in-class teaching, project work, and field and national Disaster Management Training Workshop action learning is adopted. It is anticipated that the conducted in parallel with the WSSD in August 2002 first students will register for degrees in January 2004. during the CEMSA 2002 International Conference on The programs are to be offered in partnership with the Environmental Management and Sustainable Develop- relevant faculties of the university using Wits School for ment in Africa. Participants from 10 countries (Botswana, the Environment as a conduit for cross-disciplinary stud- Egypt, Kenya, Nigeria, Netherlands, South Africa, ies. The graduate diploma in Disaster Management will Tanzania, Thailand, U.K, and Zimbabwe) deliberated for consist of six modules similar to those currently offered three days on the training and capacity-building needs as short courses but with more theoretical background in disaster management and have developed an action 208 Building Safer Cities: The Future of Disaster Risk plan that forms the basis for future DIMITRA activities. and to allow peer review. DIMITRA will also have "sup- Fund raising is currently in progress to support the porting members," strategic partners and donors who implementation of the DIMITRA Action Plan. will sit on the DIMITRA Steering Committee and guide DIMITRA is a network of university departments and its activities. research and training institutes specializing in disaster management and vulnerability reduction. Member insti- Conclusions tutions will be asked to endorse Vision for Risk-Free Environments, currently under development and aimed Disaster risk management needs to be led and based at regional integration and economic benefits for pres- within governmental authority, but its success cannot ent and future generations in Africa. be accomplished without the benefits of the widespread The mission of DIMITRA is to enhance regional capac- participation of many others. While policy direction is ity in disaster management and vulnerability reduc- crucial and legal foundations assure a continuing legit- tion through training, education, research, and outreach imacy, it is the professional and human resources deliv- by sharing the complementary expertise of its mem- ered on the ground that are the measure of success. In bers. DIMITRA member institutions will share expert- Africa the focus must thus be on the development of ise in various aspects of disaster management and human resources, which are so lacking; in addition, vulnerability reduction. The ultimate objective is to focused efforts on an international scale are urgently develop and implement a regional post-graduate pro- needed. gram in disaster management and vulnerability reduc- For this to happen, there must be coordinated effort tion aimed at educating a new generation of disaster and a systematic approach to establish sound founda- managers in Africa to address the complex and inte- tions and larger administrative and resource capabili- grated nature of disaster problems facing Africa. ties. The burden is too large for any individual country or government to handle--international efforts based on mobilizing locally available resources are required. DIMITRA Structure This is amenable to a broad network such as DIMI- TRAN, the one recently initiated, and actions are needed DIMITRA is envisaged as a membership organization. to make such a network effective. This will in turn require DIMITRA members will meet during the Annual Gen- funding and buy-in from policy makers. The Univer- eral Meeting. The first DIMITRA annual general meet- sity of the Witwatersrand considers supporting such ing will be held in the first quarter of 2003 to discuss initiatives to be its social responsibility and is taking the and adopt the constitution of DIMITRA. The AGM elects initial responsibility to make such initiatives a success. members to the steering committee that will oversee the But we cannot do it alone and support and contribu- DIMITRA secretariat. The University of Witwatersrand tions from others are needed and welcome. We look currently hosts the secretariat and is driving the initia- forward to a rewarding future. tive until funding is secured for DIMITRA to function on its own. Membership of DIMITRA is open to institutions in Bibliography Africathatareinvolvedintraining,education,andresearch Anderson, Mary B., and Peter J. Woodrow. 1989. Rising from the in fields directly related to disaster management and Ashes: Development Strategies in Times of Disaster. Boulder, vulnerability reduction, preferably at graduate and post- CO: Westview Press. graduate levels. The primary requirements of mem- Asian Disaster Preparedness Center, Bangkok. 1999. Managing bers are to subscribe to the principles of DIMITRA; to Disasters in Asia and the Pacific. A Review of Lessons Learned commit to further integrated management and preven- during the International Decade for Natural Disaster Reduction. tion, rather than relief, through sharing expertise and "Balancing the Environment and the Economy: Approaches for facilities with other members; to contribute to the devel- Mitigation." 2001. S01-13, Natural Hazards Center Research opment and maintenance of selected course modules; and Applications Workshops, session summaries, August. Living with Risk: Toward Effective Disaster Management Training in Africa 209 Blaikie, P., T. Cannon, I. Davis, and B. Wisner, B. 1996. At Risk: IDNDR (International Decade for Natural Disaster Reduction). Natural Hazards, People's Vulnerability, and Disasters. London 1999. Proceedings, IDNDR Programme Forum. and New York: Routledge. ISDR (International Strategy for Disaster Reduction). 2002. "Dis- Christoplos, Ian, J. Mitchell, and A. Liljelund. 2001. "Re-fram- aster Reduction and Sustainable Development." ISDR Back- ing Risk: The Changing Context of Disaster Mitigation and ground document for the World Summit on Sustainable Preparedness." Disasters 25 (3): 195. Development (WSSD) N°5. Disaster Management Bill. 2001. Pretoria, South Africa. Gov- SADC (Southern African Development Community). 2001a. ernment Printer. "Disaster Management Strategy." Secretariat, Gaborone. Gilbert, Roy and Alcira Kreimer. 1999. "Learning from the World SADC. 2001b. "Strategy for Flood and Drought Management in Bank's Experience of Natural Disaster Related Assistance." Dis- the SADC Region." SADC, Water Sector Coordinating Unit. aster Management Facility of the World Bank, Urban and Local Maseru. Government Working Paper Series N°2. Washington, D.C. UNEP (United Nations Environment Programme). 2002. "Global Green Paper on Disaster Management. Pretoria, South Africa. Environmental Outlook" (GEO-3 Report). A series of Government Printer. 1998. UNEP. Chapter 15 Urban Vulnerability to Disasters in Developing Countries: Managing Risks E. L. Quarantelli Some argue that the world is much riskier than it used Billions of dollars and countless hours of effort will be to be (Beck 1995, 1999). However, our age is not more wasted unnecessarily--all because we're afraid of the wrong dangerous, but the balance of risks and dangers has things.... In a frenzy of fear, we're pouring millions ... into protecting ourselves from the West Nile virus, and shifted. We live in a world where hazards created by spending only a fraction of that sum on public education ourselves are as threatening, or more so, than those encouraging people to wash their hands, which would that come from the outside (Giddens 2000:52). eliminate far more disease than killing every mosquito in "Risk" is always socially constructed by humans and America.... has no objective existence. "All perceptions of risk, In many areas, science can identify the physical haz- whether lay or expert, represent partial or selective views ards, tell us how many people are likely to be affected by each one, what various mitigations will cost and how effec- of the things and situations that threaten us" (Jasonoff tive we can expect them to be. We can rank risks and reme- 1998:91). Consistent with this is the observation that dies and put things in perspective. But we don't. Instead, "the empirical results suggest that the social processes we make policy based more on fear than fact. that construct and maintain risk in the public eye are This irrational response kills people ... the principal at least as important as, if not more important than, the underlying cause of wasteful choices is fear. But society physical and psychological dimensions of risk" (Rogers must be more rational than that .... Why not create ... an independent nongovernmental agency ... to provide 1997:745). credible guidance on risks? The institute would rank Perceptions of risk can differ dramatically. Indige- hazards so we could know which are most likely to nous residents of Bangladesh's Brahmaputra-Jamuna occur, classify risk according to their consequences, and River islands view the annual flooding in their region conduct cost-benefit studies to rank mitigation choices as part of a natural cycle that benefits the environment. by cost and effectiveness.... How do we make policy- The indigenous residents have developed a stock of local making more rational? (Ropeik 2000.) knowledge and strategies to cope with the environ- Policymakers should incorporate the views of ordi- mental conditions and take advantage of them for agri- nary people into disaster analysis and planning. Such culture, their main source of income. Western engineers, a strategy would further strengthen the partnerships on the other hand, define the flooding as an event that needed to plan for and respond to major disasters-- needs to be controlled to protect the 30,000 "helpless" consistent with the philosophy that one should plan inhabitants forced regularly from their homes (Schmuck- with people and not for them. Widmann 1998:1). This difference in perception has been observed by several researchers (Kraus and others 1992; Flynn and Urbanization others 1993; Tierney 1999). The differences between disaster "experts" and lay people often lead to a dis- Urban population growth. The world's population con- missal of lay people's perceptions. The director of risk tinues to increase, with 96 percent of this growth in communication for the Harvard Center for Risk Analy- developing countries (UNFPA 1991:3). The United sis summarized the views of experts in the following Nations projects that by 2010 there will be 511 metrop- manner: olises exceeding one million inhabitants and, for the 211 212 Building Safer Cities: The Future of Disaster Risk first time, more than 50 percent of the world's popula- a torrential rainstorm in the same area in 1952 did not tion will dwell in cities (Jones 1991:5). In addition, 40 result in a disaster since the area had been virtually unin- large cities will be added every five years, so that by habited and the park's slopes amply forested. 2025 there will be 639 metropolises with more than An unwillingness to seek safer housing is often dif- one million residents. Seventy-six percent of these will ficult or impossible, however; "people who are already be in developing countries. The metropolitan explosion barely eking out an existence will not avoid a risky flood will be greatest in Africa and Latin America. Currently, plain or the shadow of a volcano any more than they 77 percent of Latin Americans, 41 percent of Africans, will eschew the squatter settlements around a pesticide and 35 percent of Asians live in urban areas (UNFPA factory in Bhopal or a liquefied gas facility in Mexico 1991:10). City. In short, the poorest of the poor are probably likely Population growth alone does not account for all to reside in the path of both natural and technological urban growth, however. Urban migration also plays a hazards" (Bowonder and Kasperson 1988:104). large role: Industry concentration. Much of the population Declining mortality rates in rural areas of most develop- concentration is due to the concentration of industry. ing nations have not been matched with corresponding In developing countries, one or two cities often house fertility declines. The resulting increase of population the bulk of a country's industrial production. Mexico cannot be sustained by stagnating rural economics, which City, for example, generates one-third of Mexico's gross leads to growing demographic-employment opportunity national product; São Paulo, with 10 percent of Brazil's imbalances in the countryside. Migration becomes the population, contributes 25 percent of the net national only mechanism to relieve this imbalance. Rural migrants pour into the cities, exacerbating already overcrowded product; Bangkok, with 10 percent of Thailand's pop- conditions in urban subareas. The age selectivity of rural ulation, contributes 86 percent of the nation's gross migrants (largely teenagers and young adults) further con- national product in producer services and 74 percent tributes to city growth through new family formation in manufacturing (Kasarda and Crenshaw 1991:473). and natural increase (Dogan and Kasarda 1988:19). These concentrations of services and industries make Over the next 20 years, urban areas must absorb disasters in a city, a relatively small area with respect to one billion additional residents--as many urban resi- a country's total area, particularly catastrophic for a dents as there were in total in 1990 (Fuchs and others city and a country. 1994:1). There has also been a fundamental shift in In addition to the risk of losing a country's indus- the roles of cities. Once centers for collecting and pro- trial base during a single natural disaster, this concen- cessing resources and raw materials brought in from the tration of industry in and near population centers holds hinterlands, cities are now symbols of national identity additional risks. "Hazardous industries are not randomly and pride and the locus of a country's power. distributed within cities. Industrial zones in general, The urbanization process increases vulnerability to and hazardous industry sites in particular, tend to be natural disasters through the concentration of people located in less-affluent areas characterized by low socio- and assets. Increasing pressure to expand housing and economic residents less able to capably deal with, or commercial space has also accelerated the pace of vul- respond to, crises" (Britton 1991). Once again, it is the nerability. Housing complexes and industrial parks are poor who face additional risks to their life and health being rapidly constructed on unused land, formerly by the mere fact of their poverty. swamps or wetlands, in and near cities. Such land is Even when there are efforts to protect people, how- unstable for construction and property in these areas is ever, reality often intervenes. Rioters in New Delhi some of the first damaged during floods and earthquakes. torched buses and blocked roads in November 2000 Hillsides provide another unstable foundation. The to protest the closure of 90,000 small factories blamed 1999 rains and mudslides near Caracas, Venezuela killed for polluting residential areas. Rioters demanded that thousands. Many of the victims had cleared forest land the government ignore a Supreme Court order to close andbuilthomeswithintheboundariesoftheAvilaNational the factories that employed close to one million people Park overlooking the valley of Caracas. By contrast, (Dugger 2000). Urban Vulnerability to Disasters in Developing Countries: Managing Risks 213 Urban vulnerability--a reality. A decade ago, an (Alba 2000). This affiliation allows people to maintain analysis of the world's 100 most populous cities found ties with their tribes and rural areas of origin, provid- that 78 percent were exposed to one of four major nat- ing a sense of security to migrants (Gilbert and Gugler ural hazards--earthquakes, tsunamis, volcanoes, and 1992:157). A study in Nigeria found that "the evi- windstorms (not including flooding)--and 45 percent dence is overwhelming that virtually all respondents faced being struck by more than one. In developing maintain significant relationships with their rural home countries alone, 86 percent faced more than one threat community" (Gugler 1997:66). Though these migrants (Degg 1992:203­204). may be ethnically, religiously, and socially different from With their population densities and commercial the dominant group or groups in a city, understanding and industrial centers, cities have characteristics that the makeup of shantytowns and squatter settlements can magnify the impact of disasters to which they are aids in developing appropriate disaster-management prone. In developing countries in particular, many cities plans that incorporate all urban dwellers. have substandard construction, large slums, poor health Social dimensions of vulnerability. Women and care, inadequate water and sanitation services, crum- children are most likely to be victims of disasters by bling transportation networks, pollution, increasing the simple fact that they typically comprise at least 70 unemployment, and high crime rates. These challenges, percent of the population in developing countries. Until coupled with the fact that natural occurrences will recently, however, women went relatively unnoticed in continue to happen, put developing countries at a fur- disaster research and it was assumed that their behav- ther disadvantage. Even the 1995 earthquake in Kobe ior was similar to that of men. That is not necessarily overwhelmed the Japanese, who had some of the best the case. If caring for a household with its young and disaster-management policies and programs in the world. elderly occupants, women must concern themselves Growthofslumsandinformalsettlements.AUnited with saving the lives of others and themselves. A typical Nations report states that in 1990, roughly one billion "family" in some countries may include several gener- people lived in slums and squatter settlements. These ations, as well as kin normally considered "distant" in informal settlements increasingly make up large sec- the West. When the subject of women and disasters was tions of many cities: Dar es Salaam 60 percent, Lagos 58 finally studied, it was found that women have access to percent, Mumbai 57 percent, Mexico City 40 percent, fewer support systems following a disaster, are less likely Calcutta 40 percent, Nairobi 34 percent, and São Paolo to be involved in mitigation and prevention activities, 32 percent (Oberai 1993:2). When technological and are not always reached by warning systems, and have natural disasters strike these urban areas, the human not been recognized as necessary players in terms of and social costs are greater because a large percentage of disaster prevention, management, and assistance. the population is already less protected from the elements The same can be said for street children, an estimated and struggling for daily survival. Furthermore, in many 30 million worldwide (Laquian 1994:203). A study fol- cities, these informal settlements are considered illegal, lowing Hurricane Mitch found that street children fared therefore, government officials have not included them relatively worse than the general population and made in urban planning. Scarce resources are normally first up a disproportionate number of the victims. Those allocated to official neighborhoods and there is often little housed in temporary shelters following the hurricane remaining for squatter settlements. The squatters' lack suffered from severe physical, psychological, and psy- of political and municipal voting rights puts them at an chosocial problems (Reconstruction 1999:14). Further- additional disadvantage, making disaster planning for- more, in some cities, there may be a large number of midable in areas where existing social, economic, polit- individuals such as students or laborers living alone or ical, and technical constraints are severe (Davis 1987:6). rooming with nonrelatives. These unique social and Though these unofficial settlements may appear living patterns make up a city and must be adequately disorganized and chaotic to the casual observer, stud- taken into consideration in planning disaster response. ies have revealed that people often move into specific Disabled and especially vulnerable individuals. areas to live among a similar ethnic or religious group Countries in conflict or recovering from protracted 214 Building Safer Cities: The Future of Disaster Risk conflicts often have an unusually large population of throughdams,levees,andfillinginswampsandwetlands. disabled people who may need special assistance during Not only has flooding not been controlled, but in some times of crisis. Such communities may also be situated cases, the severity has increased and the ecological bal- near unmarked landmines or other hazards that make ance been upset. The Rhine River is one example: the delivery of assistance and travel to and from the The severity of flooding ... has been strongly influenced affected areas particularly dangerous. Large concentra- by many changes occurring to its floodplain since the tions of war widows with children may also need spe- beginning of the 19th century. Moors and bogs have cial attention during and following a disaster. been drained, cultivated and settled.... Since steamship The AIDS epidemic has also created new classes of navigation began in 1836, the meandering stream has been vulnerables. Those who are severely ill are especially straightened. Artificial cut-offs through meander bends vulnerable to disasters, as are concentrations of AIDS reduced the length of the Danube and caused it to incise its bed, thus increasing its fall and decreasing its width. orphans. In South Africa, some 20 percent of the pop- The construction of levees began in 1884, with the aim ulation is already HIV-positive, and it is estimated that of confining floodwaters to a narrower path.... These mod- 42 million African children will lose one or both par- ifications led to a loss of floodwater retention space, as ents to AIDS by the year 2010. In addition, the gross stagnant waters have been cut off and drained and the domestic product in many countries is expected to land filled. Rhine River forests have been eliminated, decline by as much as 20 percent as a result of the deaths wetlands drained and the infiltration capacity of both soils diminished by field clearance (Geipel 1993:112). of working-age adults (Nelson 2000; Wade 2000). In addition to the social and economic devastation, the Water is also channeled for drinking and industrial- disease means there are fewer adults and resources to ization. Chronic water shortages affect many countries, deal with other disasters that strike. particularly developing ones. "Freshwater resources are being used up at such rapid rates that groundwater sup- plies are dwindling and surface waters are fouled with Natural Disasters and Human Influence pollutants from industries, municipalities and agricul- ture. In much of Sub-Saharan Africa, the Middle East A. C. Bradley wrote nearly one hundred years ago, and parts of Asia, water consumption will reach 30­100 "[C]alamities ... do not simply happen, nor are they sent: percent of available reserves in 10­15 years--a result they proceed mainly from actions, and those the actions of population growth and inefficiencies in use" (UNFPA of men" (1906:11 cited in Hewitt 1997:ii). Floods, earth- 1991:5). quakes, and other "natural" disaster agents have social The water supply is also affected by droughts. While consequences only because of the activities of commu- droughts used to be thought of as a rural agricultural nities before, during, and after a disaster. High-density problem, this is no longer the case (Glantz and Mason population concentrations in floodplains, inadequate 1994; Vogt and Somma 2000). In Kenya, a lack of water and unenforced building codes, housing on volcanic to power hydroelectric dams resulted in the imposition slopes, and inadequate warning systems increase damage of strict power rationing in 2000 with most homes having from natural occurrences. If we think that the number electricity only three or four nights a week and factories of natural disasters has been increasing, that is not nec- receiving power every other day (Fisher 2000). Increas- essarily true. What seems to be happening is that the ingly, urban and metropolitan localities are finding them- same number of natural occurrences are resulting in selves faced with shortages or reduced water supplies. more costly remedies because humans have put them- In June 2000, the residents of São Paulo, Brazil, were told selves in nature's way. Property and economic loss, that their water would be rationed for five months, with psychological stress, and disruption of daily life are con- some three million residents required to follow the two cerns for most people rather than the actual hazards and one plan: two days with water, one day without themselves. (Wilmington News Journal, June 1, 2000:A15). Changingenvironment.Inanattempttomitigatewhat Soil erosion due to increased agricultural production we call a disaster, humans have tried to control nature also affects the environment and the impacts of disasters. Urban Vulnerability to Disasters in Developing Countries: Managing Risks 215 The use of heavy agricultural machinery compresses soils some hazards were not immediately recognized until and increases surface runoff. Crops like corn, sugar beets, after they had been in use for several years. Chemicals andturnips,whichareplantedinrows,increasethepoten- have become part of everyday life in cities, however, tial for soil erosion, especially during the spring floods, and are used to treat such essentials as drinking water. when once-scarce vegetation covers the fields. In recent Wastewater treatment also protects humans and the decades the area planted with these crops has increased. environment, though such processing can lead to the "The sectors where the levees were breached along the production of toxic sludge, further treatment of which Rhine River in 1988 contained hand-built dikes ... con- results in methane gas and carbon dioxide that con- structedtoreclaimlandforpastureandcultivation"(Geipel tribute to global warming. 1993:113). The increased use and production of chemicals has Air pollution from automobiles and industry has also resulted in an increase in accidents. Between 1917 and been blamed for increasing the severity of natural dis- 1979, there were 39 chemical disasters resulting in 20 asters. Twenty-five years ago, it was suggested that auto- or more fatalities. In just the next five years, however, mobile pollution exacerbated the incidence of tornadoes there were half as many disasters, including the pesti- in cities (Isaacs and others 1975). Although the research cide leak from the Union Carbide plant in Bhopal, India has not been universally accepted, recent studies sug- where 1,600 people died and 70,000 more were affected gest that air pollution may intensify the effects of hur- (Freitas and others 2000:21). And in February 2000, ricanes, also. It is thought that if the atmosphere's carbon the dike of a gold-extraction operation in northwest dioxide content doubles, hurricane intensity could rise Romania broke, spilling millions of gallons of cyanide- 40 to 50 percent generally and up to 60 percent in the polluted water into the Tisza River, which eventually Gulf of Mexico. The National Center for Atmospheric carried the pollution hundreds of miles into Hungary Research predicts that the atmosphere's carbon diox- and Yugoslavia (Savic 2000:2). ide content will double between 2035 and 2080, sug- Responding to technological disasters is often more gesting an accompanying increase in hurricane intensity. difficult than other disasters because the proper response Technological risks. In addition to humankind's to chemical poisoning and radiation contamination is ability to influence the impacts of natural disasters, we specialized and can strain emergency medical services have added threats entirely of our own creation. Chem- more than an "ordinary" disaster. This was the case in ical and nuclear risks are now present globally--in and Bhopal, India, where the local system was overwhelmed aroundfacilitiesandindistantlocalitieswhentransported with the numbers affected and the types of medical prob- by road, rail, and ship. Some 12 percent of all cargo in lems faced. The city's biggest hospital, Hamidia, admit- Russia is considered hazardous (Vorobiev 1998:33) as ted 1,900 seriously ill patients the first day and eventually is 5 to 15 percent of cargo on American roads. It has been treated more than 70,000 victims, though it had only estimated that more than a million residents of the 760 beds (Bowonder and others 1985:32). former Soviet Union live in contaminated areas in and Compounding the chemical risks has been an increase around 300 towns where chemical weapons were once in the sizes of trunks and tankers to transport chemi- produced, stored, tested, and destroyed (Shargorodsky cals, oil, and nuclear waste. The shipping tonnage of oil 1993). tankers has expanded sevenfold since 1960. Accidents Over the past century, the chemical industry grew involving these large tankers have resulted in signifi- exponentially and has gained political power. In the cant environmental damage as illustrated by the 1978 United States alone, the industry generates more than Amoco-Cadiz oil spill off the Brittany coast, the 1989 $200 billion per year through the manufacture of more Exxon Valdez oil spill near Alaska, and the 1982 Aegean than 4.5 million different chemicals--20,000 of them Sea tanker oil spill and fire at the harbor of La Coruña, produced annually in amounts exceeding one million Spain, a city of 250,000 people. Port cities face myriad pounds. Ensuring the safety of products that can explode, risks in addition to oil spills, however, since a large tanker burn, asphyxiate, poison, corrode, and generally wreak carrying liquefied natural gas carries the energy equiv- havoc on humans and the environment is not easy, and alent of 55 Hiroshima-type atomic bombs (Lovins and 216 Building Safer Cities: The Future of Disaster Risk Price 1981:64). Moreover, the West has begun export- weapons program, the Energy Department has con- ing its hazards by sea to developing countries, which cluded that 109 will never be clean enough to permit are desperate for the income generated. unrestricted use by the public" (Kempster 2000:A5). Nuclear hazards. The nuclear industry has a shorter Radiation fallout from the 1986 Chernobyl nuclear acci- history than the chemical industry, but its disasters can dent affected European countries that in turn exported be even more deadly. While most nuclear risks are asso- contaminated milk to Malaysia, Nepal, and Ghana ciated with cheaper and so-called "cleaner" electricity, (Adams 1998:206). In Belarus, just across the border, running a plant and disposing of spent nuclear fuel rods the incidence of thyroid cancer was found to be 24 times create significant hazards. Twenty-three nuclear inci- higher in contaminated areas than other parts of the dents were reported in the former Soviet Union between country in 1999 (ICRC 2000:94). Western Europeans 1951 and 1991, though undoubtedly there were others have expressed concern for the safety of their own nuclear that went unreported (Gentleman 2000:134). Incidents plants and the deteriorating facilities in Eastern Europe, not immediately reported nor appropriately responded particularly the six Kozlodui nuclear plants in Bulgaria to include the 1957 explosion of a tank that spewed (U.S. GAO 2000). While natural disasters may destroy 70 to 89 tons of radioactive waste in and around a buildings and alter the physical environment, techno- nuclear weapons plant near Chelyabinsk. At least 270,000 logical disasters can contaminate and render areas unfit people were exposed to the radioactive cloud, although for inhabitation. the resulting health impacts were difficult to ascertain Othertechnologies.Computertechnologyisnotinher- precisely. Twenty-three villages had to be razed and ently hazardous, although the use, or rather misuse, of 10,000 residents permanently resettled, although that technology can be. Hackers can create disasters and major was not undertaken until a year later. Seventeen thou- inconveniences by disrupting computer-dependent sys- sand acres of polluted farmland were turned, ironi- tems such as power and water supplies, and nuclear cally, into a nature reserve (Monroe 1992:535­536). facilities. They can also create viruses that interrupt This is considered the most contaminated spot on the normal systems functioning. Though many technolo- planet. gies are recent, people have become so dependent upon Even in developed countries with strict controls, them that when they fail, so do other systems. For exam- nuclear plants are not entirely risk-free. In Japan, the ple, in a chemical disaster where the computer moni- Mihama nuclear power plant released radioactivity after toring system failed, the surrounding population received a steam pipe ruptured, forcing the emergency system the warning late since there was no backup system. In to flood the reactor to prevent a meltdown (Sanger 1991). precomputer days, the warning would have been issued In the United States, the near meltdown at the Three verbally hours earlier (Quarantelli and others 1983). Mile Island facility in 1979 proved that such facilities Telecommunications infrastructure is also vulnera- were not as risk-free as claimed. ble. In 1991, 11 widespread phone system outages affect- The global arms race also ensures the proliferation ing major metropolitan areas occurred in the United of nuclear hazards. North Korea, Iran, Libya, Iraq, Israel, States. In the report accompanying those figures, it India, and Pakistan face global scrutiny for their capa- was noted that "modern fiber optics carry 10,000 bilities and activities. The former Soviet Union is attempt- times more calls than the old copper cables they replaced. ing to decommission some of its nuclear weapons, but An accidental cut of a single fiber optics cable can cut lacking adequate funding, the United States has stepped off entire metropolitan areas" (Lee 1992:8). in at some facilities to ensure weapons are properly Technology also allows us to build bigger and taller destroyed and components are not sold on the black skyscrapers, but this represents another increase in vul- market. nerability, as was tragically evident with the World Trade Cleaning up nuclear facilities is difficult if not impos- Center towers where some 2,800 people died. These sible, and lingering contamination spreads easily. A 2000 were not the first tall buildings to burn, however. Dev- report by the U.S. National Research Council states, astating fires in high-rise buildings have also occurred "Of 144 facilities that played a role in the U.S. nuclear elsewhere. In 1974, 189 died in a 25-story building in Urban Vulnerability to Disasters in Developing Countries: Managing Risks 217 São Paolo, Brazil. Such fires have prompted discussions plutonium and strontium from an earlier nuclear dis- of protective measures such as wider stairways to use aster was blown to other areas (Porfiriev 1991). Cog- in the event of an emergency, construction of heavily nizant of the chain of events, disaster risk management fire-proofed refugee floors where evacuees could gather, is increasingly taking into account the possibility of haz- and banning underground parking garages where ter- ardous chemical releases and spills after earthquake rorists could strike. shocks (Tierney 1990). Some technologies are more scientific than infra- structure-related, making planning by the average person in the event of an emergency more difficult since the Disaster Planning, Mitigation, Response, threat is often not understood. In 1979, biological toxins and Recovery were accidentally released at a Soviet research center. As many as 1,000 workers were killed, and a 20-square- Disaster planning.Thedegreeandtypeofdisasterplan- mile area around the city of Svardlovsk was contami- ning in place in developing countries varies considerably. nated by highly toxic anthrax spores (Thompson In India and Iran planning is significant, while in coun- 1987:11A; Oberg 1988). Like nuclear and chemical tries such as Venezuela and Taiwan there are no national- facilities, manufacturing facilities for biotechnological level disaster plans. In Turkey, until recently, there was purposes also have significant risks associated with the no single national coordinating agency for disaster man- production, storage, transportation, distribution, and use agement, although there was a blueprint for action and of the products involved. (For biotechnological hazards assignment of responsibilities (Gulkan and Ergunay in agriculture in developing countries, see Dommelen 1999:176). The existence of government agencies, how- 1999; the spread of biotechnology in developing ever, says nothing about the quality of their operations countries is discussed in Acharya 1999.) ineitherplanningormanaging.Theremustalsobepublic Genetic engineering may also hold risks not yet fully interest in addressing risks. Seoul was flooded in 1966, understood. For example, oil and chemical waste-eating 1972, 1984, 1987, and 1990 (Kim 1999:92) though little bacteria have been created that can be used to help clean action has been taken to address the recurrent flooding. up oil spills, although these organisms could also attack There is also considerable variation in planning for lubricants on machinery. As a National Science Foun- technological disasters. A survey of mitigation and other dation report has stated, however, "no hazard particu- preparations for disasters in developing countries in lar to genetic engineering has yet surfaced" (Schmeck Asia and the Pacific found that while there was at least 1987:7). minimal planning for natural disasters, there were almost Super disasters. With an increase in technological none for technological disasters (ADB 1991). A survey hazards and the transportation of chemicals and haz- of national disaster plans of developing countries sug- ardous waste, it logically follows that there will be an gests that governments give less attention to techno- increase in such disasters. Magnifying them could be logical disasters than natural disasters. Mexico City, for the effects of natural disasters (Showalter and Myers instance, paid little attention to planning for techno- 1994). In 1987, a powerful tornado created 14 toxic logical disasters until after the 1985 earthquake. chemical sites around Edmonton, Canada. California's It is generally accepted that the best disaster plan- Northridge earthquake caused: one train to derail and ning and management involves the following (Sorensen spill sulfuric acid and diesel fuel; nine pipeline ruptures; 1988:241­260; Quarantelli 1991): 35 breaks in natural gas transmission lines; 717 breaks · Views disasters as different from accidents and minor in distribution lines; and 15,201 natural gas leaks that emergencies; resulted in three street fires, 51 structural fires, and the · Views catastrophes as different from disasters; destruction of 172 mobile homes (Lindell and Perry · Focuses on multiple hazards and is generic rather 1995:8,11; Lindell and Perry 1998:287). In Russia in than agent-specific; 1961, windstorms in the Lake Karachay region in the · Includes all four time phases of the planning process: Southern Urals spread contamination when radioactive mitigation, preparedness, response, and recovery; 218 Building Safer Cities: The Future of Disaster Risk · Aims at multiple rather than single-hazard or risk- implemented mitigation measures can in part explain reduction goals; the great difference in the number of casualties between · Focuses on general principles rather than specific the earthquake in Armenia (10,000 to 25,000 deaths) details; and Northridge, California (61 deaths). Appropriate · Highlights a continuing process rather than an end construction materials and the enforcement of housing product, such as the production of a written plan or codes in the United States resulted in few casualties a document for mitigation; during the earthquake (Poghosyan 2000), while damage · Builds on research findings derived from systematic was heaviest in recent construction in Armenia and data in addition to personal experience; Mexico City as buildings collapsed (Krimgold 1992:217). · Emphasizes the need for coordination both within An initial requirement for mitigation planning is com- and between organizations and segments of the com- munity awareness of probable disasters in their urban munity, rather than "command and control"; and locality. This may seem obvious, but for many commu- · Distinguishes between planning and managing, nities a major disaster is a one-time occurrence in the between the strategies and the tactics necessary. course of a generation. Studies in the United States indi- Disaster planning includes developing a set of activ- cate that the majority of citizens have never directly expe- ities and systems to prepare for and predict disasters. rienced a major disaster. This is not true of disaster-prone Forecasting and warning systems, community educa- cities such as Buenos Aires, which was flooded 37 times tion, emergency operations centers, and medical and in the past 15 years (Herzer and Clichevsky 2000:34), food stockpiles are part of the preparation. This can be nor is it the case for Bangladesh and some Pacific difficult in urban areas that are struggling to develop island nations. Without personal experience, however, and where basic infrastructure and municipal services people generally do not believe they are at risk from a are lacking. Funding infrastructure and activities in these disaster. An ongoing Disaster Research Center (DRC) areas for events that may never occur may seem like a evaluation study in the United States called "Project waste of precious financial resources. Urban commu- Impact" reports that, "Repeatedly, DRC researchers have nities should put disaster preparedness planning on heard from [Project Impact] community leaders ... that their community agendas, however, and link it to it would be necessary to have a disaster hit their com- overall developmental planning since such investments munities in order to focus attention on the need to sup- need to be protected. This is especially important in port mitigation activities" (Disaster Resistant 2000:25). developing countries where disaster management agen- Even after a disaster, people can dismiss the event as a cies and awareness at the local level are rare. one-time occurrence. Thus, disaster planning, especially Mitigation. The concept of "mitigation" in the event for mitigatory measures, must take into account the expe- of a disaster has been around since ancient Egypt and rience of disasters in an area over time--perhaps over China, perhaps since 4000 B.C. (Qingzhou 1989: two generations. 193­194; Waterbury 1979). More recently in the United While some communities are unaware of their vul- States, mitigation has involved structural mitigation nerability, others are very aware, but take little action such as dams and levees to reduce flooding. because the pattern of disasters has not been studied "Mitigation" tends to be used in two major but dif- and no reliable predictions are made. Dhaka, Bangladesh, fering ways in research and professional disaster liter- has been the site of numerous floods that have inun- ature. In the United States and several other countries, dated much of the city. Two-thirds of all slum dwellings it generally refers to "away from and prior to impact were flooded in 1987 and the next year 77 percent of period" measures and includes activities to eliminate the city was flooded (Huq 1999:130). A recent analy- and lessen risks. Examples include improved building sis offered the following observation: "Historical data codes, consciousness-raising about risks and disasters, on hazards and disasters in Dhaka are almost totally targeted educational programs, relevant zoning and land- lacking. This makes it difficult to project likely pat- use rules, insurance purchase (Britton 1991), and com- terns of risks and potential losses in the expanding city plianceand enforcement of safety regulations. Successfully of the future" (Huq 1999:125). This lack of data and Urban Vulnerability to Disasters in Developing Countries: Managing Risks 219 inadequate preparedness of citizens has led residents are properly designed and run and that monitoring to respond in the only way they thought they could save systems are in place and used. A regulatory system mon- themselves--prayer. In a survey of two coastal com- itored by community authorities can increase social munities in Bangladesh struck by a cyclone, it was found pressure for proper maintenance and procedures. In the that "praying to Allah" was undertaken by 73 percent United States, however, some have contended that this of residents in one village and 90 percent in another. had led to over-regulation, particularly in the nuclear In both localities, it was the most frequent precaution- industry. This may seem like a strong counter-argument, ary measure taken (Haque and Blair 1992:225). but the result of "over-regulation" has been a safety record Awareness of potential disasters is the first step, wheretherehasneverbeenanuclearaccidentintheUnited although knowledge about possible disasters must be States resulting in casualties outside the facilities. accompanied by a willingness and an ability to imple- Restrictive land-use measures and zoning codes ment mitigation measures, since awareness does not also reduce risk, but there must be socioeconomic and automatically translate into action. In a recent survey sociopolitical support for such measures. In Brazil, vested of disaster planning in Tehran, Iran, it was noted that interests in rapid economic growth result in weaknesses the area has a history of earthquakes of magnitude 7.0 in institutional strategies for the reduction of accidents. and higher. Though a comprehensive plan for disasters Riskanalysis,land-useplanninginthesitingofhazardous has been developed by officials, "The steps that people facilities, mandatory accident notification, disaster plan- can take to protect themselves are not known and the ning, and information dissemination about risks and demand of the community to have themselves protected response measures for crisis situations are frequently is not forthcoming" (Nateghi-A. 2001:210). The same ignored (Porto and Freitas 1996:20). Even in developed seems to be true for the threat of floods in Seoul, Korea. countries, it is not the absence of knowledge about flood- City officials are aware of potential disasters, but plains that causes problems, but an unwillingness to neither flood-risk zoning nor land-use control is prac- use this knowledge in community planning. ticed, partly because of citizen "indifference" and also A local risk assessment provides the starting point because in a densely populated, fast-growing city all for understanding the most immediate threats and prepar- available land is used and low-lying areas filled in (Kim ing appropriately. The risks and preparedness measures 1999:112). identified must be communicated to the local popula- Though a community may want to minimize its expo- tion. Risk awareness, however, does not translate auto- sure to potential disasters, hazards are sometimes thrust matically into preparedness. Especially for chemical and upon them. Multinational corporations, often the nuclear risks, specific measures must be taken to pro- investors in chemical and nuclear facilities, generally tect the population and contain the threat. For exam- deal with a central government. As a result, local author- ple, there have been instances of fire departments ities frequently have little say on the siting and inspec- provoking explosions and fires by spraying water on tion of facilities, particularly in developing countries otherwise inert chemicals. In Bhopal, medical person- where community input is ignored (Towfighi 1991:107). nel, the government, and local communities should Multinationals have the added complication of public have been briefed that methyl isocyanate (MIC), a relations and avoiding bad publicity, which could hinder toxic substance, is inactivated when people breath disaster response. In Bhopal, for example, civil author- through a moist towel. This would have saved a number ities called the factory, but the staff did not confirm the of lives (Bowonder 1985:96). chemical leak, thus delaying evacuation and causing Disaster response. When a disaster can be predicted confusion among civil authorities and police trying to or is spreading, evacuation is often the official recom- respond (Bowonder 1985:250) to the mounting disaster. mendation, though it is not always followed (Perry 1985; At the community level, there should be insistence Vogt and Sorensen 1987). People are often reluctant to upon appropriate mitigatory measures, the use of safety leave their homes and farms, which might represent technology, and an effective regulatory system. Technol- their life savings. In a 1991 cyclone in Bangladesh, half ogy can increase the probability that key installations of the households in two coastal villages, all of whom 220 Building Safer Cities: The Future of Disaster Risk received early warnings of the threat, did not leave to hardest hit since protective measures may be prohibi- seek shelter (Haque and Blair 1992:217). One hundred tively expensive. Corporations tend to have more meas- and thirty-nine thousand people died in the storm. ures in place and multinationals are often best equipped Chemical and nuclear disasters also result in large-scale to pay for and carry out mitigation strategies. evacuations even when fatalities are expected to be min- Some businesses recover in the long run, although imal. For example, a 1985 toxic sulfur trioxide release those that were marginal in the pre-disaster period often in New Delhi, India forced 100,000 to evacuate and a fail or do not reopen in the aftermath of a major disas- 1989 crude oil explosion resulted in 200,000 evacuees ter. Surprisingly, prior preparedness planning seems to from Guadalupe, Mexico, though neither case resulted have little effect on recovering from a disaster, although in a significant number of deaths (Cutter 1991:280). this might be due to poor preparedness planning that Once a disaster strikes, the overwhelming majority did not take into account the special needs of businesses of search and rescue activities are carried out by friends, from the start. There is also no clear relationship between relatives, and neighbors. It is impossible to predict disaster recovery and disaster assistance provided by howpeoplewillrespond,butknowingofrisksandprepar- government agencies. In fact, "there was no relation- ing communities to face them can improve response. ship whatsoever between the number of aid sources Though some activities such as rescuing people from businesses relied on during the post-disaster period and collapsed buildings may require heavy equipment, urban the extent to which they recovered" (Webb and others officials in earthquake and other disaster zones should 1999:16). ensure that equipment and trained crews are available. A review of disaster plans from a number of urban Asian cities, however, found that, with the exception When Disaster Strikes of Singapore, few cities addressed the idea of collapsed buildings in preparedness planning. Even though some communities have put disaster man- In many countries, the military is often the key respon- agement plans in place, difficulties in implementing dent to a disaster (Walker 1992). Ironically, however, them can be illustrated by the following observations military organizations have not been systematically made in Ecuador, Peru, and Bolivia during 1997 and approached for an analytical perspective on disaster 1998: response. Neither have religious organizations. Relief provided by churches and organized religious groups While the civil defense organizations in the respective countries were the nominal "national emergency organi- frequently ranks very high as a major source of aid. At zations" ... each was rapidly pushed to the sidelines by the international level, appeals are often made to smaller one or more new but temporary governmental organiza- congregations, who donate generously. As urban offi- tion charged with supposedly managing the response. The cials prepare disaster plans, they should encourage the result was 1) confusion and duplication at the institutional participation of these experienced groups in their level, and 2) a serious loss of credibility and morale in planning. each country's civil defense structure. This is hardly the combination one would seek for optimizing institutional Disasters result in loss of life, homes, livelihoods, and readiness (Olson and others 2000:5). economic activity. Industrial areas, business districts, crops, and livestock are frequently ravaged during dis- The same scenario occurred after Hurricane Mitch, asters. Businesses and economic entities often face not though local agencies knowledgeable about an area only the loss of a place of business, inventory, and income should assume a lead role, especially if they have pre- during the days of the disasters, there are numerous lin- pared for such an event. The strengthening of their gering impacts. Suppliers may be delayed in delivering capacities for all levels of a crisis means that fewer products, people may have lost the means to patronize crises will become emergencies, fewer emergencies a business, and employees and customers may be suf- will become disasters, and fewer disasters will become fering from a lingering trauma that hampers a busi- catastrophes. If these agencies are adequately pre- ness' return to normality. Small businesses are typically pared, they will be able to respond appropriately and Urban Vulnerability to Disasters in Developing Countries: Managing Risks 221 call in reinforcements and additional expertise as nec- Center. In contrast to disasters, everyday emergen- essary. Civil defense organizations should also accept cies do not bring about such a massive convergence that disasters and catastrophes are likely to become polit- of groups. ical and plan for that eventuality (Olson and others · Organizations must adjust to a loss of autonomy 2000:36). This also applies in the disaster mitigation and freedom of action. Minor emergencies are often process for, as has been said, "The decision-making managed by public sector organizations such as fire process about the acceptability of risks is essentially a and police departments, private organizations, and political process" (Porto and Freitas 1996:24). the private sector. Individuals may also be required With local organizations at the helm, the resumption to surrender individual and property rights to "the of essential services, with which they are most familiar, greater good" as search and rescue is carried out and can more easily resume and be balanced with the pro- access to some areas is restricted. vision of emergency services. A general disaster plan- · Performance standards and norms for organizations ning and management principle is that organizations often change drastically during disasters. What is should stick to their mandates and attempt to return appropriate during periods of normalcy or minor an area to normal life as soon as possible. However, in emergencies typically becomes less relevant during all major disasters, new behaviors and groupings will the management of a major crisis. There is often a emerge as people and organizations cope with the deliberate slowdown in organizational activities multiple contingencies created that cannot be handled that does not occur in everyday emergencies. by routine service delivery. In fact, the greater the dis- · A catastrophe can have even graver consequences aster, the more improvisations will appear, accompa- than a disaster, since most or all of a community is nied by pluralistic decision-making in tasks ranging heavily impacted. In a DRC field study, it was found from evacuation and the provision of emergency med- that Hurricane Hugo destroyed or heavily damaged ical services to inter-organizational coordination and more than 90 percent of all homes in St. Croix, U.S. community priority-setting (Quarantelli 1996). Virgin Islands. The near total destruction made it Events that disrupt communities are not all of the impossible for displaced victims to seek shelter with same magnitude, since a disaster is more than an every- relatives and friends. By contrast, the 1985 Mexico day emergency. In a disaster, the core emergency and City earthquake destroyed less than 2 percent of response facilities can be destroyed. In recent catastro- residential housing stock and only 4.9 percent of phes in developing countries, small towns have had their the population reported significant damage to their medical and police personnel wiped out. Outside agen- homes. Of those displaced, most found accommo- cies later responded, though their response times were dation with relatives. longer, they were unfamiliar with the area, and there · Catastrophes have longer recovery periods than dis- were "turf wars" as agencies staked out their territories. asters. Massive destruction of buildings may make it The following describe some of the differences in impossible for some to return, while others might organizational response that can be expected in disas- have lost their places of work. This was true in the ters and catastrophes as opposed to regular emergen- 1988 Armenian earthquake. For years afterward, cies (noted in Quarantelli 1998a:5­8): there were few inhabitable buildings. The population · Organizations must relate quickly to a number of of Leninakan dropped from 230,000 to 120,000; familiar and unfamiliar groups. For example, a research Kirovakan from 170,000 to 74,000; and Spitak from team studying a massive fire near Nanticoke, Canada, 18,000 to 3,000 (Poghosyan 2000). While it is dif- identified 346 organizations inside the evacuation ficult to prepare for such a massive quake, there are perimeter of the fire--all of whom had to pass through useful lessons to be found in the recovery since the a police checkpoint (Scanlon 1992:9). Even for a areas are thriving once again. plane crash in Detroit, the DRC found 241 organiza- Once a disaster has occurred or the threat seems tions responding, including 59 fire departments and to have passed, the end of the evacuation period is 69 agencies represented at the Emergency Operations often announced. This can be more controversial than 222 Building Safer Cities: The Future of Disaster Risk evacuation, however (Stallings 1991:193). Estimating In the longer term, disasters have been found to have the remaining danger and ensuring a measure of safety impacts upon the political process. A study of 12 coun- are often impossible. Some people may feel it is safe to triesstruckbyrapid-onsetnaturaldisastersbetween1966 return home while officials may disagree, not wanting and 1980 found a positive relationship between disaster to incur expensive re-evacuation costs if the area proves severityandsubsequentpoliticalunrest(OlsonandDrury unsafe. In other cases, a government may have run out 1997; Drury and Olson 1998). Studies of Africa and Latin of funds to assist evacuees and wants people to return America found a relationship between regime type, mis- home--even though, in the case of nuclear or chemi- management, and levels of disaster damage (Davis and cal hazards, the area may not be safe from invisible Seitz 1982; Seitz and Davis 1984). In a recent study, threats and the effects might not be known for years Shefner (1999) looked at whether the sewer explosions (Stallings 1991:195). inGuadalajara,Mexicohadanypoliticalfallout.Although Recovery management. Recovery deals with activ- he found that the political activists who emerged after ities undertaken after a crisis response period is over in the disaster had nearly disappeared two years later, an attempt to return an area to normal. In the short there were longer-term effects in later elections. term, the focus is on restoring infrastructure and serv- Political issues also arise out of uncertainties. Victims ice delivery. In the long term, the attempt is to restore of technological disasters often find themselves with community life. Specific measures can include provid- health concerns for which there is no ready remedy. ing financial and technical assistance to farmers want- Their daily existence becomes a stressful guessing game ing to clean up contaminated farmland, setting up of of identifying symptoms and disease, and the uncertainty counseling services for victims, restoring urban serv- increases the duration of the disaster. This post­traumatic ices, and rebuilding damaged and destroyed facilities stress affects an unknown number of victims of tech- (see Mitchell 1996). There may also be calls for such nological disasters, although just 3.8 percent of victims an event never to be allowed to happen again. of natural disasters appear to be affected with this psy- In the aftermath of a disaster, citizens may be out- chological trauma (Brody 2000:D8). raged that certain "avoidable" incidents took place and Sometimes psychological stress is the "disaster." For local officials did not sufficiently protect or warn them. instance, there were no casualties and little property People often organize following a disaster and political damage in Mississauga, a suburb of Toronto, when activists take advantage of such an opportunity to politi- 217,000 residents had to be evacuated following a cize people around this and other issues. Where recur- train derailment that threatened the release of toxic rent threats are possible, citizens groups are even more chemicals (Scanlon and Padgham 1980). Though no likely to emerge (Quarantelli 1988). Activist groups one was killed, the threat and evacuation disrupted the appeared in Mexico City after the 1985 earthquake lives and routines of nearly a quarter of a million (Dynes and others 1990), Bhopal after the chemical poi- people for several days and caused psychological, social, soning, and Turkey after the 1999 earthquake. Groups and indirect economic impacts. Similarly, the Three Mile may also seek out those who should or are thought to Island nuclear threat "provided a dramatic demonstra- be responsible for augmenting the disaster. The dis- tion that factors besides injury, death, and property covery of building codes unenforced, warning messages damage impose serious costs." Although it caused not not passed along, and delays in disaster response can a single death and few if any latent cancer fatalities, it provoke the wrath of disaster victims. devastated the utility that owned and operated the plant Once politicized, citizens may insist upon the pas- and imposed $500 billion in costs and sanctions on sage of laws and regulations to prevent similar disasters the nuclear industry and society. Following the inci- in the future. Activist groups and NGOs may also try to dent, stricter regulations, reduced operation of reac- prepare the local population for future threats. Taking tors worldwide, greater public opposition to nuclear advantage of public interest and partnership opportu- power, increased reliance on more expensive energy nities, governments can channel resources of these groups sources, and increased costs of reactor construction and to be better prepared in the event of other emergencies. operation resulted (Slovic 1987). Urban Vulnerability to Disasters in Developing Countries: Managing Risks 223 An even more unusual example occurred in Goiania, of business and expecting them to suddenly work together Brazil, where a cancer treatment machine abandoned amicably is unrealistic. If these groups are involved in a junkyard released cesium 137. The radioactivity during the planning phases, however, and get to under- killed four people and seriously affected 44 others. But stand each other's agendas, the recovery period will run far more consequential was the perceived risk to the smoother. town's 1.2 million residents. Over 100,000 residents Additionally, to improve future disaster risk plan- underwent Geiger counter examinations to detect pos- ning, recording a disaster and response to it can pro- sible contamination, and some 8,000 formal certificates vide valuable lessons. In Bhopal, authorities and medical were issued stating that the person was not a carrier of personnel were never told of the threat that lurked in hazardous radiation. Anxiety over contamination also the Union Carbide plant. When chemicals leaked, led hotels in the country to cancel the reservations of hospitals gave only symptomatic treatment for the people from Goiania, buses and airplanes to refuse to unknown poisoning (Bowonder 1985:86). In both the take Goiania residents as passengers, and some doc- Three Mile Island and the Chernobyl nuclear accidents, tors and dentists refused new patients who did not have iodine pills that could have reduced some of the effects the formal certificates. Conventions in Goiania were also of radiation had neither been stored in sufficient quan- cancelled. One estimate was that regional tourism fell tities nor adequately distributed during the crisis. These 40 percent and it was reported that property values provide clear lessons for authorities working in areas fell, with sales levels for the entire city and state being where there are chemical or nuclear plants. affected. Some 50 percent of the state's export sales were lost for one month because the area's agricultural prod- ucts were boycotted (or purchased at 50 percent of Economic Impacts value). Even textiles and clothing manufactured in Goia- nia were affected, with some items losing nearly 40 Economic loss is often the determinant of whether an percent of their value (Petterson 1987: 3­4, 8­9, 12). event is described as a disaster or a catastrophe. Look- Recovery is often a complex economic and social ing at total dollar amount alone is not enough, how- process, and it becomes protracted when a disaster has ever since "the economic costs of disasters in poor contaminated and polluted buildings, their contents, countries often exceed 3 percent to 4 percent of the and the surrounding air, land, and water. While this is gross national product.... In ... economically vulnera- more likely in technological disasters, it can also happen ble East African countries ... the costs exceeded over with volcanic eruptions like Mount St. Helens and floods 20 percent of GNP at various times during the 1980s. such as the one that struck eastern India in August 2000, In contrast, the $24 billion loss from the 1992 Hurri- where the flood waters contaminated vast supplies of cane Andrew disaster in South Florida, which was at drinking water (Kim 2000). To the extent possible, con- the time the costliest disaster in the history of the United tamination needs to be contained, but having the appro- States, represented an almost undetectable proportion priate quick response is not often possible. of the country's $6 trillion economy" (Berke 1995:372). Agencies and groups involved in recovery also tend Others have noted that the Managua earthquake of 1972 to be different from those that participate in the pre- resulted in $5 billion in damage, nearly 40 percent of paredness and response phases of disasters. Police and Nicaragua's GNP, while the $8 billion in damage from fire personnel, emergency medical teams, and emer- the Loma Prieta earthquake represented only 0.2 percent gency and crisis managers are typically the major play- of the GNP of the United States (Hohn 1995:573). What ers during the preparedness and response phases. City is "disastrous" in a developed country might be "cata- officials such as community planners, and building and strophic" in a developing one. housing inspectors, as well as members of the private Disasters in developing countries often affect large sector such as insurance agents and bankers, are more numbers of people, also. It will forever remain unknown active during the recovery period. These two groups whether Hurricane Mitch directly affected 24.2 percent do not always communicate during the normal course of the population in Honduras or 19.5 percent in 224 Building Safer Cities: The Future of Disaster Risk Nicaragua (Reconstruction 1999:13), but clearly a sig- in terms of such losses are relatively rare. For example, nificant percentage of the population was impacted in earthquakes are infrequent. "Over 70 percent of the one way or another. The recent floods in Mozambique approximately 1.3 million earthquake related deaths affected 27 percent of the population (Swarns 2000:9). since 1900 have occurred in 12 single events.... In the By contrast, the U.S. Federal Emergency Management United States, only an estimated 1,600 deaths have been Agency (FEMA) reported in 1998 that from 1993 to attributed to earthquake since colonial times" (Jones 1998, at least 1,400,000 Americans were impacted in and others 1993:19, 20). federally declared disasters and that at least hundreds of thousands of other people were impacted by events managed entirely at state or local levels. This indicates Improved Disaster Risk Management that, over a six-year period, only half of one percent of the American population was affected. Little distinction should be made between prepared- Though disasters in developing countries affect greater ness for natural and technological disasters. Researchers numbers of people and can destroy a greater percent- summarizing the discussions at a recent international age of the domestic economy, the impacts of disasters are conference held in Japan on natural-disaster reduction no longer restricted to the areas they strike. Globaliza- noted that the core issues in risk management for nat- tion now increases vulnerability for everyone. When a ural disasters are similar for all disasters and the involve- 1999 earthquake destroyed a number of Turkish weav- ment of people in developing strategies for disaster ing factories, cotton production in Sub-Saharan Africa mitigation is crucial (Herath and Katayama 1994:1). was affected. The market for cotton was drastically Current disaster theorists argue that all disasters can reduced and unemployment rose in several African be attributed primarily to human actions and that no countries. Similarly, when an earthquake hit Taiwan in distinction between "natural" and "technological" dis- 1999, the sales of computers in the United States were asters is significant for general disaster-risk-management impacted, since critical computer components were purposes. Recently, several authors from six different manufactured in Taiwan. Also, following an explosion countries and social science disciplines agreed that at the Nisshin chemical plant in Japan, there were dif- the distinction has been increasingly abandoned by ficulties for the global semiconductor industry, since most scholars in emergency management operations the company provided the world's main source of hydrox- worldwide (Quarantelli 1998b:248). However, others-- ylamine, a chemical vital to photoresist stripping of Kroll-Smith and Couch (1991), Baum and Fleming semiconductors (Kallender 2000). (1993), Freudenburg (1997), Picou and Gill (1999), and Tourism, often a significant source of revenue and Picou (2000)--write that technological disasters are foreign exchange in developing countries, is commonly unique since they result in long-term social impacts not another victim of disasters. In Indonesia, fires created seen in natural disasters. such thick smoke pollution in 1997 that Brunei, the Yet another objection to making a natural/techno- Philippines, and other Southeast Asian countries were logical distinction is that there are disasters where no affected (Khandekar and others 2000) and the Thomas agent is clearly identifiable. Famines are the best exam- Cook travel agency refused to book vacations in affected ple of this. Also, for newer types of disasters such as com- areas, since poor visibility due to the smog curtailed, puter system breakdowns, the source of the problem delayed, and canceled flights (Tourism 1997). A number can be natural, technical, or social. That is, computer of nations lost their share that year of Southeast Asia's systems can be brought down by earthquakes, power annual $26 billion tourism industry (Mydans 1997:3). outages, or hackers. The stock market crash of October Due to complex social links in the modern world, 1987 was a disaster whose occurrence did not directly future disasters could have catastrophic potential even if result in any fatalities, but with a one-third decline in they do not result in casualties or physical impact. There the Dow Jones Industrial Average, $1 trillion vanished is a tendency to equate disasters with casualties and in four days (Barro and others 1989:127). Resulting property damage. However, events that are catastrophic economic losses around the world brought the total Urban Vulnerability to Disasters in Developing Countries: Managing Risks 225 significantly higher. Natural and technological disasters The mass media also play a role in disasters, espe- are nowhere near this costly in such a short period. cially global news organizations such as the BBC and CNN, whose reporting brings global sympathy and assis- tance to devastated regions of the world. These media Taking Action provide the lens through which most of the world views disasters (Seydlitz and others 1991; Baum and Fleming It is sometimes thought that the solution to better dis- 1993). As Rosenthal, Comfort, and Boin suggest, "If CNN aster planning and management lies in more and better defines a situation as a crisis, it will indeed be a crisis technologies. Even when addressing natural disasters, in all its consequences. Mediazation will be one of the there is a tendency to fall back on technological and driving forces in the world of future disasters and crises" engineering solutions. In Florida, six million people now (forthcoming:7). Such reporting is generally not suffi- live in counties along the ocean shore where there were cient in planning for disasters, however, since plan- 500,000 residents 60 years ago. Though hurricanes reg- ning for a disaster is not "news." ularly hit this area, the loss of life has steadily declined. Disaster management officials should take advantage This is attributed to better warning systems and evac- of the penetrating reach of local and national media to uation plans (Ingelton 1999:30) as well as improved involve them in disaster planning and management. Not construction materials. Developing countries do not only can the media carry warnings of immediate threats, have such resources, however, and there are calls by they can broadcast other aspects of disaster prepara- environmental groups to halt construction in disaster- tion and management and provide information to people prone areas. who are unable to read newspapers and disaster-man- Globally, some changes have already occurred or are agement handouts. in the process of being established, although overall Education. At a formal level, scientists and social sci- strategies are still lacking. International Strategy for Dis- entists plan for and deal with disasters. Following aster Reduction 1999 (UN/ISDR 1999) newsletter Hurricane Mitch in Central America, projects to train describes the following range of activities and programs: more meteorologists in weather forecasting were imple- · Numerous agencies and groups in the region are now mented (Reconstruction 1999:47). Schools at the pri- involved in disaster planning and management; mary and secondary levels are also good venues for · A variety of educational activities have been initiated; spreading awareness messages, since children can bring · A number of partnerships and links have been this information home to their families. created; Standardizing terms and concepts can also improve · Information networks have been established; disaster management, especially with regard to statis- · Relevant guides and reports have been published; tics. The Regional Disaster Information Center (CRID) · Numerous disaster-related meetings and conferences in San Jose, Costa Rica, has created a thesaurus of stan- have been held; dardized descriptors used in its documentation data- · Legislation for disaster prevention has been passed base (CRID 2000:4). Also, the Centre for Research on or advocated; and the Epidemiology of Disasters (CRED) has undertaken · Warning systems have been planned or put in place. major efforts to improve the data it obtains and processes, Global disaster networks offer opportunities for especially for its database containing information on information sharing. The Global Disaster Information more than 12,000 disasters that have occurred glob- Network (GDIN) and the Disaster Research Center (DRC) ally since 1900. Their efforts have led to attempts to are two organizations engaged in studying disasters and standardize terms so that statistical compilations do not bringing together professionals from the field of disas- combine incompatible phenomena. ter management. Increased use of the Internet world- Improving disaster statistics rather than docu- wide makes their research available to a growing audience menting hazards can also help researchers understand and improves access to information that can be used in the negative impacts of disasters. Officials and NGOs in disaster planning. developing countries should be trained in data collection 226 Building Safer Cities: The Future of Disaster Risk so that statistical databases of deaths, injuries, prop- have altered this demographic. Some countries have not erty damage, and social disruption can be created and recognized that, although they were once primarily rural, lessons learned. The statistics are useful in monitoring they are becoming more urbanized each day. This lag trends and predicting future disasters in hopes of saving is illustrated by the fact that government departments lives and property. dealing primarily with agricultural issues are still the Government and community action. Some com- main organizations responsible for national-disaster munities emerge stronger and spur governments to action planning and management, even where the population to better prepare for or prevent future disasters. As stated has become urbanized. earlier, respondents to disasters can only work with the There are several recommendations developing coun- existing infrastructure and social systems so change must tries and international groups might want to adopt to come before the next disaster, not after. If a flow of improve disaster readiness. The governments of devel- communication is needed to respond to a disaster and oping countries currently have various levels of disaster that communication is absent under normal circum- planning, although most are not fully ready to address stances, it will not jump into existence during a disas- their vulnerabilities in the event of major natural or tech- ter. Improved technologies cannot help either, unless nological disasters. There should be a commitment by people are trained to use them. Better radios, for exam- policy and decisionmakers to put disaster planning at ple, do not translate into better communication unless the top of their agendas. A passive willingness to accept put into use earlier as part of a normal functioning system. disasterplanningisnotenough.Theremustalsobefollow- There are several stages in the overall disaster-risk- through measures, particularly financial support. Olson management-planning process, as follows: and his colleagues note that, while a national law · Formulation of emergency response policies; addressingdisasterplanninghadbeenproposedfiveyears · Adoption of policies by government and social serv- earlier in Nicaragua, it was not until the 1999 legislative ice agencies; session that a disaster fund, with a corresponding budget · Development and adoption of appropriate region- allocation, was provided (1999:56). Disaster-related specific programs; laws without financial support are ineffective. Countries · Identification of target audiences for each level of an havetolookatallresources--local,regional,andnational-- emergency response program (these could range from tofunddisaster-managementprograms.EvenintheUnited village women to higher levels of national govern- States,Superfundlegislationtocleanuptoxicsitesispolit- mental bureaucracies); ical and subject to budget cuts from Congress. · Program implementation by government and private There is a need for specific risk analysis of the possi- organizations; ble disasters in each country. To study natural disasters, · Application of both general and specific disaster new tools such as Geographic Information Systems (GIS) principles; canbeusedinconjunctionwithsocialnetworkstounder- · Obtain feedback on whether the desired results are stand the dynamics of a particular area. The Global achieved; and Disaster Information Network and the Disaster Research · Take into account social changes that may affect the Center, mentioned earlier, serve as forums for exchang- newer threats and vulnerabilities that might occur ing information about disasters--useful to those with (adapted from Wenger and Drabek 1987; see also Internet access. A British satellite developer announced WHO 1999). plans to launch in 2002 a constellation of five satellites devoted to monitoring natural and manmade incidents. Recommendations for Developing Countries Anumberofothertechnologicaladvanceshavebeenmade that can benefit disaster-management organizations, As recently as a decade ago, approximately 80 percent provided they have access to emerging technologies of the population lived beyond the outskirts of cities with (Richharia2001:23;Hofmann-Wellenhofandothers1994). populations of 100,000 or more (Kasarda and Crenshaw Disaster planning should be part of developmental 1991:472). As discussed earlier, the shifts in population planning and should be discussed at local and national Urban Vulnerability to Disasters in Developing Countries: Managing Risks 227 levels. More than two decades ago, Long (1978) argued to developing countries, and markedly raising the dis- that disaster planning should be linked to national devel- aster consciousness of their own personnel. opment planning, but as John Mitchell notes, "Until IFIs can play a significant role in disaster planning recently, disaster relief and long-term development and management since they operate in most develop- tended to be seen as distinct entities" (1998:14). It is ing countries. The following offer several suggested perhaps equally important to look at those developing courses of action: countries where such a link has been made. · Raise the consciousness of officials at all levels of gov- Evidence suggests that emphasis on disaster-manage- ernment on disaster-management issues; ment planning and coordination at the national level is · Raise community awareness of disaster planning and the most effective (Dynes 1990). Once a policy has been mitigation measures; articulated, the center can work with local governments · Link disaster management to development objectives; to plan regional contingencies. The responsibilities of the · Assist in the development of databases through the numerous local public and private emergency response collection, analysis, and storage of standardized data organizations can be spelled out in advance and backed- on disaster losses. Assist developing countries in link- up by the resources available at the national level. The ing national disaster systems into larger networks; Federal Emergency Management System (FEMA) in the · Promote global approaches to disasters and involve United States is an example of an organization that coor- multinational corporations that cross national bound- dinates responses to natural and technological disasters aries; and makes federal disaster assistance available to severely · Focus attention on the social aspects of behavior that affected communities. FEMA also has a data bank of dis- cause disasters; and aster planning and response materials. · Conduct follow-up assessments several years after Finally, more attention needs to be paid to extreme disaster policies, programs, and plans have been disasters in two respects. First, there is a need to plan for put in place. a catastrophic disaster that could massively impact a The globalization of culture should also be taken into society and be truly national in scope. Some developing account, since knowledge of natural and technological countries have experienced such events. The impact of disasters and lessons learned from other disasters often HurricaneMitchonHonduraswascatastrophic.Thesame comes from the global media. And, since people gen- istrueofthe1988earthquakeinArmenia(seePoghosyan erally experience no more than one disaster in a life- 2000:26­29 for reasons why official statistics on casual- time, people should be prepared and not have to be ties and property losses, though high, might have been victims of such an experience to mitigate, prepare for, underestimates). Likewise, monsoon rains in India in and respond quickly to a disaster. August 2000 left millions homeless in three eastern statesandwipedoutwholevillages(Farooq2000).Second, increasingly, there will be major disasters whose origins Bibliography will be distant from where they eventually impact. Such Acharya, Rohini. 1999. The Emergence and Growth of Biotechnol- circumstances present difficulties for disaster managers ogy: Experiences in Industralised and Developing Countries. whomustgrapplewithanincreasingnumberofvariables. Chelton, England: E. Elgar. Adams, Barbara. 1998. Timescapes of Modernity: The Environment Recommendations for International Financial and Invisible Hazards. London: Routledge. Institutions ADB (Asian Development Bank). 1991. Disaster Mitigation in Asia and the Pacific. Manila. In the last decade, several international financial insti- Alba, Richard. 2000. "Ethnicity." In Edgar Borgatta and Rhonda tutions (IFIs) have taken significant steps with respect Montgomery, eds., Encyclopedia of Sociology. NY: Macmillan. to addressing disasters. These improvements include Barro, R., E. Fama, D. Fischel, A. Meltzer, R. Roll, and L. Telser. establishing environment departments, explicitly build- 1989. Black Monday and the Future of Financial Markets. Home- ing prevention requirements into recovery loans made wood, IL: Dow Jones-Irwin. 228 Building Safer Cities: The Future of Disaster Risk Baum, A. and I. Fleming. 1993. "Implications of psychological Dynes, Russell R. 1990. "Community emergency planning: False research on stress and technological accidents." American Psy- assumptions and inappropriate analogies." Preliminary chologist 48: 665­672. Paper # 145. Newark, DE: Disaster Research Center, Univer- Beck, U. 1995. Ecological Politics in an Age of Risk. Cambridge, sity of Delaware. England: Polity Press. Dynes, Russell R. Forthcoming. "Social science research: Rele- Beck, U. 1999. World Risk Society. Cambridge, England: Polity vance for policy and practice." In Claire Rubin, ed., Study on Press. Improving Earthquake Mitigation. Report to FEMA:105­124. Berke, Philip R. 1995. "Natural-hazard reduction and sustain- Dynes, R., E. L. Quarantelli, and D. Wenger. 1990. Individual and able development: A global assessment." Journal of Planning Organizational Response to the 1985 Earthquake in Mexico City. Literature 9: 370­375. Newark, DE: Disaster Research Center, University of Delaware. Bowonder, B. 1985. "The Bhopal incident: Implications for devel- Farooq, Omer. 2000. "Monsoon rains leave millions homeless oping countries." The Environmentalist 5: 89­103. in India." Wilmington News Journal August 28, 2000: A2. Bowonder, B. and Jeanne Kasperson. 1988. "Hazards in devel- Fisher, Ian. 2000. "Brownouts in Kenya highlight its chaos and oping countries: Cause for global concern." Risk Abstracts 5: misrule." New York Times June 25, 2000: 10. 103­109. Flynn, James, Paul Slovic, and C. K. Mertz. 1993. "Decidedly Bowonder, B., J. Kasperson, and R. Kasperson. 1985. "Avoiding different: Expert and public view of risks from a radioactive future Bhopals." Environment 27: 6­13, 31­37. waste repository." Risk Analysis 13: 643­648. Britton, Neil. 1991. "Hazardous industries and risky technolo- Freitas, Carlos, Marcelo Porto, and Jorge Machado, eds. 2000. gies: A new role for insurance?" Unpublished paper. Acidentes Industriais Ampliados: Desafios E Perspectivas Para O Brody, Jane. 2000. "When post-traumatic stress grips youth." Controle E A Prevencao. Rio de Janiero, Brazil: Editora Fiocruz. New York Times. March 21, 2000: D8. (In Portuguese). "CRID standardizes disaster terminology." 2000. Disasters: Pre- Freudenburg, W. R. 1997. "Contamination, corrosion and the paredness and Mitigation in the Americas 81 (July): 4. social order: An overview." Current Sociology 45: 19­40. Cutter, Susan. 1991. "Fleeing from harm: International trends Fuchs, R., C. Brennan, J. Chami, F. Lo, and J. Vitto, eds. 1994. in evacuation from chemical accidents." International Journal Mega-city Growth and the Future. Tokyo: United Nations Uni- of Mass Emergencies and Disasters 9: 267­285. versity Press. Davis, Ian. 1987. "Safe shelter within unsafe cities: Disaster vul- Geipel, Robert. 1993. "The River Danube flood of 27 March nerability and rapid urbanisation." Open House International 1986." In J. Nemec, J. Nigg, and F. Siccardi, eds., Prediction 12: 5­14. and Perception of Natural Hazards. Proceedings of Symposium. Davis, Morris and Steven Seitz. 1982. "Disasters and govern- Dordrecht: The Netherlands: Kluwer Academic Publisher: ments." Journal of Conflict Resolution 26: 547­568. 111­118. Degg, Martin. 1992. "Natural disasters: Recent trends and future Gentleman, Amelia. "Russia averts nuclear disaster." Wilming- prospects." Geography: 198­209. Disaster Resistant Commu- ton News Journal September 18, 2000: A5. nities Initiative: Evaluation of the Pilot Phase Year 2. Year 2 Report. Giddens, Anthony. 2000. Runaway World. New York: Routledge. 2000. Newark, DE: Disaster Research Center, University of Gilbert, Alan and Josef Gugler. 1992. Cities, Poverty, and Devel- Delaware. opment: Urbanization in the Third World. Oxford: Oxford Uni- Dogan, Mattei and John Kasarda, eds. 1988. The Metropolis Era: versity Press. Mega-Cities. Beverly Hills, CA: Sage. Glantz, Michael and Melanie Mason. 1994. "Urban drought." Dommelen, Ad Van. 1999. Hazard Identification of Agricultural In Stanley A. Changnon, ed., The Lake Michigan Diversion at Biotechnology: Findings From Relevant Questions. Utrecht, The Chicago and Urban Drought. Boulder, CO:UCAR:163­186. Netherlands: International Books. Gugler, Josef, ed., Cities in the Developing World: Issues, Theory, Drury, A. and Richard Olson. 1998. Disasters and political unrest: and Policy. Oxford: Oxford University Press. 1997. An empirical investigation. Journal of Contingencies and Crisis Gulkan,PolatandOktayErgunay.1999."Casestudy."InJonIngle- Management 6: 153­161. ton, ed. Natural Disaster Management: A Presentation to Com- Dugger, Celia W. 2000. "A cruel choice in New Delhi: Jobs vs. memorate the International Decade for Natural Disaster Reduction a safer environment." New York Times November 24, 2000. (IDNDR) 1990­2000. Leicester, England: Tudor Rose: 176. Urban Vulnerability to Disasters in Developing Countries: Managing Risks 229 Haque, C. Emdad and Danny Blair. 1992. "Vulnerability to trop- Khandekar, M., T. Murty, D. Scott, and W. Baird. 2000. The 1997 ical cyclones: Evidence from the April 1991 cyclone in coastal El Nino, Indonesian forest fires and the Malaysian smoke prob- Bangladesh." Disasters 16: 217­229. lem: A deadly combination of natural and man-made hazard. Herath, S. and T. Katayama, eds. 1994. Proceedings of the Work- Natural Hazards 21: 131­144. shop. Towards Natural Disaster Reduction. Tokyo, Japan: Inter- Kim, Dwi-Gon. 1999. "Flood hazard in Seoul: A preliminary national Center for Disaster-Mitigation Engineering. assessment." In James K. Mitchell, ed., Crucibles of Hazard: Herzer, Hilda and N. Clichevsky. "Floods in Buenos Aires: Learn- Mega-Cities and Disasters in Transition. Tokyo: United Nations ing from the past." 2000. In Alcira Kreimer and Margaret University Press: 92­118. Arnold, eds., Managing Disaster Risk in Emerging Economies. Kim, Susan. 2000. "Millions homeless in India flooding." Dis- Washington, D.C.: World Bank: 32­45. aster News Network August 29, 2000. Hewitt, Kenneth. 1997. Regions of Risk. Essex, England: Addison Kraus, Nancy, A. Torbjorn, and Paul Slovic. 1992. "Intuitive tox- Wesley Longman. icology: Expert and lay judgments of chemical risks." Risk Hofmann-Wellenhof, B., H. Lichtenegger, and J. Collins. 1994. Analysis 12 (1992): 215­232. Global Positioning System: Theory and Practice. New York: Krimgold, Frederick. 1992. "Modern urban infrastructure: The Springer-Verlag. Armenian case." Environmental Management and Urban Vul- Hohn, Uta. 1995. "Issues in Urban Earthquake Risk." GeoJournal nerability. Washington, D.C.: The World Bank: 215­218. 37: 386. Kroll-Smith, J. Stephen, and Stephen Couch. 1991. "What is a Huq, Saleemul. 1999. "Environmental hazards in Dhaka." In disaster? An ecological-symbolic approach to resolving the James K. Mitchell, ed., Crucibles of Hazards: Mega-Cities and definitional debate." International Journal of Mass Emergencies Disasters in Transition. Tokyo, Japan; United Nations University and Disasters 9: 355­366. Press: 119­137. Laquian, Aprodicio. 1994. "Social and welfare impacts of ICRC (International Committee of the Red Cross and Red Cres- mega-city development." In R. Fuchs, E. Brennan, J. Chamie, cent Societies). 2000. World Disasters Report. Geneva. F. Lo, and J. Uitto, eds., Mega-city Growth and the Future. Tokyo: Ingleton, Jon, ed. 1999. Natural Disaster Management: A Pre- United Nations University Press: 192­219. sentation to Commemorate the International Decade for Natural Lee, Leonard. 1992. "Making the connection." Disaster Recovery Disaster Reduction (IDNDR) 1990-2000. Leicester, England; 5: 8­9. Tudor Rose. Lindell, Michael and Ronald Perry. 1995. "Identifying and Isaacs, J., J. Stork, D. Goldstein, and G. Wick. 1975. "Effect of managing conjoint threats: Earthquake induced hazardous vorticity pollution by motor vehicles on tornadoes." Nature materials releases." Unpublished paper. 253: 254­255. Lindell, Michael and Ronald Perry. 1998. "Earthquake impacts Jasanoff, S. 1998. "The political science of risk perception." and hazard adjustments by acutely hazardous materials facil- Reliability Engineering and System Safety. 59: 91­99. ities following the Northridge earthquake." Earthquake Spec- Jones, Barclay. 1991. "Population growth, urbanization and tra 14: 285­299. disaster risk and vulnerability in metropolitan areas: A con- Long, Frank. 1978. "The impact of natural disasters on Third ceptual framework." Unpublished paper. World agriculture: An exploratory survey of the need for some Jones, N, P., E. Noji, G. Smith, and R. Wagner. 1993. "Casualty in new dimensions in development planning." American Journal earthquakes."InKathleenTierneyandJoanneNigg,eds.,Socioe- of Economics and Sociology 37: 149­163. conomicImpacts:1993NationalEarthquakeConference. Memphis, Lovins, Amory and J. H. Price. 1981. Energy Policies for Resilience TN: Central United States Earthquake Consortium: 19­68. and National Security. San Francisco: Friends of the Earth Inc. Kallender, Paul. 2000. "Disaster could endanger chip industry." Mitchell, James K. 1996. The Long Road to Recovery: Community Electronic News 46 (July 3, 2000): 1­27. Responses to Industrial Disaster. New York: United Nations Uni- Kasarda, John and Edward Crenshaw. 1991. "Third world urban- versity Press. ization: Dimensions, theories and determinants." Annual Review Mitchell, John. 1998. "Protecting development." In John Twigg, of Sociology 17: 467­501. ed., Development at Risk? Natural Disasters and the Third World. Kempster, Norman. 2000. "Nuclear sites never will be safe, study London: The U.K. National Co-ordination Committee for the says." Wilmington News Journal August 8, 2000: A5. International Decade for Natural Disaster Reduction: 14­19. 230 Building Safer Cities: The Future of Disaster Risk Monroe, Scott. 1992. "Chelyabinsk: The evolution of disaster." Porto M. and C. Freitas. 1996. "Major chemical accidents in Post-Geography 33: 533­545. industrializing countries: The socio-political amplification of Mydans, Seth. 1997. "Its mood dark as the haze, southeast Asia risk." Risk Analysis 16: 19­29. aches." New York Times October 26, 1997: 3. Qingzhou, Wu. 1989. "The protection of China's ancient cities Nateghi, Fariborz. 2001. "Earthquake scenario for the mega-city from flood damage." Disasters 13: 193­227. of Teheran." Disaster Prevention and Management 10: 95­101. Quarantelli, E. L. 1988. "Future disasters in the United States: Nelson, Benjamin. 2000. Global Health: The U.S. and U.N. Response More and worse." Preliminary Paper #125. Newark, DE: Dis- to the AIDS Crisis in Africa. 2000. Available from . Quarantelli, E. L. 1991. "Criteria for evaluating disaster plan- Oberai, A. S. 1993. Population Growth, Employment and Poverty ning in an urban setting." In Francesco M. Battisti, ed., La in Third-World Mega-Cities. New York: St. Martin's Press. Citta e l'Emergenza. Milan, Italy: Franco Angeli: 39­63. Oberg, James E. 1988. Uncovering Soviet Disasters. New York: Quarantelli, E. L. 1996. "Just as a disaster is not simply a big acci- Random House. dent, so a catastrophe is not just a bigger disaster." The Journal Olson, Richard and A. Cooper Drury. 1997. "Un-therapeutic of the American Society of Professional Emergency Planner: 68­71. communities: A cross-national analysis of post-disaster polit- Quarantelli, E. L. 1998a. "Major criteria for judging disaster plan- ical unrest." International Journal of Mass Emergencies and Dis- ning and managing: Their applicability in developing soci- asters 15: 221­238. eties." Preliminary Paper # 268. Newark, DE: University of Olson, Richard Alvarez, B. Baird, A. Estrada, V. Gawronski, and Delaware. J. Prieto. 1999. Disaster and Institutional Response: Hurricane Quarantelli, E. L., ed. 1998b. What Is a Disaster? Perspectives on Georges in the Dominican Republic and Hurricane Mitch in Hon- the Question. London: Routledge. duras and Nicaragua, September-October 1998. Folsom, CA: Quarantelli, E. L., Brenda Phillips, and D. Hutchinson. 1983. Robert Olson Associates. Evacuation Behavior: Case Study of the Taft, Louisiana Chemical Olson, Richard, Juan Pablo Sarmiento Prieto, Robert A. Olson, Tank Explosion Incident. Newark, DE: Disaster Research Center, Vincent T. Gawronski, and Amelia Estrada. 2000. The Mar- University of Delaware. ginalization of Disaster Response Institutions: The 1997­1998 El Reconstruction and Transformation of Central America After Hur- Nino Experience in Peru, Bolivia, and Ecuador. Boulder, CO: Nat- ricane Mitch. 1999. General Secretariat Central American Inte- ural Hazards Research and Applications Information Center, gration System. University of Colorado. Richharia, M. 2001. "Mobile Satellite Communications: Princi- Perry, Ronald. 1985. Comprehensive Emergency Management: Evac- ples and Trends. New York: Addison-Wesley. uating Threatened Populations. Greenwich, CT: JAI Press. Rogers, George. 1997. "The dynamics of risk perception: How Petterson, John. 1987. "The reality of perception: Demonstra- does perceived risk respond to risk events?" Risk Analysis 17: ble effects of perceived risk in Goiania, Brazil." Practicing Anthro- 745­757. pology 10: 3­4, 8­9, 12. Ropeik, David. 2000. "Let's get real about the risks we fear." Picou, J. S. 2000. "The "talking circle" as sociological practice: Wilmington News Journal August 19, 2000. Cultural transformation of chronic disaster impacts." Socio- Rosenthal, Uriel, Louise Comfort, and Arjen Boin, eds. Forth- logical Practice; A Journal of Clinical and Applied Sociology 2: coming From Crises to Contingencies: A Global Perspective. 77­97. Sanger, David. 1991. "A crack in Japan's nuclear sangfroid." Picou, J. S. and D. A. Gill. 1999. "The Exxon Valdez disaster as New York Times February 17, 1991. localized environmental catastrophe: Dissimilarities to risk Savic, Misha. 2000. "Cyanide spill killing all life in Tisa River." society theory." In M. J. Cohen, ed., Risk in the Modern Age. Wilmington News Journal February 13, 2000: 2. London: Macmillan: 143­170. Scanlon, Joseph. 1992. Convergence Revisited: A New Perspective Poghosyan, G., ed. 2000. Ten Years of Loneliness. Yerevan, on a Little Studied Topic. Boulder, CO: Institute of Behavioral Armenia: Noyan Tapan. Science, University of Colorado. Porfiriev, Boris. 1991. "Environmental dimensions of national Scanlon, Joseph and Massey Padgham. 1980. The Peel Regional security: A trial of systems analysis study." Newark, DE: Dis- Police Force and the Mississauga Evacuation. Ottawa: Canadian aster Research Center, University of Delaware. Police College. Urban Vulnerability to Disasters in Developing Countries: Managing Risks 231 Schmeck, Harold. 1987. "Panel discounts special hazards in gene "Tourism in southeast Asia hurt by fires in Borneo and Sumatra." splicing." New York Times August 15, 1987: 1, 7. New York Times September 26,1997. Schmuck-Widmann, Hanna. 1998. "Floods in Bangladesh: Towfighi, Parviz. 1991. "Integrated planning for natural and tech- Disaster or part of the life world? European engineers' per- nological disasters." In Alcira Kreimer and Mohan Munasinghe, ception, knowledge and techniques in comparison to local eds., Managing Natural Disasters and the Environment. Wash- and indigenous knowledge and strategies." Unpublished paper. ington, D.C.: Environment Department, World Bank: 106­110. Seitz, Steven and Morris Davis. 1984. "The political matrix of UNFPA (United Nations Population Fund). 1991. Population and natural disasters: Africa and Latin America." International Jour- Environment: The Challenges Ahead. New York. nal of Mass Emergencies and Disasters 2: 231­250. UN/ISDR (United Nations International Strategy for Disaster Seydlitz, Ruth, J. William Spencer, Shirley Laska, and Elizabeth Reduction). 1999. ISDR: Newsletter for Latin America and the Triche. 1991. "The effects of newspaper reports on the public's Caribbean 15: 1­40. response to a natural hazard event." International Journal of U.S. GAO (United States General Accounting Office). 2000. Mass Emergencies and Disasters 9: 5­29. Nuclear Safety: Concerns with the Continuing Operation of Soviet- Shargorodsky, S. 1993. "A lethal Soviet legacy." Wilmington Designed Nuclear Power Reactors. Washington, D.C. News Journal December 24, 1993: 2. Vogt, B. and J. Sorensen. 1987. Evacuation in Emergencies; An Shefner, Jon. 1999. "Pre- and post-disaster political instability Annotated Guide to Research. Oak Ridge, TN: Oak Ridge National and contentious supporters: A case study of political ferment." Laboratory. International Journal of Mass Emergencies and Disasters 17: Vogt, Jurgen and Francesca Somma, eds. 2000. Drought and 137­160. Drought Mitigation in Europe. Dordrecht: Kluwer. Showalter, Pamela and Mary Myers. 1994. "Natural disasters in Vorobiev, Y. L., ed. 1998. Disasters and Man: Book 1 The the United States as release agents of oil, chemicals or radio- Russian Experience of Emergency Response. Moscow: AST-LTD logical materials between 1980­1989: Analysis and recom- Publishers. mendations." Risk Analysis 14:169­182. Wade, Nicholas. 2000. "Global disasters in the making." The Slovic, Paul. "Perception of risk." Science 236 (1987): 280­285. New York Times July 18, 2000 Available from . Looking from the past into the future." In Andrew Kirby, ed., Walker, P. 1992. "Foreign military resources for disaster relief: Nothing To Fear: Risks and Hazards in American Society. An NGO perspective." Disasters 16: 153­159. Tucson, AZ: University of Arizona Press: 241­260. Waterbury, John. 1979. Hydropolitics of the Nile Valley. Syracuse, Stallings, Robert. 1991. "Ending evacuations." International Jour- NY: Syracuse University Press. nal of Mass Emergencies and Disasters 9: 183­200. Webb, Gary, Kathleen Tierney, and James Dahlhamer. 1999. Swarns, Rachel L. 2000. "Mozambique confronts huge flood "Businesses and disasters: Empirical patterns and unanswered losses but sees ray of hope." New York Times April 30, 2000: 9. questions." Preliminary Paper # 281. Newark, DE: Disaster Thompson, Mark. 1987. "Toxin killed 1,000 in U.S.S.R." The Research Center, University of Delaware. Philadelphia Inquirer March 25, 1987: 11A. Wenger, Dennis and Thomas Drabek. 1987. "Emergency pre- Tierney, Kathleen. 1990. "Risk of hazardous materials releases paredness and response planning: An overview of seismic following an earthquake." Preliminary Paper #152. Newark, policy related issues." Preliminary Paper #121. Newark, DE DE: Disaster Research Center, University of Delaware. Disaster Research Center, University of Delaware. Tierney, Kathleen. 1999. "Toward a critical sociology of risk." WHO (World Health Organization). 1999. Community Emergency Sociological Forum 14: 215­242. Preparedness: A Manual for Managers and Policy-Makers. Geneva. Chapter 16 Natural Disasters and Urban Cultural Heritage: A Reassessment June Taboroff In August 2002, central Europe saw some of the most museums to upper galleries and buildings hoped to be devastating floods of the last 200 years. Newspapers, flood-proof. One of the most spectacular rescue efforts radio, and television heralded the extreme threats to the occurred the night of August 13, when more than 4,000 historic towns of the Czech Republic, Germany, and paintings from the Old Master's gallery at the baroque Hungary by the Jahrhundertflut. The heavy rains and Zwinger Palace were shifted in just seven hours. One high floodwaters forced the evacuation of thousands of of the rescue efforts involved attaching over-sized paint- people as cities and towns were submerged and lives ings to the ceiling with ropes. were lost. Estimates by re-insurers for total flood damage It will take months, if not years, to repair the damage. in Germany, Austria, and the Czech Republic reached The workshops and tools that would normally be used $14.7 billion. to restore Dresden's art treasures were swept away by These floods struck some of central Europe's top the floods. The greatest risk now is for buildings: "Nobody cultural attractions and most valued historic urban envi- knows what the damage is exactly. It will be months ronments. The World Heritage towns of Cesky Krumlov and months to wait and millions and millions to pay."2 and Prague were damaged. Media coverage of the dis- These are the same historic buildings that were lov- aster left no doubt that the protection of cultural her- ingly and faithfully restored after being firebombed itage is of concern to society at large. BBC News during World War II. They have since become major commented from Dresden, "The people here have tourist attractions and a source of reborn civic pride been as concerned for their own properties as they are for the city. for the future of their historic buildings."1 Numerous In Prague, another cultural treasure house, officials news reports highlighted the significant losses to the at museums, palaces, and archives are coping with the Czech and German economies from the closure of cul- worst flooding of the city since 1890. Some palaces close tural sites and cancellation of tourist visits due to the to the Vltava River were flooded. Storage facilities of the floods. National Museum were heavily affected: books from the Dresden, destroyed by Allied bombing in 1945, came National Museum Library were soaked and the lower under siege once again in August 2002 when floodwa- floors of historic buildings damaged. The architecture ters swelled the Elbe to five times its normal flow, poured archive of the Prague University of Technology was into the city, and submerged seven districts. Despite the under 10 feet of water, flooding 90 percent of this impor- efforts of thousands of volunteers, army personnel, tant archive of Czech architecture. In 1940 the archive and relief workers to sandbag the river, flood waters had been relocated to the ground floor of this historic advanced through the old town and other parts of the building. Thanks to the enormous effort of volunteers, city. The city's Royal Palace was hit by floodwater. most documents were removed from the building and Many of the city's historic squares were under water. In frozen in order to allow for possible future restoration. one of the swiftest salvage operations ever mounted to An SOS Architecture Archive fund has been launched save priceless works of art, museum staff, government to restore the drawings and documents. officials, firemen, policemen, and volunteers moved Less dramatic than the floods of August 2002 but about 20,000 works of art from Dresden's galleries and cumulatively as damaging are the annual floods in Venice. 233 234 Building Safer Cities: The Future of Disaster Risk In the 20th century, the relative sea level increased by orities, improving the security of collections and more than 23 centimeters, and the number of tides over buildings, and training staff. Although many cultural 100centimetershassurged.Incontrasttothedecadeafter institutions in industrialized countries have their own 1923, when there were just 7 tides over 100 centimeters, emergency plans, they are often based on the case of in the decade after 1990, 114 floods measured over 100 fire, not floods or other types of disasters. All emer- centimeters, and in 2000 there were 19. Recent archae- gency plans need to be revised to address a range of ological studies have shown that the city is sinking much disasters. faster than previously thought. Using new carbon dating · Coordinate crisis management at both the institu- techniques, a team of archaeologists calculated the rate tional and urban/regional level. Experience indicates of relative sea-level rise over 1600 years. They found that institutional coordination has worked quite well, layer upon layer of pavements and foundations laid over for example at national museums, but regional the centuries in a constant effort to keep the city above coordination within urban and other geographic areas water.3 Such additional infrastructure increases costs has been less successful (as demonstrated by recent to the city and property owners. events in Prague). Much of Venice, especially St. Mark's Square, has a · Inform the public about the risks to cultural her- severe dampness problem. This main square lies lower itage and the need for investment in protecting her- than other parts of the city and begins to flood when itage from natural disasters. This will create an the tide is higher than 70 centimeters. Green algae grows active constituency to support preventive measures. on the porous brick and stone work of the palaces along As was evident in central Europe, the public at the Grand Canal, causing further damage. After 50 years large, through their work as emergency volunteers, of neglect and indecision, Italian authorities have finally was responsible for helping to save many treasures. begun dredging canals, raising city pavements, and The subject of urban cultural heritage is important repairing damaged sea walls. The cost is estimated at for several reasons. First, the severity and frequency of more than $40 million per year, however. The deeper natural disasters is arguably on the increase. Whether and more frequent the annual floods become, the fiercer weather-related disasters or geophysical processes, there the debate grows about how to stop them. Meanwhile, are enormous costs associated with inadequate pre- tour operators study the tides, as do shopkeepers and paredness and post-disaster reconstruction. Recent residents to calculate the opportunity and physical costs episodes of extreme weather include the high winds of flooding. that toppled trees in Versailles and the historic gardens of Paris, and the severe winter storms in 1999 that dam- aged buildings in Europe. Earthquakes regularly shatter Urban Heritage at Risk urban areas of Iran and India, destroying temples, hous- ing, and other elements of the historic environment. These recent examples present a vivid picture of what Second, cultural heritage tends to be concentrated in is at stake when natural disasters strike urban areas with urban areas where intense trading and business activi- significant cultural sites. They also highlight several key ties spawned displays of religious, civic, and private cre- points: ativityandwealth.Avastmajorityofsuchcitiesarelocated · Do not underestimate the vulnerability of cultural in coastal areas or along rivers and are therefore vul- heritage to natural disasters. Recent experience in nerable to flooding and landslides. A tally of the World central Europe starkly shows that environmental Heritage list shows that, of the cities listed, most are degradation, the inability to manage fluctuations in situated in coastal or riverine areas. There are also cities the volume of rivers, and other weather-related nat- in mountainous areas, such as Cuzco or Katmandu, that ural disasters are on the increase in many parts of are in seismically active zones. the world and can destroy precious heritage sites. Third, both capital and secondary cities are at risk. · Ensure that individual site management plans are Capital cities such as Prague and Dresden are often the adequate for disaster scenarios by establishing pri- focus of attention in times of disaster, and may be better Natural Disasters and Urban Cultural Heritage: A Reassessment 235 able to respond to emergencies. Provision of disaster risks to heritage. Representatives of 10 countries pre- information, organization of emergency services, and pared national profiles and drew up preliminary lists access to funds for reconstruction tend to be better in of heritage at risk. Only in the case of South Africa was such cities. In smaller towns such as Cesky Krumlov, the issue of risk preparedness mentioned. "Disaster man- where half of the buildings in the town's historic center agement plans for cultural property in case of fire, floods were flooded in the August 2002 floods, the physical and even the occasional earthquake, etc. are impera- and economic consequences can be equally profound tive."4 Natural disasters were cited in two reports. The or even greater, though resources for reconstruction are Ghana report mentioned sea-level change in coastal often fewer. areas as the cause of heritage loss. One example cited But even in capital cities, current procedures for pro- was encroachment of the sea at Katar in the Volta region, tecting cultural heritage have been shown to be inade- where Fort Peasantine has fallen victim to the waves. quate. A cultural heritage expert from the Czech Republic The Mali report cited drought and the southern spread filed this report: of the Saharan region as risk factors for built heritage, as they provoke population flight and the abandonment We were not prepared sufficiently. Most of the state of buildings. For Nigeria, Cameroon, Kenya, Zambia, museums, galleries and castles opened to the public had Zimbabwe, Uganda, and Botswana, risks that were iden- evacuation plans for the case of fire... In most cases a detailed program for the evacuation of the collections tified were attributable to human causes: environmen- was prepared (the priority, the succession of individual tal degradation (usually unspecified); uncontrolled items).... Nevertheless nobody was prepared for a flood development; and lack of awareness. Thus, for the major- of such immense scale. What failed completely and trag- ity of countries in southern and eastern Africa repre- ically was the forecast and quick information on the real sented at the workshop, averting the risks of natural extent of the danger (personal e-mail communication, disasters was overlooked in a discussion of heritage at October 2, 2002, from Jana Polakova, ICOMOS Czech Republic, to June Taboroff). risk, despite the fact that these countries are prone to severe natural events. With few exceptions, authorities responsible for safe- guarding cultural heritage, most often governmental Europe and the G8 bodies but also private museums and cultural institu- tions, have yet to be sufficiently prepared to respond In western and central Europe, the United Kingdom, to natural disasters. North America, and Japan, disaster-prevention measures and plans are a relatively high priority for institutions charged with safeguarding cultural heritage. The rapid Perceptions of Risk rescue of thousands of works of art by heritage and civil protection workers, assisted by large numbers of volun- Natural disasters, such as flooding and hurricanes, are teers in Dresden, Prague, and elsewhere in Germany, the recurrent rather than single events. Nevertheless, per- Czech Republic, and other affected Central European ceptions of risk to cultural heritage are highly variable countries is an inspiring demonstration of disaster coor- and provide a good indicator of readiness to tackle dis- dination.InCaliforniaaftertheearthquakeof1989,many aster prevention and mitigation. The following inter- museums and libraries were closed and rebuilt to con- national sample of current work on risk and vulnerability form to earthquake safety standards or torn down and gives an overview of thinking about heritage. replacedwithearthquake-resistantbuildings.IntheUnited States, institutions typically have individual emergency plans that are coordinated at a city- or state-wide level. The View from Southern Africa In the United Kingdom, English Heritage has the remit Contributors to the 2001 World Bank-sponsored to advise the public on ways to conserve and protect the consultative workshop on culture in Africa, held in historic built environment. It has produced, among Kimberley, South Africa, were asked to comment on other things, guidance on disaster preparedness and 236 Building Safer Cities: The Future of Disaster Risk first aid measures to be taken in the aftermath of floods. and there is little attention paid to risk preparedness. The organization can also deliver assistance through The authors remark, its regional offices and London headquarters. Natural disasters are affecting the heritage in different ways. The flooding of Antakya, ancient Antioch, caused Mediterranean Archaeological Sites serious damage to the service facilities of the historic town. In 2001 there were several earthquakes in Turkey.... The A recent effort to address risk assessment for archaeo- damage inflicted on monuments and urban areas by the logical sites in the Mediterranean, financed by the Euro- earthquakes of 1999 in Izmit and Duzce continue; some pean Commission's Euromed Heritage I Programme, of the damaged buildings have not received any care for examined natural and human risk factors (EUROMED 2 years .... Unfortunately some monuments are being 2002). Natural factors included were earthquakes, repaired without expert advice.... (ICOMOS 2002). volcanism, hydro-geological phenomena, coastal ero- sion, change in seacoasts, water erosion, climatic fac- ICOMOS Report on Risk-Preparedness tors, and biological factors. For the nine sites examined, for Cultural Heritage of which five were in urban areas, the natural risk fac- tors of significant importance were identified as hydro- At the behest of UNESCO, ICOMOS prepared a report geological (Tharros, Italy; Lixus, Morocco; Jericho, West in 1997 to assess the degree of protection to which cul- Bank); seismic (Jericho, West Bank); climatic (Tharros, tural heritage is integrated into disaster management Italy; Lixus, Morocco); and volcanic (Pompeii, Italy). ("Risk-Preparedness for Cultural Heritage," ICOMOS The report was concerned with probability of risk, and 1997). Based on responses from 14 countries, the major- gave only minimal attention to preparedness and mit- ity being industrialized countries with the exception of igation, except in general terms such as the need for Jordan, Pakistan, Zambia, and South Africa, certain pat- physical strengthening measures for standing structures terns were detected: and the need for orchestration of effort with govern- · Disaster management agencies do not generally dis- ment authorities. For these sites, representative of the tinguish cultural heritage from property. archaeological richness of the Mediterranean, little in · Disaster management for cultural heritage is handled the way of protection measures for natural disasters differently in each country. In some cases the Min- is being put in place. As a result, they remain highly istry of Culture or its equivalent works closely with vulnerable. law enforcement or civil defense agencies. In coun- tries such as New Zealand, each government insti- tution is required to establish disaster management The International Council on Monuments procedures and care for cultural heritage. In Sweden, and Sites Review of Risk the state administration is responsible for cultural The ICOMOS Heritage at Risk 2001­2002 report pro- heritage at the regional level; the disaster manage- vides a view of the opinions of heritage professionals. ment agencies responsible for cultural heritage list It is based on 75 reports received from ICOMOS mem- cultural heritage to be protected under the Hague bers, national committees, international scientific com- Convention. mittees, and affiliated groups. The 2001 report elaborates · Interest by the disaster management community in categories of sites that are most threatened: rural and integrating cultural heritage concerns into its activ- vernacular architecture; 20th century heritage sites; ities varies. Some countries display high concern for industrial heritage; religious heritage; archaeological the protection of cultural heritage in case of emer- sites; and cultural landscapes and gardens. Attention is gencies: Sweden has done so since the 1960s. Other concentrated on human risks, namely, new construction, countries may have general interest, but little in terms change of ownership, and lack of maintenance. There of follow-up procedures. is only occasional mention of natural disasters, however, · Opportunities for improving protection include new one example being Turkey's report on natural threats, legislation, coordination, and cooperation among Natural Disasters and Urban Cultural Heritage: A Reassessment 237 agencies, and the creation of one national body to U.K. Environment Agency, many of those most at risk monitor emergencies. still deny that they could be flood victims, despite dis- · Obstacles that were cited include incomplete legal turbing images of recent floods. In the autumn of frameworks; lack of equipment; lack of trained 2000, 10,000 properties were flooded in storms, although staff; and fragmented responsibilities. flood defenses are believed to have protected 280,000 · There are few fora where disaster management and properties. The survey reports that, while 95 percent cultural heritage professionals collaborate. of people in an area at risk agreed that flooding was a · Guidelines and methodologies for protecting her- serious issue, fewer than half accepted that it related to itage to improve risk preparedness are not generally them. Only one person in 20 had taken any action to available. In Belgium a special methodology was prepare for floods.5 A possible explanation for this behav- developed for archives. ior is that potential flood victims are reluctant to pro- · The type of disaster determines which professionals tect themselves against flood damage for fear it could carry out emergency response plans. This is at the reduce the value of their property. discretion of the Ministry of Interior or agency charged Certainly there are exceptions of disaster planning with disaster response, advised by heritage profes- in the field of heritage protection, but many govern- sionals. In Sweden, the Swedish Agency for Civil ments and institutions begin to put in place prepared- Emergency Planning is the focal point for discussions ness measures only after being hit by a disaster. Otherwise, among armed forces, planners, heritage specialists, the possibility of a disaster striking is considered a remote and other concerned groups. phenomenon. · There are few evaluations of emergency response for cultural heritage. The low response rate to the questionnaire (7 percent) After the Deluge? suggests that many countries are not yet convinced that disasterpreparednessisapriority.Althoughculturalinsti- When disaster strikes cultural heritage, there are spe- tutions often have emergency response plans--usually cial measures to take for historic buildings and sites. for fire--few countries have developed comprehensive These measures are extremely important because they disaster management plans, institutional structures, and can reduce further damage. The general rule is that guidelines. "remedial" work can be more damaging than the orig- This evidence from the field of cultural heritage shows inal disaster. While each type of disaster results in spe- that developing countries are even less ready to deal cific types of damage, there are certain principles that with natural disasters than wealthier countries. The guide conservation interventions. They concern record- Operations Evaluation Department of the World Bank ing, analysis of damage, and impacts on living heritage. said in a 1998 evaluation of disaster assistance: A key preventive factor is to record incidents. Various cultural heritage authorities have their own forms.6 For When the devastation caused by storms or other natural example, with flood damage it is essential not to dry out disasters in industrial and developing countries is com- old buildings too quickly with the application of heat. pared, the injury and death rates can be up to 100 times higher in the poorer developing countries.... Lack of mit- Timber paneling, boarded floors, and doors may warp igation is itself an indicator of underdevelopment, one that and twist; salts will burst out of old stone, and plaster- the Bank can help overcome (World Bank 1998). work and painted surfaces will peel and flake. The best general advice is to dry the building slowly through ven- tilation and with the aid of dehumidifiers. This can take Behavior of Denial several months but it is better than destroying original historic fabric that may be even more expensive to repli- In many countries, behavior with respect to cultural cate. Photographic recording and removal of important heritage is not unlike that of property owners in the timber elements to a cold store for freeze drying is rec- United Kingdom. In a 2002 survey conducted by the ommended for very important building elements. 238 Building Safer Cities: The Future of Disaster Risk In the case of earthquakes and other severe physical ICBS has asked that national committees be formed, damage, documentation followed by analysis of build- this initiative cannot go far beyond good intentions, ing structure and engineering is necessary before solu- since its lack of funding relegates it to a clearing house tions are proposed. The introduction of steel frames into and forwarder of information. masonry construction can cause additional and some- · UNESCO has sponsored several meetings on the sub- times fatal damage. A case in point is the Feyzullah Efendi ject of disasters. Most recently, it convened the Con- Madrasa in Istanbul, dating from the 18th century, which ference on World Heritage Mountain Cities and Natural was inappropriately repaired after the 1999 earthquake. Hazards from September 25­27, 2002. UNESCO is It is also important to highlight the human conse- discussing convening a roundtable on risk pre- quences of disasters for both physical and nonmaterial paredness in 2003. culture. The living heritage of urban centers can be · The International Center for the Study of the Preser- severely harmed by disasters, as they interrupt or erase vation and Conservation of Monuments (ICCROM) traditional customs, such as festivals, markets, craft pro- has supported several training courses on risk pre- duction, etc. The ICOMOS Heritage at Risk report (2002) paredness. Among these was a course in Central remarked, America after Hurricane Mitch. · The World Bank has directly financed several proj- But the effects on heritage places, monuments and sites ects with components targeting disaster prepared- of the loss and on indigenous language and the traditional ness and reconstruction. These include the China values, skills and knowledge that language embodies are more difficult to assess. The loss of understanding of the Yunnan Earthquake Reconstruction Project and the spiritual, intangible and cultural values of places is as dif- HondurasProFuturoproject,bothofwhichintervened ficult to document as it is irreplaceable. in historic urban areas. Under the Yunnan project, improved seismic resistance was incorporated into historic buildings in Lijiang, a World Heritage city. Recent Activities Despite the individual merits of these activities, how- ever, many sites remain unprotected and systems are At the international level, there have been recent ini- not yet in place to effectively reduce risk to heritage tiatives motivated by the need to address disaster pre- from natural disasters. vention. They are aimed at informing and motivating national governments and heritage institutions that are the guardians of heritage. Of particular interest are the Trends in Natural Disaster Preparedness following reports, meetings, training sessions, and activ- ities sponsored by international agencies, professional As the number of devastating weather-related and geo- organizations, and training centers: physical disasters and processes increases--flooding · ICOMOSHeritageatRiskannualreports.Thesereports in the United Kingdom, earthquakes in Turkey--the include individual commentaries by 75 national com- toll of urban heritage lost and damaged rises. Integrat- mittees. They are an important effort to assemble infor- ing cultural heritage concerns into national prepared- mation and examine trends. ICOMOS also set up an ness planning is the exception rather than the rule, International Committee on Risk Preparedness to especially in developing countries. develop professional guidance on risk management AnexampleofgoodpracticeistheSwissFederalOffice as an integral part of conservation practice. for Civil Protection, which includes a heritage section.7 · The International Committee of the Blue Shield (ICBS). The Swiss Federal Office has been entrusted with two Created in 1996 by the four nongovernmental organ- priority missions: providing aid in the event of disaster izations that represent professionals in the fields of and other emergencies; and protecting the public in the archives, libraries, monuments and sites, and muse- event of armed conflict. Civil protection also protects cul- ums, the ICBS works to protect cultural heritage tural heritage and participates in a regional context and threatened by natural disasters and war. Though the in cooperation with organizations specialized in disaster Natural Disasters and Urban Cultural Heritage: A Reassessment 239 rescue and relief and transfrontier rescue operations. The in vulnerability due to sea level rise, etc. Knowing Swiss Committee for the Protection of Cultural Property the extent, location, and condition of heritage is the serves as an advisory body to the Confederation, to the first step in its protection. Federal Department of Defense, Civil Protection and · Put in place risk-minimization procedures such as Sports, and to the Federal Office for Civil Protection. moving archives and collections from the basements The committee has a maximum membership of 25 indi- of buildings in flood-prone areas, upgrading museum viduals, all with an interest in the protection of cultural storage and display cases to meet seismic standards, property. In Switzerland implementation is the respon- and other measures to address known risks. sibility of the cantons in cases where responsibility does · Enforce building codes, especially in seismically active not lie directly with the confederation. They are expected zones, to reduce the likelihood of collateral damage. tocontributefinanciallytotheprotectionofculturalprop- · Ensure that heritage professionals are included in erty and take the following measures: national and local disaster and civil defense com- · Create the necessary legislative framework at the can- mittees to ensure they are notified of impending tonal level disasters and are able to work in close collaboration · Designate an office responsible for the protection of with fire and safety authorities. cultural property · When training civil defense and emergency workers, · Draw up an inventory of cultural property includemodulesonthespecialneedsofheritage,includ- · Create documentation to safeguard cultural property ing aftercare for important heritage objects and sites. · Determine the form of organization required at the · Evaluate the current experience of disaster response local level for cultural heritage to draw lessons and avoid ad-hoc · Train the staff responsible for the protection of cul- decisions. tural property. Today the realization that natural disasters could affect The requirement of the Swiss authorities for locali- cultural heritage is still not widespread. Governments ties to make financial contributions to heritage protec- are slow to mount preventive measures, and urban risk tion is especially forward looking. Sweden and the United preparedness does not always take heritage needs into Kingdom are other countries where advances have been account. Many organizations are prepared for little more made in planning for prevention. than fire emergencies. Experience in other areas of risk Public commitment to protecting heritage during dis- management suggests that "costs force customers to asters is increasing in some parts of the world. In cen- become aware of risks. They change behavior."8 The tral Europe heritage professionals, bolstered by emergency cost of loss may prove to be a persuasive argument for services and the public at large, helped save important national policymakers, especially when the cost of one treasures. In the United States a major tourism maga- lost masterpiece can climb to millions of dollars. Loss zine, Travel and Leisure, has initiated a new feature of income from tourism can also be crippling, and the page on heritage at risk after the recent floods. loss of sense of place when historic areas are destroyed is beyond dollar calculation. Improving Prevention through Policies and Planning The Way Forward Inanearlierpaper,"CulturalHeritageandNaturalDisasters: Incentives for Mitigation" (in Kreimer and Arnold 2000), Cultural heritage needs to be factored into overall dis- I discussed measures and incentives that could be adopted aster mitigation and management approaches. Cultural to better protect heritage. It is worthwhile to re-examine heritage professionals should make themselves known and update these measures and incentives. Key strategic to disaster mitigation professionals and disaster miti- elements that are suggested by current experience are: gation professionals should invite the participation of · Update and review inventories of historic places, heritage professionals in designing response systems. paying particular attention to condition and change Effective preparedness and mitigation strategies will 240 Building Safer Cities: The Future of Disaster Risk depend upon government agencies, heritage profes- 3. The Christian Science Monitor Website. "Plan to save Venice sionals, and emergency services working together to: from the sea draws praise, doubts." August 21, 2002. Available · Ensure that legislation affords heritage the neces- at . sary protection in the event of disaster 4. Conference Proceedings of the Kimberley Consultative Work- · Coordinate with disaster relief planners at local and shop on Culture in Africa, November 2001. national levels 5. FinancialTimesWebsite. "Residents fear costs of opting for flood · Determine the necessary organization for delivery defences." September 23, 2002. Available at: . · Prepare emergency plans at institutional, municipal, 6. For an example, see . regional, and national levels 7.. professionals 8. "Ready for anything, even meteorites." Financial Times, May · Provide documentation and training materials for the 20, 2002: 12. protection of cultural heritage · Design mitigation measures to ensure that heritage is properly conserved. Bibliography These are lessons that the World Bank and other inter- BBC News Website. Available from . national development institutions, bilateral donors, gov- TheChristianScienceMonitorWebsite. Availablefrom. ProVentionConsortiumcanpromoteintheirdisasterpre- Financial Times Website. Available from . paredness and mitigation work. Let us not wait until the ICOMOS (International Council on Monuments and Sites). 2002. next disaster disrupts lives and livelihoods and destroys H@R!: Heritage at Risk ICOMOS World Report on Monuments theculturalheritageofmankindbeforepreparationsbegin. and Sites in Danger 2001­2002. Paris. Kimberley Consultative Workshop on Culture in Africa. (forth- coming). Conference Proceedings. Kimberley, November 2001, Notes Government of South Africa, Government of Norway and 1. BBC News Website, News Front Page, Saturday 17 August the World Bank. 2002. Available from . in Emerging Economies. Washington, D.C.: The World Bank. 2. Museum curator, BBC News Website, Front Page, Saturday World Bank. 1998. "Disaster Assistance." World Bank Operation 17 August 2002. Available from . D.C. oys enezuela.V destr in Landslide buildings PART IV PROTECTING CRITICAL INFRASTRUCTURE FROM DISASTER IMPACTS Chapter 17 A New Structural Approach for the Study of Domino Effects between Life Support Networks Benoît Robert, Jean-Pierre Sabourin, Mathias Glaus, Frédéric Petit, Marie-Hélène Senay Life support networks allow the delivery of essential In order to respond to the limitations of emergency services such as energy, water, telecommunications, etc. measures and the evaluation of risks as a function of They are the guarantees, for a society, of properly func- the domino effect between several life support networks, tioning socioeconomic activisties, in respect to health a new structural approach was developed at the Centre and safety of the general population. These networks risque et performance of the École Polytechnique de Mon- are constituted of a multitude of interrelated infra- tréal. It is based on the characterization of the essential structure, which, upon the failure of one component, elements of the network, which allow it to carry out can provoke, by the domino effect, the failure of other the operations it is allocated. This approach focuses on components up to the point of a network-wide failure. the identification of the consequences of poor efficiency Therefore, the network can no longer carry out the oper- of the network, then on the evaluation of its vulnera- ations for which it was designed. This domino effect bilities. Thus, the domino effects can be identified and can also be observed between several networks that studied. This approach will be explained in general are interconnected. The goal of this report is to present terms after presenting the overall context of emergency a methodology for the evaluation of these domino effects measures and risk. among life support networks. This report is composed of five major parts. First, This work is in line with the current international the problems faced by life support networks will be trend toward establishing efficient and effective man- presented. Second, the general notion of risk will be agement plans for life support networks based on the explained in order to define the diverse challenges that antagonistic concepts of maximum production and min- the managers of these networks must face. Work is cur- imum risk. This approach is complex because the con- rently being done on risk and life support networks. sequences are great for populations and socioeconomic They will be fully explained and analyzed in order to activities upon failure of these networks. Thus, the need define the characteristics of the new structural approach, to have operational emergency measures based on an which will be presented in the last chapter. This will be exhaustive evaluation of risks makes itself felt follow- completed by a demonstration of the potential uses of ing the appearance of increasingly frequent catastro- the structure of this approach for risk studies, emergency phes. Analysis of the needs of emergency response response measures, and risk communication. Finally, the managers of life support networks established that the application of such an approach in developing countries behavior of these networks should have been studied will be presented. for the entire set of possible failure conditions, not only for the most extreme events. Finally, the evalua- tion of the domino effects between life support networks Context demands an extremely multidisciplinary work and the integration of technical, economic, and social expert- Life support networks are made up of a multitude of ise. Therefore, a very real problem of risk communica- civil infrastructure, which ensure the correct function- tion exists. ing of industrial activities and provide essential services 245 246 Building Safer Cities: The Future of Disaster Risk for citizens (Selcuk and Semih 1999; O'Leary 1997). Life Various life support networks are currently being support networks composed of civil infrastructure can studied by way of risk studies, the risk being the prod- be grouped according to the following categories uct of the probability of recurrence of an event by the (Isenberg 1991; Lau 1995): consequences that the event begets (Kaplan 1997). These · Electricity (generation, transportation, distribution studies generally only consider precise events (scenar- infrastructure, etc.) ios) that can provoke a failure (CAN\CSA 1991). They · Natural gas and liquid fuels (storage, transporta- are normally focused on obtaining a unique result tion, distribution infrastructure, etc.) (economic, technical, or industrial risk) and base them- · Potable water and wastewater (collection, treatment, selves on the study of a finite number of natural or tech- storage, transportation, distribution infrastructure, nical events (Stedinger and others 1996). etc.) Nevertheless, analysis of past catastrophes indicates · Telecommunications (broadcasting, cable transmis- that the actual methods of risk evaluation only par- sion, distribution infrastructure, etc.) tially reflect the real risk of failure of these networks. · Transportation (road systems, public transportation The origin of the events is rather of an anthropic nature; systems, etc.). in other words, they combine natural events with tech- Thus, these networks fulfill fundamental roles for the nical malfunctions and human intervention. In effect, proper functioning of a society by ensuring essential not only are the life support networks subject to unfore- services concerning the health and safety of populations seen natural turns of events, but equally the infrastructure and the proper functioning of the economy. When a that they are made up of differ in age, state, nature, life support network fails, the human and socioeco- design, etc. The management methods used call upon nomic repercussions are very significant. Therefore, they both automated systems and human intervention; how- cannot be ignored (Hubbert and Ledoux 1999). ever, several catastrophes are due or amplified by human Life support networks are made up of a set of strongly error (Hubbert and Ledoux 1999; Reason 1990). interrelated components. The components, which are On the other hand, current risk studies aim, for the directly related to the role of the network, form a pri- most part, to put a figure on the risk of failure, so that mary network, but parallel or secondary networks-- the identified consequences must, generally, be expressed such as remote control and computer networks--exist. as a number, often a dollar value. Social impacts, although The set forms an integrated system whose reliability quite tangible, are broached very lightly or not at all, depends on the set of its components. As soon as one while environmental impacts, even more intangible, are component exhibits a failure, the impact on the system often ignored. This approach, therefore, has a ten- not only depends on the importance of the component dency to minimize the consequences of a potential fail- but also on the importance of the components linked ure of a life support network. It does not permit a realistic to the faulty one. In addition, the components of sev- portrayal of the situation or the development of efficient eral networks can be linked, provoking repercussions and effective emergency and mitigation measures. In by way of the domino effect (Allen 1997; Plate 1996; addition, the interrelations between several life support Moses 1998; Lemperiere 1999). networks are often neglected because certain links that A good understanding of the dynamics of these net- unify these networks are not identified. works is therefore essential in order to avoid benign Finally, the current risk studies are generally carried malfunctions transforming into major crises. The numer- out without regard for the highly multidisciplinary nature ous catastrophic events observed throughout the of these works. In effect, to realize these works ade- world over the last few years demonstrate just how much quately, it is important to gather experts from all tech- the safety of the human and the natural environment nical, social, and emergency measures disciplines (as depend on the proper functioning of infrastructure. In much from the life support networks being studied as addition, the events of September 11th, 2001 in the from the communities affected by failures of the networks) United States show that infrastructure is vulnerable to as well as the various competent authorities (BIT 1990) varied acts of malevolence. and the other interconnected life support networks. A New Structural Approach for the Study of Domino Effects between Life Support Networks 247 Unfortunately, the communication between these con- other disciplines, such as economics, law, and commu- cerned parties is underestimated during the course of nication, must be taken into consideration. these studies. Risk Communication Risk: General Concepts Risk communication is generally described as an exchange (Covello and Merkhofer 1994; Leiss 1990), bilateral The notion of risk encompasses several concepts, which (CAN/CSA 1997) or interactive (U.S. NRC 1989), of must be clarified. This section sets out the general guid- information, perceptions, opinions, and preoccupations. ing principles of the management, evaluation, percep- "As a discipline, risk communication tries to achieve tion, and acceptability of risks. They allow a clearer view an adequate understanding of the communication of the context of the problem presented by risk com- processes in the risk area, an understanding that responds munication, which is essential in the study of domino to its inherent complexity and array of participants. In effects between life support networks. terms of its practical orientation, risk communication seeks to improve the workings of these processes, and so to reduce the level of mistrust among participants. Risk Management The ultimate objective is to assist in the formation of Risk management is a complex process during which the reasonable consensus in contemporary society on decisions concerning risks are made by reconciling how to assess and manage risk" (Leiss 1989). notions of analytic procedures that are necessary for the There are two types of information circulation: within efficient management of risk and the human, legal, a domain or a network and between several domains administrative, political, and organizational dimensions or networks. Furthermore, a model proposed by Leiss of the decision-making process (Covello 1986). Fur- (1989) suggests that the difficulties of communication thermore, efficient risk management corresponds to effi- will be even more considerable once information passes cient allocation of limited resources (Dynes 1994). It is from one domain to another. This process of commu- a question of identification, estimation, evaluation, nication is channeled, constrained, by the nature of risk reduction, and control of risks (Petts 1992), allowing itself, but also by the institutional communication chan- the reinforcement of the ability of those involved to nels already in place. In this way, the circulation of infor- interact in an organized manner during the prevention mation relies on two approaches: structured and and preparation phases. In this way, the goals of risk nonstructured. The structured approach translates management can be to control and reduce the risks to into committees with sitting members being govern- an acceptable level, to reduce the level of uncertainty ment representatives, risk generators, experts, and in the decision-making process regarding risks, and to representatives of the public. The nonstructured approach increase the confidence of the public or other concerned relies on the circulation of information arising from the parties in the decisions taken (Gutteling and Wiegman routine contact between those involved. These routine 1996). contacts are a good starting point to put into motion a A life support network is a risk generator. It can be more complete approach. However, they are not enough considered the starting point of this approach; the infor- for the risk communication approach and must be mation the network possesses confers upon it a major, enriched for a greater efficiency. even essential, role. Moreover, it is faced with an imper- Communication must, therefore, play a role in medi- ative: its mission. In this context, it tries to reconcile its ation that serves the entire group of concerned parties, obligations and constraints with the interests of the where each one gains by integrating into this media- community where its production facilities are located. tion space if it wants its point of view known and if it The identification and the estimation of risks take from, wants to exert some influence on the overall process of withoutcontradiction,thetechnicalandscientificdomains. risk management. This mediation, based on the exchange From a managerial point of view, expertise coming from of information and perceptions, represents the very 248 Building Safer Cities: The Future of Disaster Risk essence of risk communication. The explanation of the (e.g., children) of the threatened population (Covello nature of risks, as well as the preventive and response and others 1994; U.S. NRC 1989; Leiss and Chociolko measures put in place for the concerned public, will be 1994). accomplished in a language understood by all and pre- Therefore, the domain of risk perception depends on viously validated by representatives of concerned par- the social context, the political decision-making process, ties. "In the revealed-preference and risk compendia and the influence of the population on the competent approaches favored by technical experts, a technology authorities that are part of the process of risk manage- is judged to be socially acceptable if the risks of death ment. The public sphere gathers interested parties associated with it do not exceed the risks of death asso- who are normally more reactive when facing risks and ciated with comparable technologies. These approaches, the actions or inactions of governments or risk gener- therefore, yield an absolute number of "acceptable" fatal- ators. The understanding of the public, in the face of ities (a so-called `acceptable risk' level)" (Covello and risks, is generally limited by the perception of these others 1986). risks, resulting from their disturbing nature and from It becomes clear that the formation of committees the way this information is presented. intended to improve the communication and the man- The acceptability of risks relies upon the presump- agement of risk must include technicians and special- tion that there is a probability that an event might not ists capable of judging the scientific and technical data occur. Thus, the population, tacitly or explicitly, accepts (expert sphere), as well as citizens (public sphere) capa- the existence of these risks if they are below a certain ble of rigorously validating the proposed measures. threshold and if the benefits related to their existence "While expert knowledge of many different kinds is exceed the perceived risks (Leiss 1994). It seems that called upon to address the complexity of such risks, the population pays more attention to the qualitative non-experts are also called upon to approve or disap- characteristics of the risk, while the sphere of expert- prove, ultimately through political processes, deci- ise is more concerned with the assessment of the level sions based on accumulated facts and reasoning" of death and disease potentially linked to the risk (Leiss 1989). (Covello 1986). Moreover, the population seems effec- The task of clear communication is further com- tively ready to accept a certain level of risk if they per- pounded by the presence of not one, but many publics, ceive it as being justified or if the risk allows them to characterized by a public mood that fluctuates, a public reach some goal or provide certain advantages (e.g., perception that is inconsistent throughout the popula- generate a job). Alternatively, it will be much more dif- tion, and a public view that is difficult to measure ficult for a risk to be accepted if it seems to be imposed (Middlekauf 1989). Despite these constraints, a dem- or if it is in opposition with certain values. Therefore, ocratic society must find ways to place specialized knowl- the acceptability of risks will be part of the negotiation edge into the service of public choice. Moreover, it must between the interested parties, according to their respec- be perceived to do so (Powell 1996). tive perceptions of risks (Renn 1998). The decisions coming from the competent author- ities will be made while considering the acceptability Acceptability and Perception of Risks and perception of risks by society in an economic, social, Generally speaking, the notion of perception is associ- and political context. To not take into account the ated with the public, media, and special interest groups acceptability and the perception is to risk controversy belonging to the public sphere. It is defined as an intu- and failure. In addition, two types of language are itive judgment regarding the nature and importance that used regarding risk, namely technical language and a risk presents for health. Nevertheless, it can be perceptible language. Competent authorities are strad- defined and characterized by factors that influence it, dled between these two languages that they must master. such as the level of understanding of the risk in ques- Thus, they have an essential role to play in the under- tion, the fear resulting from the expected rate of death standing of the scientific and technical problems of risk, and disease, as well as the size and the characteristics while paying attention to the perceptions of the public A New Structural Approach for the Study of Domino Effects between Life Support Networks 249 sphere (BIT 1990). The competent authorities have water supply network, etc.). These results can be illus- no choice other than to consider the risk in question trated using maps or balance sheets (Quach and others in a social, human, technological, political, and eco- 2000). nomic global perspective (Leiss 1989). Conventional risk analysis mainly focuses on extreme events, of which the probability of occurrence is extremely low. In such a context, the event is considered a Risk Studies: Traditional Approach hazard, which is defined as being an unforeseeable or a disruptive event (United Nations 1984). Within the Risk definitions found in literature are numerous and framework of risk analysis, a hazard can be defined as diverse, and content themselves to rule on the nature a generic class gathering a potential set of causes, or as of risk (Seidou 2002). Among these definitions, that of a source of causes. For example, a hurricane is a hazard Rowe (1997) best summarizes the approach used at that can generate causes like flooding of an underground present. In this approach, risk is defined as being the public transportation network, or strong winds that can potential of incidence of unwanted and negative con- damage power lines. sequences of an event. In technical fields, definitions include a clause allowing the calculation of a value, gen- Hazards erally by multiplying the probability of occurrence of harmful events by the severity of these events, expressed, A risk or a hazard is an event. A cause is defined as being generally, in monetary units. These formulas can apply the origin, the reason, for which an event occurs. It is themselves to sequences of events (scenarios) by sum- generally accepted that causes are of a natural or an ming, for all the events, the products of the consequences anthropic origin. The term "natural" refers to the phys- and the probability (Kaplan 1997; CAN\CSA 1991; ical world, except for humans and their works, and to Stedinger 1996). Even though the use of probabilities what occurs spontaneously in the universe without the to assess risks related to extreme events is currently ques- intervention of calculation, reflection, or will, which are tioned (Zielinski 2001; Robert and others 2002a), the considered the prerogative of humans. The term "an- approach that studies a finite number of scenarios remains thropic" relates to human industry, that which is made largely privileged in a highly linear process (figure 17.1). by humans, that which is due to the existence and the This approach is based on the prior establishment presence of humans. Based on these definitions, it is of boundary conditions. For every modification of the possible to say that an anthropic cause, due to the exis- initial conditions, the process is repeated, thus estab- tence of humans, is the cause that will bring about an lishing as many scenarios (Law and Kelton 1991). Over- event (failure, malfunction) affecting the studied sector. all, this approach consists in first defining the source A natural cause, contrary to an anthropic cause, is not event of the potential risk. The characteristics of the at all influenced by human activities. event make up the source data of the model. This model Today, it is increasingly difficult to define an excep- is established in order to simulate the behavior and the tional event as exclusively of natural origin (IPCC 2001). effects of the event in a specific context (study area, Risks are in interaction with one another (Zielinski 2001; delimited geographical space, etc.) with specific objec- Denis 2002). In this context, no event said to be natu- tives (scale of representation, time progress, etc.). The ral can be completely independent from human activ- results of the simulation show the consequences engen- ities, particularly with regard to extreme weather dered by the event in the study area (city, region, potable phenomena (tornadoes, hurricanes, ice storms, tor- rential precipitations, floods). It is, however, admitted Figure 17.1 Risk scenario: a linear process that a natural event can be considered a trigger, where the human activity and the anthropic risk that result from it are defined as aggravating elements. In addition, Event Model Consequences the anthropic causes are mostly internal to a system, while natural causes are external to it. 250 Building Safer Cities: The Future of Disaster Risk Although risk studies of either of the two types of advantage of offering a global vision of the expected events (natural or anthropic) follow the same generic impacts of the hazards in a given territory. However, the approach, it appears important to make a distinction approach generally followed during these evaluations in their specific treatment. Effectively, this distinction does not integrate, a priori, the concerns of the man- allows the differentiation of hazards coming from out- agers of the life support networks located in the studied side a life support network from those propagated inside territory (figure 17.2) this same network. Therefore, for those in charge of the networks (par- ticipants B and C, figure 17.2), the risk evaluation for their infrastructure requires a specific approach for the Natural Events interpretation of the generic results obtained by par- Particularities of exceptional natural events lie in their ticipant A. This process of nonintegrated sequential low predictability. Therefore, the evaluation of natural causes is essentially carried out using quantitative and Figure 17.2 Evaluation of the impacts of a natural predictive methods based on statistical analyses and sto- hazard and use of the results chastic modeling (IPCC 2001; Zielinski 2001; Bier and others 1999). In this case, the establishment of a model Participant A draws, on the one hand, on physical laws that govern the components of the event (e.g., hydraulic laws in the case of a flood) and, on the other hand, on the analysis Natural of prior recorded events in order to validate the model. hazards Historical, morphological, and cartographic studies as causes well as studies more specific to the studied phenome- non (seismic risk, floods, etc.) are fundamental in order to collect data necessary to the proper operation of sto- chastic models. It is a matter of better understanding Model phenomena in order to better predict them. Consequently, it is possible to simulate a multitude of synthetic events (scenarios). The results of these sto- Generic chastic models are generally illustrated as maps, graphs, results and tables (RNC 2002). They allow reporting of the effects of the event upon the studied geographical site. As an example, table 17.1 shows three cases of the evaluation of the consequences of natural causes. Participant B Participant C These generic results are generally obtained in the form of maps that integrate, for example, water levels, ... Processing of the generic results tornado corridors, and wind speeds or even seismic according to their fields of activity ... zones. This way of expressing the results has the Table 17.1 Cases of evaluation of consequences of natural events Official Application Model(s) Reference U.S. Geological Survey Earthquakes in the Mississippi Embayment http://www.ceri.memphis.edu/usgs/model/ United States reference model index.shtml U.S. Federal Emergency National Flood Insurance Numerical models accepted http://www.fema.gov/mit/tsd/en_modl.htm Management Agency Program (NFIP) National Hurricane Tropical cyclone track and Operational track guidance http://www.nhc.noaa.gov/aboutmodels.html Center intensity guidance models models A New Structural Approach for the Study of Domino Effects between Life Support Networks 251 steps highlights two main limitations. The first is related Figure 17.3 Summary of the psychological varieties of to the single-sequence approach, generally observed by unsafe acts authors of natural hazard studies (participant A). The BASIC ERROR Attentional failures second limitation is related to the need of those in charge TYPES Intrusion Omission of life support networks (participants B and C) to eval- SLIP Reversal uate not only the most unfavorable scenario, but also Misordering UNINTENDED Mistiming the set of potential situations with which they can be ACTION Memory failures confronted, in order to be able to understand the dynamic LAPSE Omitting planned items Place-losing behavior of their network and thus anticipate some mit- UNSAFE Forgetting intentions igation or emergency measures. ACTS Rule-based mistakes MISTAKE Misapplication of good rule Application of bad rule INTENDED Knowledge-based ACTION mistakes Anthropic Events Many variable forms Anthropic events refer to two main categories: techni- VIOLATION Routine violations Exceptional violations cal malfunctions and human reliability. If the first cat- Acts of sabotage egory is relatively well known and integrated into Source: Reason 1990. analyses, the second one is in many regards minimized in risk studies. border between violations and faults is often difficult It is, however, important to differentiate between reli- to determine." ability and human error (Nicolet and Celier 1985). A second approach consists of differentiating latent Human reliability is the probability that an individual, malfunctions from active malfunctions (Reason 2000). a team, or a human organization, will accomplish a mis- Active malfunctions have almost an instant effect and sion, under given conditions, within acceptable limits, are often related to the operators of the system being within a certain length of time. A human error is a behav- considered. Latent malfunctions, as their name indi- ior that exceeds acceptable limits (variation between cates, are more insidious since they can go by unno- what has been done, perceived, understood and what ticed until an event or a combination of events makes should have been done). These two definitions demon- them active. This last type of malfunction is related to strate that the study of human errors takes place after individuals, such as managers, who are remote from the the event occurs. In contrast, studies of human relia- control interface (Reason 2000). Due to their nature, bility take place before the event takes place. latent malfunctions, which are more likely to affect a The study of human reliability is complex since it system, are more difficult to assess than active errors. can be approached in two different ways. A first approach Therefore, many studies tend only to consider active consists of defining human reliability as the gathering failures in their evaluation methods. This differentia- of human errors, violations, and positive actions that tion between types of malfunctions presupposes that could be carried out by participants. Human errors arise only a series of latent and active human malfunctions from intended actions (mistakes) and unintended actions can lead to the failure of the system. (slips, lapses). As shown in figure 17.3, drawn from When it becomes a matter of integrating anthropic Reason (1990), a slip corresponds to a failure of atten- events as causes of potential risk, the approach proves tion, while lapses are failures of memory. Mistakes to be complex. However, these events can be taken, refer to the incorrect application of a good rule or to not only as a direct cause, but also as a potential tech- the correct application of a bad rule. Violations, accord- nical cause. Technical malfunctions are principally found ing to Reason (1990), "...can be defined as deliberate, at the design, construction, operation, and maintenance but not necessarily reprehensible, deviations from those stages. These malfunctions are related to human factors. practices deemed necessary to maintain the safe oper- It is not easy to differentiate between the two, and for ation of a potentially hazardous system. Therefore, the this reason, the notion of sociotechnological risk is often 252 Building Safer Cities: The Future of Disaster Risk employed (Denis 1998). Even in the case of a mechan- uncertainty that exits in the area of the study of extreme ical breakdown, human intervention cannot be disre- events (data and models). Presently, the form under which garded, because it may be the result of an error in design risk study results are presented makes it difficult to take or maintenance. Therefore, the theoretical approach into account this fundamental dimension of uncertainty. normally relates human causes with technological causes. For example, analysis of the risk of dam rupture (tech- In this category, there exists a multitude of tools, nological event potentially engendered by a natural hazard) which are more or less specialized and dedicated to allows the establishment of a flood map for the regions activities such as design, operation, maintenance, and downstream of the infrastructure as a function of a fixed management of a network. In particular, certain models scenario.However,takingintoaccountsourcesofexpressed permit the simulation of a technological risk (for uncertainty, the results (water levels) do not allow the example, breakdown of a transformer station in a power managers of essential infrastructure, potentially affected grid) and the measurement of the behavior of the net- by the flood, to grasp the entire set of potential effects of work faced with the risk. The organizations responsi- the event on the components of their infrastructure in ble for civil infrastructure generally have at their order to evaluate the consequences on the different mis- disposition these models, which allow the simulation sions of the essential infrastructure. For example, the mis- and operation of their networks. sion of a potable water network is to deliver a volume of water, of sufficient quality, with a minimum pressure. Analysis, a posteriori, of so-called natural catastro- Limitations of Current Risk Studies phes shows, more often than not, that extreme events, Profiting from the never-ending increase in perform- which on one hand are difficult to foresee according to ance of computers, these software tools allow a large the models being used and the previously defined bound- number of parameters and variables to be taken into ary conditions (recovery time, initial conditions, etc.), account. Nevertheless, the limitations of these models can combine with other events (technological risks, have less to do with the computing power than with etc.), which amplify the effects and the consequences fundamental understanding of natural phenomena, such (Lavallée 2000). Scenarios consider, in general, single- as meteorological events (hurricanes, tornadoes, or event conditions (simple hazards). This characteristic cyclones). Models of natural phenomena collide with is implicitly required for the statistical treatment of the complexity of the interactions, between environ- previous events. So studies of catastrophes highlight the mental elements as much as between technological synergetic effect that can exist between generators of elements. At present, models are mathematical relations natural, technological, and human risks (Denis 2002). simplified from the rules of behavior of natural phe- For example, a hurricane engenders not only direct con- nomena. In effect, the majority of models rely on a com- sequences on populations, but also on essential infra- bination of empirical and theoretical relations (Tyagi structure (flood of an underground public transportation 2001), which by their very nature imply incertitude as network, felled telecommunication relays or electric much for the input values as for the algorithm of the power distribution pylons, etc.). Out of service or mal- model being used. Therefore, these sources of uncer- functioning parts of the infrastructure, by definition, tainty are reflected in the results (output). have negative effects on one another and finally on the In addition to the incertitude engendered by the functioning of civil society. This statement and the result- ignorance of the behavioral mechanisms and interrela- ing consequences are repeated more and more often, tions between different components, conditions of a particularly in industrialized societies with essential scenario are defined by historical data and the statistical infrastructures that are increasingly complex and analysis of this data. However, certain authors (Bier and interrelated. However, even if there are differences others 1999) highlight the specific character of extreme between societies in industrialized countries and devel- events. They define them as events that are at once oping countries, the development of life support net- severe and outside the normal sequence of occurrence works always tends toward an increased complexity and of the system in question. This statement highlights the interdependence. A New Structural Approach for the Study of Domino Effects between Life Support Networks 253 To be able to treat overall the multiple and complex results has been maintained by removing all informa- relations that encompass tangible and intangible param- tion that could identify the zones studied. eters, one of the advocated approaches is to refocus the analysis on essential infrastructure. In effect, simulation of natural hazards does not integrate, a priori, the expec- Characterization of a Life Support Network tations of the managers of these infrastructures. In this context, the reference should no longer be the causal Fundamentally, a life support network is created in order event, but the infrastructure itself, which, if it does not to fulfill certain missions or functions directly related fulfill its mission completely or in part, engenders con- to human activity. To do this, certain operations are sequences on civil society, and other infrastructure. From available to managers of this network. In terms of engi- here, research of potential causes that generate devia- neering, a network is composed of a set of important tion from the missions is accomplished with respect to infrastructure components, of which some are essen- the preoccupations of the managers, regardless of the tial in terms of vulnerability of the network. A network type of risk: natural, technological, or human. In this fails once one mission cannot be fulfilled in whole or context, study of the domino effects between life sup- in part, following the failure of one or several opera- port networks is clearly necessary in order to be able tions or of infrastructure. Therefore, there is a notion to evaluate the risk and to contemplate emergency meas- of degree of efficiency of a mission. The description of ures in order to respond adequately to catastrophes. the three principal elements--mission, operation, and essential infrastructure--follows below. Mission: the mission of a life support network cor- New Structural Approach responds to a function for which it was designed and built. The classical approach, which consists of analyzing a life As an example, the mission of a potable water supply support network from a finite number of potential fail- network is to: ure causes, does not allow consideration of the entire · maintain a quality of water suitable for human set of situations. In addition, the domino effects between consumption; networks are rarely identified and considered. With · provide water (volume and pressure) in order to regard to the very nature of life support networks, which fight fires; have human and socioeconomic consequences result- · provide a volume of water; ing from their failure, it is dangerous not to have a com- · etc. plete picture of the possible situations. To respond to The mission of an electrical network is to: this disquieting situation, a new approach is proposed · maintain constant electrical power; to carry out an exhaustive study of life support networks. · maintain a minimum electrical power; This analysis will act as a catalyst for the study of the · provide at least some organizations (hospitals...) domino effect between several life support networks. with electrical power; The proposed approach relies on a precise character- · etc. ization of the functions of a life support network, its Operation: an operation is a technical process modes of operation, and the infrastructure that composes allowing direct or indirect actions on part or all of it. The analysis of a life support network is based on two the network in order to accomplish the mission; these study methodologies. The first is an evaluation of the actions can be automated or manual. consequences of failure of the network, while the second As an example, the operations of a potable water is a determination of its vulnerabilities. All of this infor- supply network include: mation will serve the study of the domino effects between · pumping; several life support networks. Case studies carried out · treatment; by the École Polytechnique de Montréal will illustrate · storage; the proposed methodologies; the confidentiality of these · etc. 254 Building Safer Cities: The Future of Disaster Risk The operations of an electrical network include: If an element cannot fulfill all of the tasks, its efficiency · energy generation; decreases. When an element is out of service, its effi- · transportation; ciency becomes zero. Therefore, efficiency can vary · distribution; between 100 percent and zero. This efficiency is · etc. evaluated differently depending on the elements: · Essential infrastructure: infrastructure corresponds · Mission: a mission is 100 percent efficient if all of to installations that are necessary to accomplish the the conditions that characterize it are present and operations. Infrastructure is essential when it has a the set of clients affected by the mission is served. primary role in the proper functioning of the network. · Operation: an operation is 100 percent efficient if As an example, the essential infrastructure of a potable the management procedures are valid, adequately water supply network is: applied, and known by all of the personnel. · ozonator; · Essential infrastructure: an essential infrastructure · reservoir; is 100 percent efficient if it is in good condition, · pump; functions correctly, is used within the design norms, · etc. was designed according to the existing norms, etc. The essential infrastructure of an electrical network Importance: characterizes how an element is essen- is: tial for a network · transformer station; · Mission: a mission is essential if it represents a · hydroelectric generating station; priority that whose fulfillment is imperative; some · high voltage power lines; missions can be secondary. · etc. · Operation: rules of operation are essential if they To fulfill a mission, several operations are available; an ensure a minimum level of service, while some operation uses a certain amount of essential infrastructure. rules of operation can be identified as refine- Figure 17.4 shows a schematic of these elements and ments or can be used to ensure an optimization their potential interrelations. For example, the essen- of operations. tial infrastructure (E.I. 1) allows the network to fulfill · Essential structure: the importance of an essential its mission (M. 2) using the operations (O. 1 and O. 3). structure is a direct measure of its role in the net- These operations depend on the specificity of each work. In the case of a failure, or cessation of serv- life support network and must be established with the ices, the consequences on the correct functioning managers, design engineers, and the technical and oper- of the network and the accomplishment of the ations managers. The correct identification of these missions are significant. elements is essential to defining their efficiency and their Thus, a network is functional if the entire set of its importance. These notions can be defined as follows. missions is fulfilled with a degree of efficiency of 100 Efficiency: an element is 100 percent efficient when percent. Once a mission is no longer efficient, the net- it adequately fulfills the task for which it was designed. work fails. A failure results from a vulnerability of the network, in other words, once an operation and/or an essential infrastructure is no longer efficient. The size Figure 17.4 Diagram of the characterization of a life support network of the vulnerability depends on the importance to the mission of the nonefficient elements. Essential infrastructure There are two distinct notions, in terms of analysis, E.I. 1 E.I. 2 E.I. 3 ... E.I. i O. 1 Operations related to correct functioning of the network: failure O. 2 O. 3 and vulnerability. Failures of a network provoke con- ... O. i sequences on populations and socioeconomic activities M. 1 and will, therefore, be studied from the angle of the M. 2 M. 3 Missions study of consequences. Vulnerability of a network will ... M. i be analyzed by way of the study of vulnerabilities, which A New Structural Approach for the Study of Domino Effects between Life Support Networks 255 allows the affected elements to be established. These allow evaluation of the resources to establish those two studies are described in the following pages. that are available and those that must be sought. 3. Failure study related to the variation of efficiency of a mission Consequence Studies Once a mission is not respected, the potential fail- Analysis of current risk studies shows that the approach ures are numerous. For example, the mission of a by scenarios is too restrictive for studying life support potable water supply network can be inefficient fol- networks. In effect, these scenarios generally advocate lowing multiple failures, such as chemical, biologi- a macrosequential approach, based on the evaluation cal, or bacteriological contamination, etc. The of past events, composed of natural or technological identification and characterization of these failures risks, where human reliability is rarely integrated. Finally, are important in order to evaluate the potential con- they consider extreme events to the detriment of the sequences properly. Failures can be grouped as a func- intermediate situations that affect a set of well-defined tion of consequences. Equally, failures can be described infrastructure, usually coming from one network. These in terms of variation between no failure and maxi- studies bring about a generalization of knowledge, mum failure, with respect to the variation of the effi- whereas only a part of the entire set of possible situa- ciency of the mission. tions has been analyzed. It follows that there is an appar- 4. Exhaustive evaluation of the consequences ent gap in the evaluation of vulnerabilities of a network A study of consequences is focused on the com- and the domino effects resulting from the multiple fail- plete analysis of a failure and the set of resulting ures that can arise. To alleviate these major weak- consequences. The entire set of possible and plausi- nesses, during risk studies and the planning of emergency ble situations is studied while considering all of the measures, a new approach is proposed: the study of graduations of the failure. Such an exhaustive approach consequences (Robert 2001b, 2002b). allows a progressive picture of the repercussions or A study of consequences is based on the following the consequences generated by the occurrence of a fundamental principles: failure to be established. For example, following the · A life support network fails when one of the missions failure of a hydroelectric reservoir, a flood is cre- for which it was designed is no longer fulfilled with ated. A study of the consequences covers all the water an efficiency of 100 percent. levels between the start of the flood and the maxi- · The reduction of efficiency of the mission of a life sup- mum water level. The progression of the flood is port network is studied without consideration for therefore studied as well as the increase in the the causes that could have provoked the reduction. resulting consequences. · All of the degrees of efficiency of a mission are ana- 5. Drawing of consequence curves lyzed from 100 percent to zero, in a manner that The results of a study of consequences are pre- covers all of the possible situations. sented in the form of curves or diagrams in order to From the preceding principles, a study of consequences show the progression of the phenomena studied. of a life support network is carried from the six princi- 6. Identification of life support networks affected by pal steps, described generally below: the failures 1. Evaluation of variations in efficiency of missions First, life support networks affected by the failure of Each mission of a life support network is analyzed a network are identified. The conditions of failure of the in technical terms in order to identify experts who first network are also identified in order to establish should be involved in the project to characterize the the initial conditions provoking the domino effects. failures and evaluate the consequences. Communication mechanisms between these networks 2. Sequential evaluation of missions according to are initiated in accordance with the concepts of risk their importance communication. Prioritization of the study of consequences as a Actually, the consequence curves are established to function of the their importance. This step will cover the human socioeconomic consequences of the 256 Building Safer Cities: The Future of Disaster Risk failure of a life support network. These curves were based on specific events. In effect, such risk studies allow established principally for the municipal emergency a result representing a precise state of vulnerability, which measures managers directly affected by consequences corresponds with a point on the consequence curve, to for their citizens and municipal affairs. Some examples be obtained. For the different planning stages of emer- of produced curves: gency measures, these graphics allow the entire set of · Number of people affected by the failure (flood and potential situations to be covered and establish thresh- water contamination) olds that correspond to the emergencies. · Damage to buildings (flood) · Number of sensitive elements (hospitals, schools, daycare centers, etc.) affected by the failure (flood, Vulnerability Studies contamination, loss of electricity) · Dangerous industries affected by the failure (flood, The failure of a network is a unique event generally loss of electricity) resulting from the conjunction of several distinct causal · Public transportation affected by the failure (flood, chains. Therefore, the proposed approach consists of loss of electricity). studying the potential failures of the missions of a net- Figure 17.5 presents an example of a consequence work in order to define their causes, both natural and curve for a municipality expressed in terms of number anthropic. The goal is not to determine a probability of of people affected by a flood. Analysis of such a curve, occurrence of an event, which often proves to be sub- in addition to providing all the knowledge for a good jective, but to define the series of hazards that can lead plan of the mechanisms of evacuation, allows the estab- to the failure of the network. For example, for the bac- lishment of the levels of severity of the emergency gen- teriological contamination of a water supply network, erated by these consequences (Robert 2001b). it is a matter of determining the contamination points These graphical results, coming from a progressive of the network while taking into account the operating and systematic approach, are easily used in a risk study procedures, the condition of the network, and the pos- sible contamination of groundwater by natural phe- nomena or technological accidents. Figure 17.5 Consequence curve for a municipality First, the methodology of the study of vulnerabili- ties is presented overall. It is based on the identifica- 350 tion of the vulnerabilities of a life support network. Affected population Explained next are the general concepts of evaluation 300 of the possibility of the vulnerabilities, since these data are essentially for the study of risk. 250 Methodology 200 individuals Network engineers must determine the conditions nec- of 150 essary to reach a certain state, taking into account that external risks can arise and that components of the Number network (operations and essential infrastructure) can 100 become inefficient. Figure 17.6 diagrams these condi- tions, which are described below. 50 External hazards: act as triggers by affecting the net- work at the level of essential infrastructure. Three 0 causes of external hazards have been identified. 30 32.25 34.5 38.75 39 · Natural causes: subdivided into two groups: Water surface elevation (meters) ° Instantaneous phenomena, such as earthquakes A New Structural Approach for the Study of Domino Effects between Life Support Networks 257 Figure 17.6 Diagram of the conditions for decreasing the efficiency of a mission Mission efficiency Decreasing Active human failures Chain of internal risks -Errors Mission Latent human failures -Management External risks Technical Natural event failure -Design -Maintenance Transfer of -Breakdown Operations vulnerability Malevolence Essential infrastructure ° Predictable phenomena, such as the seasonal · External human causes: are exclusively active human surge of a river. failures such as acts of malevolence (sabotage). In all cases, the natural events studied are excep- Decreasing efficiency of the components of a net- tional phenomena. work: can only have two origins, either technical or · External technical malfunctions: correspond to the human. transfer of vulnerability between life support net- The methodology of evaluation of overall vulnerability works following the domino effects between them. is presented because it requires highly multidisciplinary 258 Building Safer Cities: The Future of Disaster Risk work, with a whole set of technical, managerial and Figure 17.7 Vulnerability curve for a transformer station human resource participants. The concept of risk com- munication will serve to identify the participants and Efficiency of the infrastructure the precise structure of the steps to follow as a function 100% 95% 80% 75% 0% Accessibility Out of of the specifics of each life support network. Three main Limit service steps can be identified. 1. Evaluation of vulnerabilities Normal operation Failure zone zone The analysis of consequence curves allows the identification of the missions where efficiency has 23 24 25 26 27 28 29 30 decreased. Therefore, it is a matter of determining Water level at the station (meters) the causes bringing about the decrease in efficiency. To do this, each element linked to the inefficient mis- · An evaluation, numerical or not, of the possi- sion is analyzed. bility of occurrence of these conditions. · First, the operations are studied. It is a matter of Next, an overall analysis is carried out in order to evaluating whether human error can diminish the obtain a complete picture of failure possibilities. The efficiency of an operation. For this, how and where variation in efficiency of a mission is analyzed to know human interaction takes place must be identi- the progression of the magnitude of the failures. fied. Next, management procedures must be ana- 2. Creation of vulnerability curves lyzed in terms of response to an exceptional The results of the evaluation of vulnerabilities can be situation. presented in the form of curves or diagrams that illus- · Essential infrastructure is then analyzed in terms trate the progression of the phenomena. These curves of internal efficiency. To do this, it is a matter of are associated with procedures of operation and essen- evaluating whether the actual condition of an infra- tial infrastructure. They indicate variation in the degree structure is satisfactory, then establishing the poten- of efficiency. Figure 17.7 gives an example of a vulner- tial conditions for diminished efficiency, for example ability curve of a transformer station, essential infra- the minimum level of maintenance necessary to structure of a hydroelectric network, having to control affect an infrastructure. floods as a mission (Robert 2001b). · External hazards are then studied. It is a matter 3. Identification of domino effects between several of establishing the natural events that can arise life support networks and affect the efficiency of the infrastructure, while Consequence studies highlight the life support net- assuring that the entire set of possible situations works affected by failures of a first network. At this stage, is covered. it is a matter of identifying the essential infrastructure, Technological causes coming from the failure of which has links with other life support networks, and another life support network are analyzed from the evaluating at which threshold the efficiency of the first failure curves of the network and from the trans- network affects the second. fer function as described in the section on domino effects. This transfer of vulnerability allows direct evaluation of a decrease in efficiency of infrastructure. Possibility of Occurrence of Vulnerabilities Acts of sabotage must be considered possible, To carry out a risk evaluation, it is necessary to evalu- requiring an evaluation of the protection of the ate the possibility of the occurrence of a hazard. Numer- infrastructure. ous techniques exist for evaluating overall reliability The result is an exhaustive evaluation of the ele- (human and technical) (Dhillon 1985, 1989) of a system ments related to a mission, with the following infor- as well as the effects of natural events on it. Based on mation for each one: the classification proposed by the Institut de Sûreté de · A description of the conditions bringing about Fonctionnement (ISDF 1994), it is possible to define a decrease in efficiency; and five major classes of methods: A New Structural Approach for the Study of Domino Effects between Life Support Networks 259 · Quantitative and predictive methods (Kirwan 1994; The objective of systematic design analysis methods El Shahhat and others 1995); is to consider an entire design project as much from a · Descriptive analysis methods (Hadipriono and others technical as an organizational plan. These methods take 1986; Leverenz and others 1996); both human and technical factors into account at once. · Combination approaches (Suokas 1993; Ridley and It is important to note that several methods can obvi- Andrews 2000); ously be combined according to the situation and desired · Structural and descriptive analysis methods (NASA result, as long as they integrate into a common approach. 1995; Modarres and others 1999); and The classical approaches to risk analysis advocate · Systematic design analysis methods (Garin 1994; probabilistic evaluations of risk, which essentially rely Cazaubon and others 1997). on quantitative and predictive methods. These meth- The objective of quantitative and predictive meth- ods rest on the use of elaborate scenarios built from data ods is the evaluation of the probability of occurrence collected during observed decreases in efficiency of net- of human errors, technical failures, and natural events works. The proposed approach distinguishes itself from that are observable under the given conditions. The the scenario approach by addressing the problem with majority of these methods call upon the opinions of the study of consequences of the decrease of efficiency experts, which allows the various aspects--both natu- of a mission. Defining risks remains important, notably ral and anthropic--that can bring on a failure to be con- for the design criteria and mitigation measures, with- sidered. However, from these methods, probabilities out associating probabilities, which often prove to be that are often subjective result, since they are strongly subjective. correlated to the opinions of the experts who partici- Different risks are strongly dependent upon one pated in working them out and to the variables used. another. Therefore, it appears necessary to define an The objective of descriptive analysis methods is to overall methodology integrating all of these causes to tackle the origin of the decrease in efficiency of a mis- better predict and manage the risks that can affect the sion and provide reliable collection and failure analy- network being studied. This approach will have to com- sis tools. These methods are used once the event has bine the evaluation methods of three large risks. Cer- arisen, which is at the same time their strength and tain methods integrate two of the risks (natural and their weakness. In effect, they allow the hazards technological, technological and human), but none inte- bringing about a decrease in efficiency to be described, grating the three seem to have been developed. Another but they cannot serve to evaluate the network before difficulty lies in the integration of the evaluation of dif- the failure. ferent simultaneous natural phenomena, like for The objective of combination approaches is to describe example, the probability of occurrence of an earthquake and analyze in order to anticipate and even predict and a flood. Therefore, it must be determined, among failures. These methods, such as cause/consequence the large group of previously presented methods, which trees, are very interesting. Starting from a potential con- technique or group of techniques would be the most sequence, it is possible to rebuild the series of hazards appropriate for evaluating the set of risks, both anthropic that have or could bring about a decrease in efficiency and natural, that could lead to the failure of a life sup- of a mission. port network. While allowing the combination of dif- The goal of structural and descriptive analysis is to ferent factors, the chosen evaluation methods must be identify the structural characteristics of a network, in easily understood, quickly put into place, and adapt- other words, its boundaries, components, and inter- able to networks of different natures. Methods with a connections. Because of their mode of representation low degree of subjectivity will have to be found, from in block form, the use of such methods can be prob- which mitigation measures can be implemented. lematic, even impossible, for complex systems such as An overall methodology for risk evaluation will life support networks. However, these methods are inter- have to be based on a good understanding of the net- esting for defining, in an efficient manner, the entire work being studied, its vulnerabilities, its environment, network and its interrelations with other systems. impacts that could affect it in case of decreased efficiency 260 Building Safer Cities: The Future of Disaster Risk of another network, and equally the infrastructure that of repercussion of vulnerabilities. These links are direc- it could influence in case of a failure (vulnerability trans- tional, defined by a source and a destination. The fer). The evaluation of events that can engender risk source is associated with an essential infrastructure spe- must be carried out in a systematic manner, not to elim- cific to the originating network, whereas the destina- inate risk, but to diminish the vulnerability of the set tion is associated with an essential infrastructure of the of life support networks. destination network. Two types of links can be identi- fied: (1) a direct link if the two components are physi- cally or mechanically connected (like an electric cable); Study of Domino Effects between Life Support and (2) an indirect link if there is no clearly identifiable Networks connection. The type of link directly influences the reper- cussion of vulnerabilities function. However, the method- Consequence and vulnerability studies highlight that ology for the study of these links is the same. It uses the the failure of a life support network can provoke con- following three steps. sequences on another network by affecting its essential 1. Definition of the source vulnerability infrastructure (external technological risk). The links The source vulnerability represents a combination uniting several life support networks are numerous and of consequence and vulnerability studies carried varied. A good understanding of the dynamic of these out on the source life support network. The conse- networks is therefore essential to avoid benign mal- quences of multiple failures of the network are functions transforming into major crises (Allen 1997; identified as well as the other life support networks Hubbert and Ledoux 1999; Lemperiere 1999). Figure affected by way of the links between these net- 17.8 illustrates this principle of links. works. A methodology was developed at the École Poly- The study of vulnerabilities of the source net- technique de Montréal to study the links that united sev- work allows the essential infrastructure where the eral life support networks (Robert 2001a, 2002b). It is efficiency has decreased to be identified. If this based on the idea that these links, which represent a con- infrastructure is the source of a direct or indirect link, nection between different networks, allow vulnerabili- there is a potential domino effect. ties of one network to reflect on another through a function The set of degrees of efficiency of the essential infra- structure must be analyzed and passed on to the des- Figure 17.8 Definition of links tination network. Equally, it is important to integrate the consequences of these variations of efficiency. A Source Destination graphical representation illustrates this graduation network network of the vulnerability of a network. Figure 17.9 shows Essential Direct link Essential infrastructure infrastructure an example of a vulnerability curve for a hydroelec- I.E. 1 I.E. 1 tric installation. On this curve, the different compo- Essential Essential nents are identified and characterized in terms of infrastructure infrastructure technical efficiency. In addition, the multiple conse- I.E. 2 I.E. 2 quences of these failures are represented. Such a curve ... ... synthesizes all of the information that must be com- municated to the other networks affected by these Essential Indirect link Essential infrastructure infrastructure failures. I.E. i I.E. I 2. Definition of the repercussion function The consequences of the variations in efficiency Repercussion function of of the source essential infrastructure must be propa- vulnerabilities Vulnerability Vulnerability gated, modified, or transformed before reaching the source destination destination component. The repercussion functions Example of a link are generally established by simulations, which attempt A New Structural Approach for the Study of Domino Effects between Life Support Networks 261 Figure 17.9 Affected components as a function of the water level rise in the upstream storage basin 49.5 49.00 49 + 49.00 48.5 48.50 48.16 m 48.16 m (meters) 48 47.70 47.70 47.80 + 47.85 level 47.50 m 47.50 47.50 m 47.5 + voir 47.13 + 47.00 47.40 47.28 47.30 eserr 47 46.94 m 46.94 m 47.00 eam Failure 46.5 46.60 + Overtopping Upstr 46.50 Dam break 46 Critical point - high risk 45.5 Very high risk Loss of control 45 5 ee 2 ee ee ee 5 ee ee 2 e ee and degr degr bank degr intake degr and /weir shor 2, Dam Dam Spillway 2, e degr e degr e degr 1, /left 1, Canal north # 1/minor # 2/minor bank, Canal # 1/serious /right # 2/serious y sever 4, y sever # 1/ver # 2/sever Dam # 2/ver Embankments Canal Embankments Municipality Municipality Municipality Municipality Municipality Municipality Municipality to evaluate the consequences--direct or indirect-- flood. Therefore it is important to know the of variations in the efficiency of the infrastructure flooded areas. Graphically, this function is on the environment. The type of link, direct or shown in figure 17.10, which shows a curve indirect, influences these studies. relating water levels between a reservoir a. Direct link: the repercussion function depends and a transformer station. on the physical nature of the link. For elec- 3. Definition of the destination vulnerability trical links, the function can be a variation in The source vulnerability transferred to the essen- power. In general, these functions are well doc- tial infrastructure of the destination network is con- umented, numerically modeled, and known sidered an external technological risk that is going to by the network managers and engineers. influence its efficiency as described in vulnerability b. Indirect link: the definition of a repercussion studies. Figure 17.7 shows the result obtained on a function for an indirect link is more com- transformer station. plex in terms of technical expertise, because The study of domino effects between several life sup- it requires specific studies and calculations, port networks demands the collaboration of multiple which are not regularly carried out by net- experts and requires complex studies, especially for work managers and engineers. For example, indirect links. It is a multidisciplinary work, which will for a hydroelectric installation, source vul- have to be carried out so that the entire set of human nerability is shown in figure 17.9. The reper- and socioeconomic consequences can be considered cussion function of this vulnerability in the establishment of efficient and effective emer- characterizes the behavior of water that over- gency measures and the performance of complete risk flows a reservoir and affects the infrastructure studies. The diverse uses of these studies are pre- of the other network affected by the same sented below. 262 Building Safer Cities: The Future of Disaster Risk Figure 17.10 Example of a repercussion function: these measures permits them to be integrated into financ- relationship between the water level in a storage ing mechanisms and therefore ensures their existence basin and at a transformer substation and efficiency. 30 29.5 Risk Studies: Scenario Approach 29 (meters) 28.5 A risk study based on the consideration of a finite number the 28 at of scenarios does not allow the entire set of possible 27.5 level situations to be considered, but maintains a certain util- 27 substation ity, especially in establishing the data necessary for design ater 26.5 W 26 and rehabilitation. Equally, it can serve to evaluate a set 25.5 of similar infrastructure in a given territory. This type transformer 25 of risk study is widely used for hydroelectric installa- 24.5 46.5 46.8 47 47.3 47.5 47.8 48 48.3 48.5 48.8 tions (Plate 1996; Lemperiere 1993; Quach and others Water level in the storage basin (meters) 2000). The result of such a risk study is generally rep- resented by the product of the probability of occurrence of a scenario and the consequences, as defined at the Applications and Uses beginning of this report. The proposed vulnerability studies coupled with The results of these studies are particularly well suited the precise definition of a scenario will permit the prob- for risk studies and planning of emergency measures. ability of a scenario arising to be established with For risk studies, two methods are possible: an approach much greater precision. In effect, the scenarios studied by scenario and an approach by consequences. The can be situated among a set of potential situations, which choice of the method depends on the desired use of is a plus when the sum of the probabilities must the results. For emergency measures, the proposed stud- remain equal to 1. Once the scenarios have been stud- ies are particularly useful for the planning and identi- ied and simulated, the potential consequences gener- fication of mitigation measures. A section specific to risk ated by the scenarios will have been exhaustively studied communication will be presented, since it acts as the by consequence studies, including the domino effects basis for the previous activities by ensuring the involve- on the other life support networks. Therefore, the uncer- ment of all concerned parties. The diverse uses of these tainties of risk calculations will be minimized. studies will be presented below. The applications described below allow the evalua- Risk Studies: Consequences Approach tion of risks related to life support networks facing nat- ural hazards. In the context of financing these networks, A way of evaluating the overall risk related to a life sup- the quality of the information obtained from conse- port network is to study the entire set of potential situ- quence and vulnerability studies will allow a better eval- ations. In the face of infinite possible solutions, a finite uation of the opportunity to develop these networks number of conditions can be analyzed to trace a risk with regard to consequences during a failure. It will also curve that corresponds to the variation in efficiency of be possible to evaluate whether the structures and the a mission. To determine the study points for this risk networks, as planned, are adequate in terms of risks and curve it is a matter of analyzing the consequence curves especially in terms of consequences for populations. and establishing different levels of progression. Each Alternative solutions can therefore be proposed and level corresponds to a significant and abrupt increase in developed. Proactive solutions for the protection of pop- the consequences, therefore a network failure and a ulations can also be directly associated with the financ- decrease in the efficiency of a mission. Vulnerability stud- ing of networks by way of imposed emergency measures ies allow the vulnerabilities of the network that pro- centered on the network missions and the mecha- voke the decrease in efficiency of a mission to be nisms of risk communication. The initial planning of determined. They identify the concerned elements at the A New Structural Approach for the Study of Domino Effects between Life Support Networks 263 Figure 17.11 Schematization of a risk curve neglect a potential situation. It is one of the principal drivers of the study of consequences. Concerning the establishment of mitigation meas- ures, a combination of consequence and vulnerability studies is necessary. Similar to the risk studies done according to the consequences approach, levels of esentation consequence must be established. These levels are deter- eprr mined from an evaluation of consequences that are unac- Risk ceptable for the population. Next, vulnerability studies allow vulnerable elements at the operations level and the essential infrastructure to be identified. It is, there- fore, possible to determine the corrective measures to M. 1 M. 2 M. 1 M. 1 M. 2 M. 2 M. 1 M. 2 M. 2 apply--whether they are structural or not, permanent (90) (85) (50) (20) (60) (45) (0) (15) (0) Efficiency of the missions (percent) or not--as much on the essential infrastructure level as for the operational procedures, all the while consid- ering external risk. operations level and the essential infrastructure. There- fore, it is possible to evaluate the plausibility of occur- rence of these causes and to calculate a risk in terms of Risk Communication the couple of plausibility of occurrence and consequences Several studies have attempted to prove that the man- (this result can be purely numerical or can have a sym- agement of risks, as well as the process of risk com- bolic value). This risk calculation is carried out for each munication that accompanies it, reduces uncertainty mission of a network. A set of risk values is therefore and increases the acceptability of risks. According to a associated with decreases in efficiency of missions and systematic approach, risk communication allows the the results can be expressed, if necessary, as a graph. risks to personnel, populations, and material, indus- Figure 17.11 shows a risk curve established for two mis- trial, technological, and financial resources to be reduced sions (M. 1 and M. 2). (Bates and Fitzpatrick 1994). Such a risk study allows the maximum risk of a net- In the proposed structure of life support networks, work to be determined, and also the total risk, which as soon as the domino effects have been identified at represents the sum of all the intermediate risks. These the level of consequence and vulnerability studies, a results can serve the network managers, and also the structure for risk communication must be created. It insurers, the financial organizations, and the compe- ensures the transfer of information between concerned tent authorities for the evaluation of the risks. networks, with the goals of planning emergency meas- ures, and reducing and controlling risks. Of course, it must not be forgotten that these networks are part of a Emergency Measures larger set of networks and that information concerning Consequence studies and consequence curves produced risks and their consequences will have to be exchanged are particularly well suited to emergency measures. In with partners. effect, they synthesize the information to be used for Risk communication must ensure an interactive the preparation of an emergency plan. For the man- exchange of structured information between concerned ager of a life support network, preparing for an emer- parties. Therefore, setting up a Risk Management Com- gency means evaluating all possible situations to ensure mittee seems like an essential tool for exchange, dialogue, the best protection to the people potentially touched by mediation, and decisionmaking. All aspects relative to a failure of this network, however remote the chance prevention and preparation will be treated there. of such a situation occurring. In effect, the human and The Risk Management Committee is composed of rep- socioeconomic consequences are too important to resentatives of the two networks: source and destination, 264 Building Safer Cities: The Future of Disaster Risk aswellasrepresentativesoftheconcernedauthorities.The · Transmit information quickly and without dis- committee must consist of different hierarchal levels and tortion expertise. It is also advisable to create working subcom- · Transmit information that sets mitigation meas- mittees for precise domains of expertise, whose results are ures in motion submitted to the committee for discussion and approval. · Transmit information that integrates with the oper- The Risk Management Committee will have to first ations of the destination networks approach certain points, which are described below. · Establish a direct link between personnel of the Prevention: the first responsibility of the committee same hierarchical and operational levels is to ensure that the steps of vulnerability transfer · Create robust, redundant, and compatible links are adequately carried out and then establish the level between the networks. All mechanical and elec- of acceptable consequences for the concerned par- tronic means can be considered. ties. Finally, the important question concerning the The committee will have to look into the nature of disclosure of the risks to the population must be the information needing to be transmitted, and assure addressed. If the anticipated consequences predict a itself of the robustness of these preferred communi- direct danger for the population, it will have to be cation links. brought to the attention of the committee. State of readiness: the committee has the responsi- Preparation: each life support network has the respon- bility for maintaining the systems and functions put sibility of reaching a level of preparedness that per- in place. Therefore, the division of responsibilities mits it to maintain or re-establish, in the shortest must be agreed upon for: possible time, the functions that allow it to fulfill its · Maintenance of the communication channels mission during a disaster or catastrophe. Certain ele- · Verification of the robustness of these channels ments of this preparation must be planned jointly · Training of personnel who intervene in emergency with the partners of the Risk Management Commit- situations tee, including: · Preparation of joint exercises, allowing the readi- · Criteria and warning processes in case of an inci- ness of all participants to be verified. dent · Protocol for these exchanges between networks Application to Developing Countries · Channels of communication used in such cases · Encoding and decoding of transmitted informa- Current State of the Problematic tion as well as the feedback process · The implementation of mitigation measures at the Independent of the location and the operation and level of operations and infrastructure maintenance conditions, each type of life support net- · The decisionmaking levels required and involved work has its own behavior. In effect, the accomplish- in these exchanges of information. ment of the missions of a life support network is Intervention: unlike the prevention and preparation based on the needs (potable water, electricity, etc.) of phases, where information to be exchanged and dis- the community it serves. This statement defines the cussed passes via the Risk Management Committee, structural as well as the functional aspects of the net- the protocol for information exchange is very dif- work. The goal of the current work is to put into place ferent during intervention. Therefore, direct links a theoretical basis for a methodology to understand life between managers and experts must be anticipated support networks in terms of potential failures and to in order to allow an exchange of precise information evaluate their consequences on other life support net- and to favor coordination between networks. The works. Therefore, this approach can be applied suc- preferred channels of communication must adhere cessfully in developed and developing countries. to certain precise criteria: However, the use of this approach in developing coun- · Transport high-quality, concise, precise, and tan- tries requires the prior acceptance of the three princi- gible information ples described below. A New Structural Approach for the Study of Domino Effects between Life Support Networks 265 · Life support networks with a certain degree of com- (age, resistance, grid density, etc.) as well as on a suffi- plexity must be present. This condition, as a general cient knowledge of the aforementioned to ensure ade- rule, corresponds to large urban zones, which are quate planning of network management and operations. particularly sensitive to the failure of these networks. The recent development of disaster management programs "The concern over the risk to megacities, particularly by a number of countries emphasizes establishing national in the developing world, is their growing vulnerabil- policies and initiatives but seldom targets the specific ity caused by their hyper-concentrations of popula- conditions pertaining to large cities. In most cases, the cen- tion, dependence on complex and aging infrastructure, tral government retains the authority for disaster man- and unprepared local institutions" (Bendimerad 2000). agement programs that are often focused on developing · The use of this method requires a regional method- response capabilities, instead of proactive mitigation. With the regulatory environment concentrating decisionmak- ological approach. For example Hurricane Hugo hit ing authority and resources at the central level, the diffi- the Caribbean and the Carolinas in 1989 (Badolato and culty in predicting, assessing, and controlling the impact others 1990; Denis 2002). Blackouts caused the stop- of catastrophes on large cities, with competing priorities page of potable water treatment plants; the lack of water and limited resources, and local government officials defer- lasted from three days to two weeks. The telephone ring decisions and responsibility to the central govern- network on the island of St. Croix, Virgin Islands was ment, the result is that local government action for disaster management is often ineffective. totally destroyed, and rebuilding took several months. Yet experience and modern disaster management prac- To evaluate the consequences on populations, tice recognize the importance of a strong and well-struc- emergency plan measures and the establishment of tured local disaster management capacity, and the need adequate mitigation measures are necessary to ana- for decentralized authority to achieve an effective response. lyze regional networks. A national approach does not During a disaster, local governments are immediately con- permit such a refinement of knowledge. fronted with the responsibility of providing relief to vic- tims but often do not have the means or adequate legislative · The study of these networks, as advocated, implies authority to mobilize these resources. Local governments the consideration of exceptional but noncatastrophic also have difficulties in accessing mitigation funds because events. In effect, if the natural events that trigger the funding and relief agencies typically work directly with failure of these networks inflict considerable damage central governments. Reaching out to local governments on populations and destroy all social infrastructure, to help them build local capacity, acquire knowledge and it is illusory to specifically study life support net- resources and providing them with authority for decisions are essential policies for reducing losses (Bendimerad works. In this case it is preferable to concentrate on 2000). the essentials, the protection of populations. There- fore, there must be a global vision of the risks, Risk Reduction: Avoiding hazards and reducing vulnera- which cover all possible scenarios, not only extreme bility to disasters result when an extreme natural or tech- ones. In terms of planning and preparation of emer- nological event coincides with a vulnerable human gency and mitigation measures it is important to pre- settlement. Reducing disaster risk requires that all stake- pare for intermediate emergency situations so that they holders change their perceptions and behavior to place a donotbecomeexceptionalsituations.The1995Sakhalin high priority on safety in planning and development. Effec- tive risk reduction involves mitigation measures in hazard- earthquake in the Russian Far East (Porfiriev 1996; prone developing countries. Such measures include land Denis 2002) is an example where the consequences use planning, structural design and construction prac- of a large noncatastrophic natural event (rated inter- tices, and disaster warning systems. In addition to employ- mediate) were strongly amplified due to poor knowl- ing scientific and technical knowledge, risk reduction may edge of the communication and energy distribution also involve overcoming the socioeconomic, institutional networks and the domino effects resulting from their and political barriers to the adoption of effective risk- reduction strategies and measures in developing coun- destruction. tries. This may be accomplished through projects analyzing Fundamentally, the difference between life support the possible roles of government, nongovernment, and networks in developed and developing countries resides private sector organizations in risk reduction, local and in the physical characteristics of the infrastructure regional workshops and conferences aimed at heightening 266 Building Safer Cities: The Future of Disaster Risk the awareness of stakeholders to the threat of natural dis- measures to be put in place to protect populations and asters and what can be done about it, and educational and ensure a quick return to normalcy. training activities that increase the understanding of pol- Thus, in this context, it is possible to systematize this icymakers, decision makers and practitioners about dis- approach by missions by simplifying the information to aster management" (World Bank 2000). be transmitted to take into consideration the resources The proposed approach for the study of domino available for the collection of this data. Table 17.2 pres- effects is perfectly consistent with this perspective. ents the essential information for gathering and collat- ing. The information must be provided by the local actors Answer to the Problematic of Developing responsible for the management of the networks and Countries completed by the engineers that designed or rehabili- tated them. In the table, fictional examples taken from It becomes apparent that the principal differentiation several different networks are presented by way of illus- between life support networks in developing countries tration. The essential information is described below. and those in developed countries resides in the physi- · Identification: the life support network is identified. cal characteristics and the planning of emergency meas- · Mission: this information is primary as it forces all ures. These two principal differences will be identified actors to define it together. It is a multidisciplinary below and the answer provided by the method of the work that puts in place the basis for future activity study of domino effects will be made explicit. for planning emergency measures. · Essential infrastructure: it is a matter of determining, with the engineers, the infrastructure that is essen- Physical Characteristics tial for the network to function. For example, one Life support networks are in variable condition, designed identification criterion would be that, faced with a according to nonuniform criteria by multiple contribu- total loss of this infrastructure, the efficiency of the tors originating from varied countries. The situation varies mission would be reduced by at least 75 percent. The according to the country, even the region. Therefore, con- list of these essential infrastructure components should trary to developed countries, there is often a lack of homo- be limited. Operations are not considered at this stage, geneity in the design and operational criteria of the life since the focus is on infrastructure without regard support networks that serve the same community. to human intervention, which is more complex to In this context, the proposed approach adapts itself study. perfectly to the problematic because it is centered on · Links with other networks: The most important, or the definition of the missions of a life support network, essential, direct and indirect links should also be the principal characteristics of the network being linked identified and associated with the mission to which to these missions. It is possible to adapt itself very pre- they are attached. As for essential infrastructure, only cisely to the specifics of a country. the links that significantly influence the networks in Populations must know the missions of the life sup- question need to be identified. It is a matter of iden- port networks and their importance to understand the tifying the type of link (direct or indirect) and the inherent risks. The competent authorities, managers, network connected to this link (destination network). and technical support personnel of these networks The consequences for the destination network can can, in association with the populations, foresee miti- be expressed globally. Next, the people responsible gation measures in the event of failure. These measures for this destination network can be identified. must be in agreement with the local resources and social · All basic information found in this table is available particularities. In terms of emergency measures, in a after the completion of consequence studies, without crisis situation, multiple actors, governmental and others, regard to the causes of the failures or the probability often from abroad, are called on location. The initial that they occur. Obtaining this information does not knowledge of all the missions the networks allows, in require significant resources, and it is available the absence of an emergency plan, permits minimal locally. In addition, the studies or steps necessary to A New Structural Approach for the Study of Domino Effects between Life Support Networks 267 Table 17.2 Example of essential information relative to life support networks Network being studied: Potable water supply Mission: Maintain a minimum pressure Essential infrastructure: Pumping station 5 main conduits (no. 1, no. 2, no. 3, no. 4 and no. 5) Links Identification Type Connected Consequences Responsible (D/I) networks personnel Main water conduit no. 3 I gas A leak in a main conduit provokes Mr. Smith (secondary road X the destruction of a gas conduit, Facade of building Y) which creates a risk of explosion. Mission: Provide quality water Essential infrastructure: chlorination pool reservoir Links: none Network being studied: Electrical power supply Mission: Maintain a minimum voltage Essential infrastructure: Thermal power plant 3 transformer stations (no. 1, no. 2, and no. 3) 2 main transmission lines (no. 1 and no. 2) Links Identification Type Connected Consequences responsible (D/I) Network personnel Transformer station D Potable water supply The loss of electrical power in the transformer Mrs. Smith no. 2 (near the water station provokes the cessation of treatment treatment plant) of water. There is no electrical transmission line backup to the treatment plant. There is a generator at the treatment plan, but it is not in good condition. obtain this information require contributions from the cussion functions can be carried out on the essential principal actors, who must be gathered for the plan- infrastructure affected by the domino effect. ning of emergency measures and crisis management. This information allows concrete and tangible mitiga- tion measures to be established, as much at the level of essential infrastructure as at the level of links generat- Planning Emergency Measures ing the domino effect. In effect, the establishment of any It becomes apparent that knowledge of life support net- prevention and mitigation measures requires a better works is not enough to plan emergency measures ade- understanding of the consequences. The protection of quately. Basic information, such as the minimal example essential infrastructure and the identified links corre- presented in table 17.2, also shows the basic informa- sponds to minimal but indispensable (considering the tion necessary for planning emergency measures. consequences for the population) emergency measures. The preceding information issuing from consequence In effect, the failure of these networks following excep- studies is fundamental because it identifies the essen- tional natural events will greatly amplify the conse- tial infrastructure and the domino effects between life quences for the populations at risk. support networks. These results allow, at the very Preferred channels of communication between those least, to alert those in charge to problems and put in in charge of networks (identified in table 17.2) should place adequate protection measures and emergency be put in place to ensure that concrete measures are measures. Minimal studies of vulnerability and reper- taken. They will be based on routine contact between 268 Building Safer Cities: The Future of Disaster Risk the individuals identified. Later, they can be enriched populations and avoid network failures provoking a sig- and serve as a basis for real risk communication. These nificant amplification of consequences. steps represent a minimal measure of protection for pop- Finally, systematic steps are actually carried out to ulations, especially as the favored measures essentially reduce natural risk in developing countries. The ProVen- aim to reduce socioeconomic impacts. tion Consortium is an example. "The ProVention Con- The principles of communication and management of sortium (2002) is a global coalition of governments, risk do not change, whether we are in a developing coun- international organizations, academic institutions, the try or not. Analysis and management methods and the private sector, and civil society organizations aimed at principle of information exchange remain the same. Only reducing disaster impacts in developing countries. The the techniques of information transmission during an Consortium functions as a network to share knowl- emergency situation can cause a safety problem. Dialogue edge and to connect and leverage resources to reduce must be instituted based on local practices and structures disaster risk. It focuses on synergy and coordination so and a minimal exchange of information must be carried that efforts, and benefits, are shared." The conse- out between life support networks, despite the level of quence studies complete the work of these groups that sophistication of existing means of communication. focus on natural hazards. However, in our approach The development (including construction and reha- these hazards correspond to external trigger events bilitation) of a life support network in a developing that affect essential infrastructure and the links between country should consider, from the start, not only the networks. From that point, the socioeconomic conse- possible consequences of a failure of this network for quences are evaluated more rigorously and actions for other networks, but also the consequences of failures a return to normal are better planned. Information on of other networks on this network. Also, measures essential infrastructure, links, and consequences permit designed to improve the network's own robustness should the planning of replacement actions in response to be anticipated with the principle that the missions of each total or partial destruction. This planning includes the network can be ensured. The identification of informa- hierarchical classification of the infrastructure to be reha- tion needing to be exchanged therefore relies on the bilitated to ensure minimal efficiency for missions judged consequence criteria: what must be demanded from the to be of primary importance; technical, human, and risk generator? The risk generator must carry out a risk financial resources necessary for the replacement of and consequence study and provide the results. At a infrastructure and direct links; and the timeframe for a minimum, national and regional authorities must be return to normal activities. made aware of the potential domino effects of certain This process requires will on the part of the inter- risks on other life support networks. national community and local actors to work together to establish minimal efficient and operational emergency The Role of International Financial Organizations measures to cope with potential failures of the multi- ple missions of life support networks. Organizations that finance these networks can play an initiating role in putting into place, maintaining, and refining the information presented in table 17.2, and Conclusions in developing concrete, efficient, and operational actions of emergency measures planning. Understanding risks relative to life support networks Better control of risk ensures a substantial reduction is fundamental for populations to establish protection in costs and better protection of populations in the event measures in the face of potential failures. When con- of a catastrophe. This is also true in developing coun- sidering the interconnection of life support networks, tries, which are often subject to large-scale natural haz- the domino effects between them are supplementary ards with considerable socioeconomic consequences. phenomena to which a certain importance must be It is therefore mandatory for the people who develop attached. Traditional risk studies based on scenarios and finance life support networks in urban areas vul- composed of extreme natural events do not allow the nerable to natural hazards to begin working to protect entire set of possible situations to be considered. A New Structural Approach for the Study of Domino Effects between Life Support Networks 269 The proposed structural approach to life support net- Focusing the analysis of a network on the missions works is innovative in the sense that it is based on the that it must fulfill offers numerous advantages relative study of the missions that the networks must fulfill. If to the particular context of these networks. In effect, these missions are not respected, the consequences the construction, reconstruction, and rehabilitation of will be analyzed and synthesized without regard for life support networks are generally considered large the causes, allowing operational and efficient emergency projects, and their financing comes from multiple inter- measures to be put in place. For risk evaluation, knowl- national organizations. The management and mainte- edge of the vulnerabilities of the network is critical. nance of these networks involves diverse authorities The proposed structure allows not only the infrastruc- and often nongovernmental organizations (NGOs). ture in question to be evaluated, but also the operation Therefore, it is important that all participants know an of the network that can diminish system efficiency. infrastructure's mission and its relevance to other It is essential to foresee that life support networks are infrastructure. It is the same for all people affected by subject to external and internal risks. External risks, these networks. This knowledge allows missions to be combinations of natural events, technological events by optimized and avoid becoming contradictory or ill- way of the domino effect, and acts of malevolence act suited for a local social context and to adapt financing as triggers, which affect the network at the essential to local realities. infrastructure level. These result in a decrease in oper- Knowledge of consequences allows better planning ational efficiency of the network's mission. This approach, of minimum emergency measures. In effect, consequence combining natural and anthropic events, is in line with studies being a representation of human and socioeco- a new international tendency in risk studies that does nomic risks, it is thus possible to concentrate on the not uniquely consider scenarios based on natural events essential, as much on the level of protection provided of a known frequency. The domino effects between life for populations as for the mitigation measures to put in support networks can now be studied systematically by place. For organizations that finance these networks, transferring the vulnerability of a network, the source knowledge of these basic consequences allows the costs of failure, to another network. Thus, this transferred that will be assumed for rehabilitation and reconstruc- failure becomes an external technological cause that tion following a failure to be evaluated. A basic conse- influences the efficiency of one or sets of essential quence study for each mission of a life support network infrastructure of a destination network. This structural should be systematically carried out, especially consid- approach allows, finally, risk communication mecha- ering that it does not require significant resources. nisms between life support networks, structured by way The proposed approach focuses on the missions of of Risk Management Committees, to be put in place. life support networks, coupled with consequence and These committees complete the emergency measures vulnerability studies, and has to its credit the ability to by bringing together the concerned networks and assur- identify the sensitive points of networks to put mini- ing efficient transmission of precise and tangible infor- mum emergency measures in place. For the competent mation, allowing mitigation measures to be put in place authorities who participated in financing these net- during a crisis. works, this approach offers a concrete picture of the This structural approach is particularly well suited potential risks and, therefore, favors the implementa- for developed countries with complex life support infra- tion of efficient mitigation measures for populations. structure and emergency measures. The results of this work will allow the state of preparedness of the vari- ous participants to improve. Concerning developing Bibliography countries, situations vary from country to country. Allen H. 1997. "Making a business of dam safety." International Life support networks are in various conditions and Water Power Dam Construction 49(8): 20­21. designed according to nonuniform criteria. Emergency Badolato E.V., J. Bleiweis, J.D. Craig, and H.W. Fleming. 1990. measures are not always efficient. The proposed approach, Hurricane Hugo: Lessons Learned in Energy Emergency Pre- in this context, offers several interesting alternatives. paredness. Clemson, South Carolina: The Clemson University 270 Building Safer Cities: The Future of Disaster Risk Strom Thurmond Institute of Government and Public El-Shahhat A.M., D.V. Rosowsky, and W.F. Chen.1995. "Account- Affairs. ing for human error during design and construction." Jour- Bendimerad F. 2000. "The risk of natural disasters occurring in nal of Architectural Engineering 1(2): 84­92. megacities is growing. Proper planning by local governments Garin H. 1994. "AMDEC/AMDE/AEEL: l'essentiel de la méth- can reduce the human and financial loss to cities." Available from ode." Association française de normalisation, Paris: 39. . nication. University of Twente: Kluwer Academic Publisher: 213. Bier V.M., Y.Y. Haimes, J.H. Lambert, N.C. Matalas, and R. Hadipriono F. C., C.L. Lim, and K.H. Wong. 1986. "Event tree Zimmerman. 1999. "A Survey of Approaches for Assessing and analysis to prevent failures in temporary structures." Journal Managing the Risk of Extremes." Risk Analysis 19(1): 83­94. of Construction Engineering and Management 112(4), Decem- BIT (Bureau International du Travail). 1990. Prévention des acci- ber, 500­513. dents industriels majeurs. Genève:: 117 Hubbert G. and B. Ledoux. 1999. Le coût du risque... L'évalua- CAN\CSA. 1991. Risk analysis requirements and guidelines: qual- tion des impacts socio-économiques des inondations. Paris: Presses ity management. A national standard for Canada. Conseil cana- de l'École Nationale des Ponts et Chaussées: 232. dien des normes, Association canadienne de normalisation, IPCC (Intergovernmental Panel on Climate Change). 2001. "Cli- CAN\CSA Q850: 49. mate Change 2001: The Scientific Basis." Contribution of CAN\CSA 1997. Gestion des risques : Guide à l'intention des décideurs, Working Group I to the Third Assessment Report of the Norme nationale du Canada. Conseil canadien des normes, Intergovernmental Panel on Climate Change. J. T. Houghton, Association canadienne de normalisation, CAN\CSA-Q850- Y. Ding, D.J. Griggs, M. Noguer, P. J. van der Linden, and D. 97: 49. Xiaosu, eds. Cambridge: Cambridge University Press: 944. Cazaubon C., G. Gramacia, and G. Massard. 1997. "Manage- ISDF (Institut de Sûreté de Fonctionnement). 1994. "L'état de ment de projet technique: méthodes et outils." Collection tech- l'art dans le domaine de la fiabilité humaine." Octares édi- nosup, génie de production. Paris: Ellipses: 182. tions, collection ISDF: 446. Covello, V.T., J. Menkes, and J. Mumpower. 1986. Risk Evalua- Isenberg J. 1991. "Performance of lifeline and emergency response tion and Management. New York: Plenum Press: 435­457. in Watsonville, CA, to Loma Prieta earthquake." Proceedings Covello, V.T. and M.W. Merkhofer. 1993. Risk Assessment Meth- of 4th US Workshop Earthquake Disaster Prevention Lifeline ods: Approaches for Assessing Health and Environmental Risks. System: 381­390. New York: Plenum Press: 319. Kaplan S. 1997. "The world of risk analysis." Risk analysis 17: Denis H. 1998. "Comprendre et gérer les risques sociotech- 407­417. nologiques majeurs." Éditions de l'École Polytechnique de Kirwan B. 1994. A guide to practical human reliability assessment. Montréal: 342. London: Taylor and Francis: 592. Denis H. 2002. La réponse aux catastrophes, quand l'impossible Lau D.L. and others. 1995. "Performance of lifeline during the survient. Montréal, Canada: Presses Internationales Polytech- 1994 Northridge earthquake." CJCE 22(2): 438­451. nique: 318. Lavallée D., C. Marche, and T. Quach. 2000. "De nouvelles Dhillon B. S. and S.N. Rayapati. 1985. "Analysis of redundant approches pour quantifier le risque de rupture d'une retenue systems with human errors." Proceedings of the Annual Relia- d'eau." Canadian Journal of Civil Engineering 27:1217­1229. bility and Maintainability Symposium, Philadelphia, PA: 315­321. Law, Averill M. and David W. Kelton. 1991. Simulation Modeling Dhillon B. S. 1989. "Modeling human errors in repairable sys- and Analysis. Columbus, OH: McGraw Hill: 221. tems." Proceedings of the Annual Reliability and Maintainability Leiss, William. 1996. "Three phases in risk communication prac- Symposium, January 24­26 1989, Atlanta, GA: 418­424. tice." In Annals of the Academy of Political and Social Science, Dynes, Russell S. 1994. "Community Emergency Planning: False Special Issue, Challenges in Risk Assessment and Risk Manage- Assumptions and Inappropriate Analogies." International Jour- ment. Thousand Oaks: 15­30. nal of Mass Emergencies and Disasters 12 (2): 141­158. Leiss, W. and C. Chociolko. 1994. RiskandResponsibility. Montreal: Dynes, R. and K. Tierney, eds. 1994. Disasters, Collective Behav- McGill-Queen's University Press: 405. ior, and Social Organization. London and Toronto: University Leiss, W. and D. Krewski. 1989. "Risk Communication: Theory of Delaware Press, 71­84, 145­159. and Practice." In W. Leiss, ed., Prospects and Problems in Risk A New Structural Approach for the Study of Domino Effects between Life Support Networks 271 Communication. Waterloo: University of Waterloo Press, Water- Reason J. 1990. Human Error. Cambridge: Cambridge Univer- loo: 90­111. sity Press: 302. Lemperiere F. 1993. "Dams that have failed by flooding: an analy- Reason J. 2000. "Human error: models and management." British sis of 70 failures." International Water Power Dam Construction Medical Journal 320: 768­770. 45(9­10): 19­24. Renn, O. 1998. "The role of risk communication and public Lemperiere F. 1999. "Risk analysis: What sort should be dialogue for improving risk management." Risk Decision and applied and to which dams?" International Journal of Hydropower Policy 3: 5­30. Dams 6(4): 128­132. Ridley, L.M. and J.D. Andrews. 2000. "Reliability of Sequential Leverenz F. L., Chadwell G. Bradley, and S.D. Unwin. 1996. Systems Using the Cause-Consequence Diagram Method." "Process event tree based hazop studies." Battelle Memorial Proceedings of the 14th Advances in Reliability Technology Sym- Institute. http://www.battelle.org/environment/irm/pdfs/pet- posium, Manchester, November. Available from . Middelkauf, R.D. 1989. "Regulating the safety of food." Food RNC (Ressources Naturelles Cananda). 2002. "Programme de Tech 43: 296­307. Séismologie des Tremblements de Terre (PSTT)." Ressources Modarres M., M. Kaminskiy, and V. Krivtsov. 1999. Reliability Naturelles Canada. Available from . Marcel Dekker: 542. Robert, B. 2001a. "A Method for the Study of Cascading Effects Moses F. 1998. "Probabilistic-based structural specifications." Within Lifeline Networks." Workshop on Mitigating the Vul- Risk Analysis 18(4): 445­454. nerability of Critical Infrastructure to Catastrophic Failures, NASA. 1995. "System reliability assessment using block dia- Alexandria, Virginia, September 10­11. Available from gramming methods." Preferred Reliability Practices, Practice . No. PD-AP-1313: 5. Robert, B. 2001b. "Outils d'aide à la décision pour les plans Nicolet J. L., and J. Celier. 1985. La fiabilité humaine dans l'en- de gestion des aménagements hydroélectriques." Proceedings treprise. Collection le nouvel ordre économique. Paris: Masson: of CDA Annual Conference. Fredericton, Canada, October: 302. 39­46. O'Leary S. 1997. "Concrete rehabilitation for dams." Pro- Robert, B., C. Marche, and J. Rousselle. 2002a. "Computeriza- ceedings of International Conference on Hydropower 1610­ tion and use of flood impact curves." ICLR Research, Paper 1622. Series XX. Institute for Catastrophic Loss Reduction, April. Petts, J. 1992. "Incineration risks perception and public con- Robert, B. 2002b. "Étude des risques et des effets domino pro- cern: Experience in the UK improving risk communication." duits par la défaillance d'un aménagement hydroélectrique." Waste Management and Research 10: 169­182. Proceedings of Annual Conference of the Canadian Society for Civil Plate E.J. 1996. "Dams and safety management at downstream Engineering, June. valleys." Proceedings International NATO Workshop Dams Safety Rowe, W.D. 1997. An Anatomy of Risk. Wiley Series in System Manage Downstream Valley: 27­43. Engineering and Analysis. New York: John Wiley and Sons: Porfiriev B.N. 1996. "Social aftermath and organizational response 256. to a major disaster: the case of the 1995 Sakhalin earthquake Seidou, O. 2002. "Intégration du risque dans les modes de ges- in Russia." Journal of Contingencies and Crisis Management 4(4): tion des systèmes hydriques." Philosophiae Doctor Thesis 218­227. (Ph.D.), École Polytechnique de Montréal, Canada: 251. Powell, D. 1996. An Introduction to Risk Communication and the Selcuk A.S. and Y.M. Semih. 1999. "Reliability of lifeline net- Perception of Risk. Ontario: University of Guelph: 17. works under seismic hazard." Reliability Engineering System ProVention Consortium. 2002. . Stedinger, J.R., D.C. Heath, and K. Thompson. 1996. "Risk analy- Quach T.T., J. Gagnon, B. Robert, and C. Marche. 2000. "Upper sis and safety evaluation: hydrologic risk." Report No. 96-13- Ottawa River: A Four Step Risk Assessment." 20th Congress, 16:51. Institute of Water Resources. International Commission on Large Dams, Beijing, China, Sep- Suokas J. and V. Rouhiainen. 1993. Quality Management of tember 19­20, 2000. Q76-R32: 479­498. Safety and Risk Analysis. Elsevier Science Publishers BV: 296. 272 Building Safer Cities: The Future of Disaster Risk Tyagi A. and C.T. Haan. 2001. "Reliability, Risk, and Uncer- ception and Communication. Washington, D.C.: National tainty Analysis Using Generic Expectation Functions." Jour- Academy Press: 332. nal of Environmental Engineering 127(10), October: 938­945. World Bank. 2000. World Development Report 2000/2001: Attack- United Nations. 1984. "Proceedings, Engineering Foundation ing Poverty. Available from . Resources Series No. 58. New York: United Nations: 117. Zielinski, P.A. 2001. "Flood frequency analysis in dam safety US NRC (United States National Research Council). 1989. assessment." Proceedings of CDA Annual Conference, Frederic- "Improving risk communication." Committee on Risk Per- ton, Canada, October: 79­86. Chapter 18 Mitigating the Vulnerability of Critical Infrastructure in Developing Countries Lamine Mili Critical infrastructure is a collection of facilities and development of the Brazilian energy crisis of 2001 that institutions that provide vital services to people and stem from the combined effects of the energy policies economies. Its disturbance would impair the defense or undertaken by the Brazilian government since the 1980s economic security of a country. It includes the follow- and the severe drought that steadily reinforced its grip ing systems: on several regions of the country since 1998 are ana- · telecommunications lyzed. Specifically, the lack of diversification in energy · electric power resources and the meager investment in transmission · gas and oil storage and transportation and generation expansion since the late 1980s exposed · water supply the Brazilian electric power infrastructure to a series of · transportation system-wide blackouts while leading the country to a · banking and financial services painful energy rationing. The latter will certainly have a · emergency-service institutions, including hospitals, negative impact on economic growth in the years to come. police stations, fire and rescue departments To increase the security and efficiency of the power · government institutions. infrastructure at a minimum cost, we advocate: (1) the The previous definition applies to any country, indus- implementation of a defense plan against catastrophic trialized or in the process of industrialization. Here the failures through islanding initiated by special protec- telecommunication systems comprise not only the cable tion systems; (2) the development of a fault detection, and wireless telephone networks, but also the commu- isolation, and restoration scheme for the power distri- nication and earth observation satellite systems and bution system that aims to decrease the time of recovery computer networks, including the Internet. This infra- after a blackout; and (3) the installation of computer- structure can be regarded as the backbone of the aided software programs at control centers to seek the economies of developed and the developing countries best topology of the distribution network and yield min- worldwide since they provide material support for the imum electric losses. delivery of basic services to all segments of a society. Finally, we promote the development of a global infor- Their maintenance, expansion, and protection against mation infrastructure that provides effective disaster all types of threats, manmade, technological, or natural, risk management against extreme weather conditions are of great concern to state executives and govern- at a global scale. Using satellite technologies and the ment constituencies (Clinton 1996). Internet, this infrastructure links together disaster cen- In this report, attention is restricted to telecommu- ters strategically located around the globe to enable the nications and electric power systems. Their key role in twin tasks of: (1) detecting any incipient disasters by economic growth in developing countries and growing processing the huge volume of data received from the vulnerability to technological failure and extreme nat- earth observation satellites; and (2) broadcasting early ural hazards due to the global warming are high- warnings to a wide public through the Internet and lighted. To substantiate these claims, the genesis and other telecommunications media. 273 274 Building Safer Cities: The Future of Disaster Risk Positive Impact of the Electricity and provides a cleaner living environment since electric Telecommunication Services pumps deliver running water to communities. They also reduce the time spent on household chores and increase As indicated by the recent World Bank report (World opportunities for learning since electricity allows people Markets Research Center 2001: 119) entitled "World to read more easily at night and access broadcast infor- Infrastructure and Development," electricity and telecom- mation and cultural programs offered by radio, televi- munication services in developing countries can sig- sion, and now, the Internet. nificantly contribute to economic development by Since electricity is instrumental in the development providing new opportunities for small businesses to of national economic markets, most developing coun- develop and grow, and therein to propose new jobs tries have implemented ambitious programs of rural and services. Not only can the local market enlarge electrification since the early 1970s. Distribution power and consolidate its direct geographical basis, but it networks have been expanded to reach rural towns may also find ways to improve the productivity of its and villages. The installations of decentralized renew- workforce to the point that it can compete at the regional able electric energy sources have flourished where the and the national level. It may even target the global econ- cost of power grid connection is too high (World Energy omy via the Internet and the multitude of services that Council 1999). In China, small hydropower stations it conveys should computer infrastructure exist. This with a total capacity of 17 GW supply nearly 20 per- is exemplified by the US$8.3 billion software industry cent of rural electricity. In Kenya, a dynamic private in India whose exports, mainly to the United States sector has so far provided more than 1 MW of photo- and Europe, have shown impressive growth rates since voltaic panels to 20,000 households. With the finan- 1995 to reach US$6.2 billion in 2001, placing it in fourth cial help of the World Bank, the government of Argentina place after Microsoft, Oracle, and SAP (Ravisankar 2002). has recently encouraged the emergence of a private com- This industry employs more than 500,000 profession- petitive market consisting of small distributed electric als and accounts for nearly 2 percent of the Gross Domes- energy resources. The latter comprise solar photovoltaic tic Product (GDP) of the country (NASSCOM 2002). panels, micro-hydropower units, small wind-powered While China also has a fast-growing software industry, stations, and diesel generating units. These are only a it cannot yet compete with that of India. few examples of a long list of success stories in Asia, Tourism industries, traditional handicraft businesses, Africa, and Latin America and the Caribbean. and educational institutions are also taking advantage of the Internet to market their products worldwide. Three successful initiatives have been undertaken by Low Level of Rural Electrification in the Developing the Naushad trading company of Kenya (Naushad 2002), Countries the Internet Bazaar of India (Internet Bazaar 2002) and the African virtual university, which links 25 learning Despite significant efforts undertaken so far, the vast centers in 15 African countries, a project funded by majority of rural populations in developing countries the World Bank in Kenya (Phombeah 2000). Many other are still relying on wood, dung, agricultural wastes, and businesses mushroomed throughout Africa, Asia, other biomass as accessible cheap fuel for cooking, heat- Latin America, and the Caribbean thanks to the PAN ing, and lighting (World Energy Council 1999). One Networking Program initiated and funded by the extreme case is provided by Sub-Saharan Africa, where International Development Research Centre (IDRC), fewer than 8 percent of households are electrified. In fact, Ottawa, Canada (IDRC 2002). The PAN networking this figure masks a wide disparity between countries, groups provide basic training and financial support to between rural areas and cities, and between poor and rich rural businesses in information technologies and Web people (World Markets Research Center 2001). India pro- site development, among other activities. vides another instructive example. Power networks there For poor communities, another beneficial impact of reach about 80 percent of all the villages, though only electricity and telecommunication services is the enhance- 31 percent of households can afford the connection ment of social welfare. The availability of these services costs. Poor maintenance of equipment and excessive Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 275 energy losses, which may be as high as 30 percent or Such pollutants pose a serious threat to the health more if unpaid electricity usage is included, afflict the of exposed populations, primarily children and the eld- power industry and raise the cost of production and erly. They have been blamed for hundreds of thou- distribution of electric energy (Word Bank 1998). sands of premature deaths in poor suburbs and rural To diversify the cost of electrification, thereby areas of Africa, India, China, and Southeast Asia. They allowing the poor to afford its service, some econo- are also destabilizing the planet's climate through the mists have questioned the high quality standards that greenhouse effect that deflects sunlight from the earth's have been enforced by most developing countries, which surface and traps heat. These undesirable effects in turn are often simple copies of those adopted in developed may cause in the near future a significant decline in agri- countries. As indicated by the World Bank Infrastruc- cultural productivity, jeopardizing the economic growth ture Forum of May 2001 (World Markets Research Center of many countries, especially in South and East Asia. In 2001), this diversification of price could be accom- India, a decrease of 5 to 10 percent in rice productiv- plished by replacing the monopolies of the electric energy ity is anticipated shortly (Topfer and Ramanathan 2002). sector with private energy providers that would com- In view of the increasing number of extreme events pete to sell power at least cost. In fact, such competi- occurring on the planet, climate change is undoubt- tive electric markets are currently emerging worldwide edly taking place (Working Group II of IPCC 2001). as a result of aggressive restructuring policies enacted Long-lasting droughts and extreme hot weather are by an increasing number of governments of both devel- afflicting many developing countries, putting exceed- oped and developing countries. The main motivation ingly high stress on their water reserves and electric of the latter is to create a favorable market environment power infrastructure. Western Asia and North Africa that would encourage private national and international are suffering from severe water shortages, while Brazil companies to invest in the expansion of the electric and India experienced large-scale blackouts in 2002 power system. This expansion responds to the impor- due to an overload of their electric power systems. Of tant increase in consumption that results from the eco- a more destructive nature are the cyclones and floods nomic and population growth in these countries. that recently ravaged Central America and some parts of Asia. These natural hazards are occurring with unprece- dented magnitudes and frequencies as testified by the The Need to Reconcile Economic Development four successive hurricanes that struck countries of Cen- and the Environment tral America and the Caribbean from August 1998 to Unfortunately, the form of economic development cur- November 1999. These were Hurricanes George, Mitch, rently adopted worldwide comes at the expense of the Floyd, and Lenny in August 1998, October 1998, July environment. It not only induces deforestation and deser- 1999, and November 1999, respectively. Of these four, tification over vast regions of the globe, but also gives Hurricane Mitch was the most destructive. It devastated rise to enormous air pollution. This pollution stems from large segments of the Honduran and Nicaraguan trans- the release into the atmosphere of large volumes of green- portation and communication infrastructure, the recon- house gases such as ozone, dioxide, and carbon monox- struction of which was still underway two years later ide, among others. As recently discovered, it also results (USGS 1998). By the magnitude of its impact, Hurri- from a huge amount of anthropogenic aerosol, termed cane Mitch ranks among those that occur once in one the Asian brown cloud (Topfer and Ramanathan 2002). or two centuries (USAID 2001). According to clima- This is a three-kilometer-deep haze of dust, black carbon, tologists, El Niño, La Niña, and the Asian Brown and sulfates, which stretches over a large part of Asia. Its Cloud are to be blamed for the high frequency and mag- existence has been uncovered by an international team nitude of these natural disasters. of scientists working under the umbrella of the Center In the light of the foregoing discussions, a burning for Clouds, Chemistry and Climate on the Indian Ocean question arises: How to reconcile economic develop- Experiment (INDOEX) project commissioned by the ment with the preservation of the environment? Are United Nations Environment Programme (UNEP). these two incompatible? This issue is at the center of 276 Building Safer Cities: The Future of Disaster Risk most of debates taking place in workshops and con- telephone system can be regarded as the first infra- ferences sponsored by the United Nations and the World structure to reach a global scale since a point-to-point Bank. Two key concepts are emerging: (1) adaptation connection can be established between any pair of to climate change, and (2) economic sustainability. telephones around the world. This is achieved via an The former calls for the implementation of technolog- adequate combination of cables and wireless technolo- ical, economic, and social measures to mitigate the gies, including several constellations of satellite systems impact of extreme events, while the latter requires the at low, medium, and geostationary earth orbits. Com- adoption of new forms of economic development that puter networks have a similar propensity to globaliza- promote environmentally-friendly technologies and the tion as more and more countries equip themselves eradication of extreme poverty in the world. Any pro- with Internet technologies. posed solutions will revolve around key questions related While being recognized as the most complex system to the development of critical infrastructure, mainte- ever built by human beings, electrical power systems nance, and access to all. In the next sections, we will have not yet attained a global scale. However, some such focus mainly on the former problem by proposing mit- as the North American and European interconnected igation measures for electric power and telecommuni- power systems, reach continental size. Note that the cation infrastructure. latter system can be regarded as a single system servic- ing not only Western- and Eastern-Continental Europe, but also the United Kingdom, Scandinavia, and Russia. The Growing Vulnerability of Critical In addition, it is now expanding its reach through the Infrastructure Strait of Gibraltar and the Bosphorus to interconnect in the near future with all power grids in the Mediter- Main Characteristics of Networked Infrastructures ranean region. The first phase of this project has already Critical infrastructure systems such as water, energy, been completed with the installation of a 400-kV AC telecommunications, computer, and networked banking submarine cable between Spain and Morocco (Janssens are becoming increasingly interdependent as the digital and Kamagate 2002). Future projects to link the Russ- society matures on a global scale. Consequently, the vul- ian power system to the high-voltage power grid of Japan nerability of these stratified networks is raising major con- and Alaska (Koshcheev 2001) also exist. However, we cerns worldwide. For instance, the normal operation of should recognize that the advantages brought by these water, telecommunications, and banking systems is main- interconnections also come with possible technical prob- tained only if there is a steady supply of electric energy. lems such as sustained loop flows around seas and moun- On the other hand, the generation and delivery of elec- tains and low-frequency oscillations across long tric power cannot be ensured without the provision of transmission lines connecting generating units to remote fuel, water, and various telecommunications and com- load centers. The latter is more thoroughly outlined in puter services for data transfer and control purposes to the emerging technologies section of this paper. the power plants and networks. These interdependen- cies are strengthening their grip as the usage of the Inter- Increasing Vulnerabilities to Extreme Natural Hazards net and other computer networks becomes prevalent. Energy and telecommunications infrastructure each Newspapers report daily the devastating effects of form a complex interconnected networked system that cyclones,floods,anddroughtsonpeopleandinfrastructure stretches over a large geographical area to reach in worldwide. These extreme events strain critical infra- principle every household and economic entity in a structure in countries most at risk, especially less devel- region. Their functions in a society are similar to those oped countries. In Central America, floods and hurricanes of arteries and veins that branch out through the human that take many lives and damage infrastructure in body to nourish every cell with vital nutrients. Infra- these mostly developing countries have been recurring structure's covered geographical areas may include a at a higher rate than usual since the early 1990s. India continent or even the whole world. In fact, the public experiences small-scale blackouts on a regular basis Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 277 throughout the country. This adversity does not spare Spain and several other Western European countries, its frenetic capital, New Delhi, where electricity cuts killing dozens of people and damaging infrastructure. lasting several hours are not uncommon. However, from The impact of the wind was so intense in France that time to time, a huge blackout occurs across several Indian it inflicted considerable damage to the French public states, affecting millions of people and producing large transport network and the electric power networks, economic losses. This was precisely the case during leaving millions of people without electricity for several the blackout of July 30, 2002, which spread over five days (ABC News 1999). Like the Canadian disaster, the states of Western India. It was due to a heat surge that restoration of the French power system required sev- followed several months of intense drought that afflicted eral weeks to complete and came at a high price. many regions of the country. The blackout was triggered by a load increase that went beyond the maximum capa- Increasing Risks of Cascading Failures Triggered bility of the transmission system (Rediff News 2002). by Local Disturbances One of its victims was the Indian railway system, where service was interrupted for several hours. The integrity of critical infrastructure is at risk world- The fate of India is not isolated. Brazil faces similar wide not only because of the growing frequency of difficulties due to the lasting drought the country has extreme events of natural causes, but also because they suffered from since 1998. The drought has induced an are increasingly vulnerable to local disturbances. This energy crisis resulting from the heavy reliance of the is in part due to the strong reliance of critical infrastruc- Brazilian power system on complex water reservoirs ture systems on one another, which may turn a local feeding numerous hydroelectric plants placed in cas- disturbance in one system into a large-scale failure via cade which account for more than 90 percent of the cascading events that has catastrophic consequences on installed generation capacity (Augusto and others 2002). society as a whole. It is also in part due to the current Despite a strict rationing of 20 percent of its electric trend to operate critical networked systems closer to consumption, instated with the active participation of their stability or capacity limits. One compelling reason the population from June 2001 to February 2002, ten for this practice is, of course, economics. Providing states located in the south, southeast, and center-east this infrastructure with some degree of resiliency comes of the country experienced widespread blackouts on at a price, which is entailed by the required level of January 21, 2002 (Latin America Regional Reports 2002). redundancy in the equipment that needs to be achieved. These are the most populous and industrious states of This is even more true in developing countries, where the country. Because the energy crisis may jeopardize its the expansion of critical infrastructure systems does not economic growth in the long run, Brazil is seeking to diver- keep pace with rapid growth in demand. sify its electric energy production, mainly by boosting Another reason for the degradation in infrastructure the construction of gas-fired plants (The Economist reliability is the detrimental role played by hidden fail- 2002; Prates 2001). The Brazilian energy crisis will be ures in the equipment (Thorp, Phadke, Horowitz, and further analyzed in a later section. Tamronglak 1998). Hidden failures are hardware or soft- Developed countries are also facing unusually harsh ware failures that are only exposed when a system or a weather conditions that expose critical infrastructure to portion of a system is highly stressed due to congestion the risk of partial or total destruction. For example, in or fault. In other terms, hidden failures cannot be revealed North America, an ice storm of an unprecedented before the system is perturbed. In particular, routine magnitude knocked down a sizable segment of the elec- maintenance testing may not detect them or even worse, tric transmission grid of Canada and the North Eastern may induce them. This was precisely the case in the region of the United States from January 5­10, 1998. 1977 New York blackout (Sugarman 1978; Wilson The restoration of the Canadian network took months and Zarakas 1978) where a protection relay was dam- and cost the taxpayers several billion dollars. In Europe aged during a testing procedure a few weeks before the on December 17 and 26, 1999, two consecutive waves power system failure. Another cause of hidden relay fail- of fierce windstorms hit France, Germany, Switzerland, ures is the present practice in electric power systems to 278 Building Safer Cities: The Future of Disaster Risk favor dependability over security in relay settings to affect the present and future states of these systems. For- ensure the isolation of a fault with high probability while tunately, these measurements are becoming available allowing the tripping of non-faulty devices from time as more Earth Observation Satellites (EOS) are placed to time. Hence, it should not come as a surprise that a in orbit. These space platforms measure such things as North American Electric Reliability Council report (NERC cloud cover, crust tectonics, ocean movement, polar ice 1988) blamed hidden failures in the protection sys- surface and displacement, vegetation density and cov- tems for aggravating the situation in 73.5 percent of sig- erage, and soil moisture level. nificant disturbances that were investigated in the U.S. In 1972, NASA launched LANSAT-1, the first earth electric power transmission network. This is a sizable observation satellite. Because NASA made the images portion of major failures in power systems that should taken by LANSAT-1 available to the public, many devel- call out for the development of mitigation measures oped and developing countries built and operated ground not only in the United States, but also in other devel- stations to receive them. Developing countries that were oped and developing countries. able to take advantage of the images included Brazil, Zaire, Another example of a hidden failure that wreaked India, Iran, and China. Since then, many EOS satellites havoc with the normal operation of critical infrastruc- were put into orbit by the United States, Japan, France, ture was a software bug that existed in the switching and other members of the European community. As an systems of the AT&T long-distance public telephone example, figure 18.1 shows the Advanced Land Observ- network before its general breakdown via cascading fail- ing Satellite (ALOS) of Japan (NASDA 2002). Among ures on January 15, 1990 (Neumann 1995, 1998; Mason developing countries, only China and India have so far 1990; Travis 1990). The triggering event was a failure founded their own satellite programs that include in a switch whose detection was passed to all its neigh- EOS. Brazil and Argentina have established space bors. As a result, the latter switches correctly took action agencies, while Kenya hosts the United Nations Envi- not to forward calls through the failed one. Unfortu- ronment Programme and South Africa has established nately, a software bug in all of the switches prompted the Satellite Application Center/Council for Scientific each of them to mistakenly notify their neighbors that and Independent Research. All process the huge data- they were failing and calls were not to have been for- base created from the measurements sent by EOS, which warded through them. This domino effect resulted in entitle them to be members or associates of the presti- the interruption of the long-distance telephone service gious and very active Committee on Earth Observation in the United States for nine hours. Satellites (CEOS). Since its foundation by the Group of 7 at the 1984 Summit of Industrialized Nations, about 40 EOS were put into orbit and more than 50 more mis- Satellite and Information Technologies sions are anticipated in the near future (CEOS 2002). for Natural Disaster Risk Management There are two types of EOS: low-earth-orbit (LEO) satellites and geostationary-earth-orbit (GEO) satellites. The Earth Observation Satellites The former revolve on polar orbits at an altitude that The management of natural disasters consists of five ranges from 500 to 1000 kilometers and carry either main tasks that need to be achieved in a cyclic manner: optical sensors that cannot penetrate clouds or microwave prevention, mitigation, warning, response, and recov- sensors that have the ability to operate in a cloudy and ery (Wood 2002). Prevention and mitigation require foggy environment. Among these sensors, synthetic assessing the risk of various potential hazards and taking aperture radars provide the best image resolutions, which action in the long- and mid-term to mitigate their harm- can be as high as one inch. On the other hand, the ful impacts. A premise of the disaster-risk manage- GEO satellites are placed on a 24-hour equatorial orbit ment process is the availability of good models of the that allows them to remain stationary with respect to earth's climate and geological systems. The development the earth's surface at an altitude of 36,000 kilometers. and validation of these models rely on the accessibility These are meteorological satellites that take images in the of a comprehensive measurement set of all variables that visible or infrared spectrum at intervals of 15 minutes Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 279 for U.S. satellites GOES-8 and GOES-10 and 30 minutes maps that are geographically referenced. For disaster- for all the others, including the Indian National Satellite, risk-management applications, these images are those INSAT (Wood 2002). provided by the EOS satellites and other remote-sensing technologies. They are extremely useful during the warn- ing and disaster relief phases, since they allow rescue The Communication and Navigation Satellites personnel to have handy spatial views of predicted or Two other types of satellites, communication and nav- current impacts of hazards. igation satellites, play a key role in disaster management, especially during the warning and relief phases. Their Development of a Global Information Infrastructure advantage over alternative technologies comes from the for Disaster Management fact that they are less prone to failure during the impact of natural disasters. As their name indicates, communi- In an informative paper, Jayaraman, Chandrasekhar, cation satellites retransmit electromagnetic signals that and Rao (1997) of the Indian Space Research Organi- they receive from and to a ground station or another satel- zation convincingly argued that the EOS and commu- lite. On the other hand, navigation satellites, commonly nication and navigation satellites in addition to the known as Global Positioning Systems (GPS), allow a Internet could greatly help in managing natural disas- GPS receiver to calculate its position, velocity, and time. ters on a global scale. To this end, they called for the Since 1995, the most accurate receivers have provided development of a Global Information Infrastructure measurements within 16 meters of a position, 0.1 meters (GII) with the capability to transfer via satellite to dis- per second for velocity, and 100 nanoseconds for time aster management centers at high speed large volumes (Graham 1995). The GPS systems form a constellation of data sensed by the EOS. The feasibility of this tech- of 24 geostationary-Earth-orbit satellites that have a nology was recently demonstrated by the Trans-Pacific 12-hour revolution at an altitude of 20,900 kilometers. High-Data-Rate Satellite-Communications experiments. These satellites consist of six groups of four platforms These experiments were conducted by the Disaster- placed on six 55°-inclined orbital planes 60° apart. This Observation-Satellite working group as one of the configuration allows four satellites to be visible any time many tasks of the Japan-U.S. cooperation in space from any location on earth. projects. Distributed at strategic locations around the The GPS systems found broad application world- globe, the GII centers will process satellite data to detect wide. In particular, GPS receivers are placed on board any incipient disasters and widely disseminate the infor- military and civil aircraft, ships, and land vehicles. Fol- mation to governments and research institutions, which lowing in the steps of developed countries, India recently in turn should make them available to a broad public announced that its Civil Air Navigation system will be via the Internet and other telecommunication media. based on a GPS- and GEO-Augmented Navigation The GII will also help doctors practice telemedicine system, called GAGAN (Singh 2002). However, a recent during the disaster response and relief phases. Telemed- report released by the Office of the Assistant Secretary icine is being successfully applied in India to overcome for Transportation Policy of the U.S. Department of the acute lack of health infrastructure and medical per- Transportation warned the general public of potential sonnel in 600,000 villages (Harris 2002). The country vulnerabilities in the transportation infrastructure that has a ratio of only 1 bed for 1,333 Indians. This is sig- relies on GPS. It indicates that GPS signals are prone to nificantly less than the 1-bed-to-600-people ratio in the various intentional and unintentional interference, Philippines, for example. Thanks to the initiative of the including atmospheric interference. To mitigate these Indian Space Research Organization (ISRO) and the Min- vulnerabilities, the report calls for the usage of backup istry of Information Technology, 200 telemedicine sites systems. interconnected via 128-kbps integrated-service digital One important application of GPS is the Geographical networks and communication satellites are already oper- Information System (GIS). This is a computer software ating in the country. As a first step in the extension of program that processes and displays spatial images and this network nationwide, ISRO is targeting the state of 280 Building Safer Cities: The Future of Disaster Risk Orissa to develop an ambitious statewide telemedicine sion and distribution networks, a growing number of network in the near future. A successful story of the countries have embarked on a structural reform of the application of telemedicine was provided by the Gujarat power industry geared toward the emergence of com- earthquake of January 27, 2001. During and after that petitive markets. The first to take action was Chile. As earthquake, telemedicine was used by medical doctors early as 1982, the Chilean government enacted a law in Ahmedabad to provide their services to patients in that dismantled public utilities into three independent remote areas. businesses and allowed private companies to enter the market. A privatization program was put in place that prompted the Chilean private pension funds and Vulnerability Assessment of the Brazilian Electric multinationals to play a greater role. U.S., Spanish, French, Power Infrastructure U.K., and Canadian companies have responded favor- ably to the appeal. A few years later, some Chilean com- Positive Impact of the Power Industry Reforms panies strengthened their financial positions to the extent in South and Central America that they could expand throughout South America in Since the early 1990s, a steady degradation in the reli- countries that had opened their markets to competition. ability of electric power systems in developing coun- The positive experience of Chile has found many tries, especially in South America and Asia, has taken emulators not only in South America, but also in Europe, place. There are several reasons why this has happened. North America, Asia, and the Pacific region (Einhorn First, the expansion of electric power generation assets and Siddiqi 1996; Ilic, Galiana, and Fink 1998; Hunt and high-voltage transmission networks has not kept and Shuttleworth 1996; Tabors 1996). This trend was pace with the significant increase in load demand pushed enforced by the World Bank through a strict policy of higher by robust economic growth and an improvement binding loans to the initiation of reforms. Argentina in the standard of living for the middle class. This is then followed Chile's example and deregulated its power mainly due to the national debt crisis that struck many market in 1992. Peru deregulated in 1993 and Bolivia of these countries in the early 1980s. The crisis pre- and Colombia in 1994. Most Central American coun- vented them from borrowing from the international tries followed in 1997. financial market at a favorable rate to finance capital- Brazil initiated its reform process with the privatiza- intensive investment in their power infrastructure. tion of its distribution companies enforced by Law 8987, Second, climate change, described as the El Niño effect, enacted February 13, 1995. In just two years, 73 percent has also affected these countries. The latter forced the of distribution companies were divested (Rudnick and Colombian Government to decree electric energy rationing others 1998). However, despite this success, several from 1992 to 1993 following a surge in load demand genuine attempts made by the Brazilian government to induced by unusual weather conditions (Rudnick and encompass generation plants in this reform failed. As others 1998). Third, there is poor management of elec- of 2002, only a tiny fraction of the public fleet was sold, tric power assets by public utility companies that showed a not very encouraging prospect for private investors. a lack of interest in bringing productivity and effi- In fact, this meager accomplishment stems from the exis- ciency to international standards. In many developing tence of strong political opposition to the whole reform countries, energy losses in distribution networks have process that gained strength as the energy crisis worsened. been extremely high, productivity of the workforce too The energy crisis in California and the severe drought of low, and unpaid electric energy consumption endemic. 2001 finally prompted the government to interrupt the The latter may vary from 20 percent to 50 percent in privatization process in 2001 and, via the revitalization some utilities in Venezuela and India. committee, revise the general regulatory schemes by To improve capital and human productivity at elec- incorporating market-based rules (Prates 2001; Barroso tric utilities and encourage the national and international and others 2002). Because of its impact on the future private sector to take part in badly needed investments economic growth of the entire region, the Brazilian energy in generating units and the expansion of the transmis- crisis will be examined. Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 281 Fromabroadperspective,itisgenerallyrecognizedthat Figure 18.1 Major regions and river basins in Brazil the reform of the electric power sector in South America has been a great success. Besides a non-negligible capi- tal investment performed by private companies and funds, North competition and stringent quality standards enforced Amazones Rv Belem Manaus by laws in many countries have brought a noticeable Sao Luis improvement in the operational efficiency and reliability Iocirbns Rv North East of electric power systems. This improvement accompa- Receife Slo Fruence Rv nied a decrease in the price of electricity, especially in kngd the wholesale market. According to Rudnick, Varela, and South East Salvador Hogan (1997), Chilean companies were able to cut Brasfilia energy losses by half in seven years while Argentinean Pverola Rv NURUAS companies achieved this level in three years following Belo Horizonte Owale Rv the reforms. The latter dramatically ameliorated work- Cities Sao Paulo Hydro Power Plants force productivity and enhanced the efficiency and reli- MAPU Rio de janeiro Curitiba Thermal Power Plants Tuoce Rv ability of generating units via an improvement of their Rv River Ungual Rv maintenance programs and the usage of better tech- South Porto alegre nologies such as combined-cycle units. This in turn has raised their availability for operation from 47 percent to (Courtesy of Jardini and others, IEEE Power Engineering 70 percent just two years after the reforms. Therefore, it Review 2002, ©2002 IEEE) came as no surprise that the monthly wholesale electricity prices in Argentina went down by more than 40 percent. In Brazil, too, productivity of the newly privatized dis- has not met government expectations. Three main rea- tribution companies rose significantly while energy losses sons may be put forward to explain this lack of enthu- decreased by 9 percent (Prates 2001). siasm: (1) the nonexistence of a competitive market in generation that would have allowed private companies to sign long-term contracts with load-serving entities and therein secure returns of investment; (2) the implemen- Structural Vulnerabilities of the Brazilian Electric tation of large devaluations in the Brazilian currency Power System that have put generation companies at great financial risk, Following the external debt crisis of the 1980s, the since they have to sell electricity in the national currency Brazilian government decided not to invest heavily in the while buying the fuel in U.S. dollars; and (3) the pres- expansion of power systems. No major project was under- ence of a transitional business environment character- taken except the interconnection of the north-northeast ized by precarious rules and regulations. By 1999, this to the south-southeast power grids that were completed unfavorable environment had resulted in the delay or sus- in 1999 (see figure 18.1 for the four main regions of Brazil). pension of many planned projects. In May 2001, one of Following enactment of the reform laws of 1995, 1996, the last companies to suspend all investments was the and 1998 that provided a legal framework for imple- U.S.-basedAES(BusinessNewsAmericas2002).Aswarned mentation of the privatization process and creation of bytheBrazilianplanners,thesecurityofthenationalpower a wholesale market managed by an independent system infrastructure was in jeopardy and the imminence of a operator, the government anticipated that the private shortage in electric energy production loomed ahead. sector would rapidly take the lead in the investment in Another major vulnerability of the Brazilian electric generation and transmission expansion to respond to power system is the lack of diversification in energy a forecasted load growth ranging from 5.3 percent to sources since the power generation system is over- 6.5 percent for the next 10 years (Jardini and others 2002). whelmingly dominated by a host of reservoir-based Unfortunately, the involvement of private companies hydroelectric plants in cascade along major river basins 282 Building Safer Cities: The Future of Disaster Risk of the country. As of early 2001, the installed capacity Fortunately, thanks to the active participation of the of the hydroelectric generating units amounted to 59.6 population, the objectives of rationing were met and GW for a total capacity of 65.7 GW, yielding a percentage major blackouts were avoided. The end of the drought of 90.7 percent. The remaining capacity was shared and the return to more normal rainfall levels during between conventional thermal-power units at a modest the 2001­02 wet season prompted the government to level of 6.4 percent and nuclear units at 2.9 percent. Being declare an end to the rationing in February 28, 2002. aware that the structural vulnerabilities of the power Also, to overcome a glaring lack in electricity production, infrastructure might escalate into a major energy crisis, the government began a new program of generation the government initiated in early 2000 an emergency expansion on July 6, 2001. By 2003 this is expected to plan termed the Thermoelectricity Priority Plan. This plan increase capacity by 26 percent, including 15 thermal advocated the construction by 2003 of 12 GW of gas- plants of 6.4 GW. The U.S. power company AES decided fired power plants mainly fed by the Bolivian-Brazil to reconsider its decision to pull out by contributing to gas pipeline (Jardini and others 2002). Unfortunately, the national effort with a US$1.2 billion investment in this plan came too late to prevent the energy crisis from a set of 3000-MW thermoelectric plants (Business News striking, following a severe drought in early 2001. Americas 2002). However, this good news came with a warning signal that announced the fragility of the Brazilian power system: the warning took the form of Causes of the Brazilian Energy Crisis of 2001­02 a major blackout on January 21, 2002. From 1998 until 2001, the November to April wet season resulted in steadily declining water inflows into the The Blackout of January 21, 2002, in Brazil reservoir systems. Due to a severe drought during the 2000­01 wet season, the water inflows to the northeast/ The blackout spread in the south, southeast, and center- southeast reservoirs were seriously insufficient. By the west regions and caused large economic losses in 10 of end of April 2001, the reservoir was only one-third of the most populated and industrialized states. Business its maximum capacity, well below the recommended News Americas (2002) revealed that the system break- half-capacity level. To avoid depleting the reservoirs and down was triggered by a fault on a transmission line facing blackouts during the dry season, the Brazilian belonging to the company CTEEP. It consisted of a government called for the rapid implementation of an sequence of cascading outages that originated from the electricity rationing plan, effective June 1, 2002. This plan faulted line and spread sequentially from one location advocated an energy reduction of 20 percent for the res- to the other over an increasingly larger region of the net- idential and commercial sectors, including office build- work via the tripping of the protection relays. In the ings, 15 percent to 25 percent for the industrial sector, course of their actions, the latter isolated many electric and 35 percent in public street illumination (McClellan generation plants in a sequence that started with the 2001; Wheatley 2001). Itaipu hydroelectric plant, followed by the CESP Parana's Interestingly, while the northern and southern regions plant and then the Angra nuclear plants. received enough water during this period, they were The details of these cascading events have not yet unable to transfer the excess electric power to the other been made available to a broad public. However, one two energy-starved regions because of a lack of avail- typical scenario could be as follows. The triggering event able capacity in the transmission lines. As result, a great could have resulted from a short-circuit that occurred deal of water was spilled. The situation was exacerbated on one of the network's transmission lines. The relays by the failure of one 1500-MVA transformer in the São of that line sent tripping signals to its circuit breakers. Paulo area and by a delay due to environmental con- Before the faulted line opened, the short-circuit cur- straints of the third high-voltage transmission line that rent was sensed by a certain number of relays located would have connected the Itaipu hydroelectric plant within the region of influence of the fault. The latter located in the south to the national power grid (Jardini region is defined as the union of the regions of vulner- and others 2002). ability of all the relays whose hidden failures are exposed Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 283 by the fault. Consequently, each of these relays may have of well-known multinational automakers such as VW, unnecessarily opened an unfaulty line if it suffered from GM, Fiat, Ford, and Honda announced that it would a hidden failure. Therefore, in addition to the faulted line, maintain its forecasted production of 1.9 million vehi- we may have two, three, or more simultaneous line open- cles for 2001, an increase of 14 percent with respect to ings, usually (but not necessarily) located in the vicinity 2000 (McClellan 2001). In response to the energy short- of the fault. As a result, the power that used to pass through age, many automobile manufacturers are shifting or will the tripped lines found its way through other links in shift to gas, a more abundant energy source. The future the network, which in turn may have overloaded some is therefore not as dim as it might appear at first glance. of them. If any of the overloaded currents is larger than the setting of the overcurrent relays, then the latter will open the associated unfaulty line, putting additional Risk Mitigation of Blackouts in Electric stress on the network. As a domino effect, this sequence Power Systems of line tripping followed by line overloading may propagate throughout the network until either the line New Technologies for Power System Transmission overloading vanishes, or the stability limits or the volt- and Generation age collapse limits are reached. The risk of these cascading events leading to black- The shift toward a competitive electricity market was outs increases dramatically when the power system is made possible by technological advances not only in operated closer to its stability limits, which was precisely telecommunications and computers, but also in power the situation of the Brazilian power system. The system generation, power electronics, fiber optics, and sensors. has small reserves in generation and transmission, making The economies of scale that once characterized large the congestion and equipment overloading likely to occur power plants were no longer valid with the advent of following major disturbances and contingencies. smaller and more efficient generating units in the 1980s. This cleared the way for growing use of small-scale dis- tributed generation technologies such as combustion Impact of the Energy Crisis on the Brazilian Economy turbines, small hydro-turbines, wind turbines, photo- Electricity rationing has had an immediate negative voltaic panels, and fuel cells. These are generally owned impact on Brazilian society, and its effect has changed by independent power producers that compete among the economic and political trends for the years to come-- themselves to provide generation and auxiliary serv- but by how much? To get a sense of the magnitude of ices at both the distribution and the transmission levels. the disaster, McClellan (2001) referred to a study that The power electronics technology is not a newcomer anticipates the loss of 800,000 jobs if the decrease to power systems. In the mid-1970s, Static Var Com- in energy is pursued. Another study revealed that pensators were already being utilized for voltage con- economists contemplate a decline in economic growth trol and to a certain extent for stability control. Also, from a previously projected 4.5 percent to less than HVDC converter stations and high-voltage DC lines have 3 percent. In a poll conducted in São Paulo, 65 percent been incorporated into the transmission grid to allow of manufacturers said that they would shrink or post- for the transfer of large amounts of power over long dis- pone investments and 25 percent are contemplating lay- tances and to allow the coupling of two power systems offs. Wheatley (2001) mentioned that on May 25, 2001, with different frequency regulations (i.e., the East-West Flextronics International Ltd., a Singapore-based com- European connections as indicated in the IEEE Sub- pany, announced the cancellation of an $85 million committee on International Practices, 1996). However, addition to its industrial plant at Sorocaba, a suburb of the 1990s have witnessed the advent of a broad range São Paulo, thereby erasing the prospect of 500 new jobs. of power-electronic apparatuses, termed Flexible AC Fortunately, the impact of the crisis on the automobile TransmissionSystem(FACTS)devicesbyHingurani(1993). industry, an important component of Brazilian GDP and These devices meet the urgent need of utility compa- exports, seems to be minor since Anfavea, a consortium nies to increase the loadability of their transmission 284 Building Safer Cities: The Future of Disaster Risk network following the environmental regulations imposed for damping inter-area oscillations and steering the around the world. Issued in the late 1980s, these regu- system away from transient instabilities or voltage lations impose stringent limitations on the construction collapse. of new overhead transmission lines, which in turn increase the stress put on the existing transmission network. As Enhancing the Efficiency of a Distribution System a result, the power systems are being pushed dangerously via Distribution Automation closer to their operating limits, making the occurrence of major disturbances more likely. This is substantiated Distribution automation is being incorporated in the by the growing number of incidents recently reported distribution management system of an increasing number in the literature such as voltage collapses, angular insta- of utilities worldwide (His and Chen 1998; Su and bilities, and cascading outages leading to brownouts or others 2000; Humphreys 1998; Choi and Kim 2000). blackouts (IEEE Working Group on Voltage Stability Typically, distribution automation consists of a collection 1990; Van Cutsem and Vournas 1998; IEEE Working of functions that act either on the utility feeders and equip- Group on Special Stability Controls 1990). ment or on the customer appliances and metering system Driven by the rapid development of the microprocessor (Williams and Walden 1994; IEEE Communication Pro- and fiber optic technology, protection and sensors are tocols Subcommittee 1995). The feeder automation also experiencing technological improvements. For exam- functions seek to improve the performance of the dis- ple, computer relays are replacing the old electro-mag- tribution system through switching and other control netic and electronic protection devices as they become actions. One common control action is the switching of less costly (Phadke and Thorp 1988). One important circuit breakers to reconfigure the distribution network feature that makes them attractive is a higher reliability in order to transfer loads among feeders to minimize net- due to a self-checking and self-monitoring capability. work losses while maintaining phase imbalance at a rea- Also, in the event of a detected defect, they can take them- sonable level during peak and off-peak periods. This selves out of service after informing a central control switching may also be initiated to isolate the smallest system. Another interesting feature is their ability to adapt possible segments of the network during fault conditions, to different operating conditions of a power system by knownasfaultdetection,isolation,andrestorationschemes. changing their settings, yielding the so-called adaptive These network reconfigurations must be carried out relaying. Algorithms may also be implemented on their regularly due to the diversity and time-varying charac- microprocessors to perform filtering, fault location, and teristics of the load (Huddleston 1990); a configuration various network calculations for control purposes (i.e., that is optimized at peak load may do poorly at off-peak damping oscillations). Fiber optics are advocated as load and vice versa. the most effective medium to transfer a large volume of Another contribution of distribution automation is data to a central location without any electro-magnetic in the area of power quality. Research in power quality interference. enhancement has resulted in the development of a Another computer device that is attracting growing variety of devices to mitigate the effects of transients on attention is the microprocessor-based phasor measure- customer equipment and utility transmission and dis- ment unit, or PMU (Phadke and Thorp 1988). Invented tribution networks. Transients include voltage sags and at Virginia Tech in the mid-1980s, they are presently swells, flickers, surges, and harmonics. They may cause installed on the transmission grids of many electric util- interruption of service or have a damaging effect on util- ities in the United States. These utilities are interested ity and customer equipment. Presently, the focus is on in the ability of the PMUs to measure and time tag with finding solutions for voltage magnifications and voltage great accuracy all the voltage and current magnitudes and sags (Adams and others 1998; Middlekauff and Collins phase angles of the buses where they are located. Because 1998). Induced by capacitor switching, voltage mag- they sample at a sufficiently high rate the various non- nifications may cause the loss of audio signals and a loss linear oscillations that can take place in the system, the of power to video signals of local television stations. PMUs are being incorporated in various control schemes Several devices that can cope with these problems are Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 285 being developed and marketed (Adams and others 1998; the efficiency and the quality of service of the trans- Nelson and others 1995). mission and distribution grids. A priority list of actions includes the following: · Implementation of a plan of defense for the surviv- Overcoming the Transmission Capacity Shortage ability of electric power network transmission fol- via Distributed Resources lowing a major disturbance. This defense plan aims According to the U.S.-based Electric Power Research at confining the cascading failures to a small region Institute (EPRI), it is anticipated that by 2010 more than via the breakup of the transmission grid into islands. 20 percent of U.S. growth in electricity will be sup- This breakup should achieve, as far as possible, a bal- plied by distributed resources. This trend is also fore- ance between power generation and load demand seen in Europe and many developing countries. This is in every island. It should be carried out by a collec- a sizable amount of newly installed power capacity that tion of special protection schemes having objectives would put a significant strain on the distribution net- that range from the prevention of angular and volt- works since many of these units will be connected along age instabilities to the relief of overloaded lines and distribution feeders, down to the low voltage levels. transformers. To meet these requirements, protection Obviously, the design of this network should be schemes may initiate generation and load shedding rethought completely (Hadjsaid and others 1999). A and line switching across the network. Interestingly, shift toward a more meshed network is probably required a pioneer implementation of such a defense was for stability and security. The redesign should also recently carried out by the Taiwan Power Company account for the inclusion of various distribution automa- following a series of blackouts in Taiwan, the last one tion devices such as fault detection, isolation and restora- occurring July 29, 1999 (Hsiao and others 2002). tion; automatic voltage controllers; distribution static · Implementation of the fault detection, isolation, compensators; and dynamic voltage restorers for power and restoration scheme for the distribution system quality purposes. Monitoring and control of the dis- to decrease the time of recovery after a blackout. One tribution system, including the use of automated important component of this computer-based scheme mapping facilities, would require the design and instal- is the Outage Management System (OMS). Currently lation of a distributed architecture of remotely dis- implemented by a growing number of U.S. utilities, tributed control centers spread across the service areas. the OMS makes use of the geographical information These centers would communicate among each other system to map the location of all devices that com- and share a common database through fast communi- pose the distribution network and achieve func- cation links such as fiber optics, radio communica- tions such as customer call management, connected tions, and low orbit satellites. The design of such circuit models, switching, outage/regulatory, and distributed computer architecture has already been remote-user access (Blew 2001). envisioned by researchers in Taiwan, yet without includ- · Implementation of computer-aided software pro- ing a large number of distributed resources and auto- grams at control centers that seek the best topology matic devices (Lu and others 1997; His and Chen 1998; of the distribution network associated with mini- Su and others 2000). mum electric losses. These losses constitute a siz- able fraction of the total losses in an electric power system. Their minimization will result in millions of Technology Deployment for Disaster Risk dollars of savings per year while simultaneously Management in Electric Power Systems improving the quality of service of the system. The deployment of the technologies previously out- lined, coupled with advanced control schemes, has the Bibliography ability to enhance the resiliency of a power system to ABC News. 1999. "France, Austria Battered Again," Paris, Dec. catastrophic failures and decrease the time of recovery 28. Available from . 286 Building Safer Cities: The Future of Disaster Risk Adams, R. A., S.W. Middlekauff, E. H. Camm, and J.A. McGee. Hadjsaid, N., J.-F. Canard, and F. Dumas. 1999. "Dispersed Gen- 1998. "Solving Customer Power Quality Problems Due to Volt- eration Impact on Distribution Networks." IEEE Computer age Magnification." Paper No. PE-384-PWRD-0-11-1997, Applications in Power 12 (2): 22­28. IEEE Transactions on Power Delivery 11. Harris. 2002. "India: Telemedicine's Great New Frontier: An Adibi, M. M. 2000. Power System Restoration: Methodology and Indigenous Technology Effort is Wiring Up its Healthcare Implementation Strategies. IEEE Press. System." IEEE Spectrum, April: 16­17. Barkovich, B. R. and D.V. Hawk. 1996. "Charting a New Course Hingorani, N. G. 1993. "Flexible AC Transmission." IEEE Spec- in California." IEEE Spectrum July: 26­31. trum April: 40­45. Barroso, L. A. N., M. V. F. Pereira, R. Kelman, P. Lino, and J. His, P.-H. and S.-L. Chen. 1998. "Distribution Automation: Com- Rosenblat. 2002. "Can Brazil Learn from California?" IEEE munication Infrastructure." IEEE Transactions on Power Deliv- Power Engineering Review, August. ery 13 (3): 728­734. Blew, D.S. 2001. "Outage Management Systems: Surviving the Hsiao, T.-Y., C.-N. Lu, and Y. Liu. 2002. "Defense Plan Design Implementation." Proceedings of the Workshop on Mitigation of in a Longitudinal Power System." Proceedings of the 14th the Vulnerabilities of Critical Infrastructures to Catastrophic Fail- Power Systems Computer Computation, Paper 5 of Session 25, ures, Alexandria Research Institute, Virginia Tech, Alexandria, Seville, June 24­28. Virginia, September 10­11. Available from . 1990. "Reconfiguration Algorithm for Minimizing Losses in Business News Americas. 2002. "Latin America Electric Power Radial Electric Distribution Systems." Electric Power Systems Weekly Wrap-Regional." Financial Times Information, January Research 18: 57­66. 25. Available from . Humphreys, S. 1998. "Substation Automation Systems in Review." Cauley, G., P. Hirsch, A. Vojdani, T. Saxton, and F. Cleveland. IEEE Computer Applications in Power April: 24­30. 1996. "Information Network Supports Open Access." IEEE Hunt, S. and S. Shuttleworth. 1996. "Unlocking the Grid." Computer Applications in Power July: 12­19. IEEE Spectrum, July: 20­25. CEOS. 2002. Working Group on Information Systems and Ser- IDRC (International Development Research Centre of Canada). vices. Available from . 2002. Ottawa, Canada. . Choi, J.-H. and J.C. Kim. 2001. "The Online Voltage Control of IEEE (Institute of Electrical and Electronics Engineers) Com- ULTC Transformer for Distribution Voltage Regulation." munication Protocols Subcommittee. 1995. "IEEE Tutorial Electric Power and Energy Systems 23: 91­98. Course: Communication Protocols." IEEE Power Engineering Clendenin, M. 2001. "After Plague of Blackouts, Hsinchu Hears Society, Report 95 TP 103. Vow of Better Power in 2001." Electronic Engineering Times, IEEE (Institute of Electrical and Electronics Engineers) Sub- CMP Media, Inc., Jan. 8. committee on International Practices. 1996. "East and Clinton, W. J. 1996. "Executive Order 13010-Critical Infra- Central European Policy on Electricity Infrastructure, Inter- structure Protection." The White House, July 17. Available connections and Electricity Exchanges." IEEE Report 96 TP from . 113-0. Dana, P. H. 2000. "Global Position System Overview." Colorado IEEE (Institute of Electrical and Electronics Engineers). 1990. University. Available from . and Frequency Domain Methods for System Dynamic Performance. Economist, The. 2002. "A Struggle for Power." February 2. IEEE Special Publication 90TH0292-3-PWR. Einhorn, M. and Siddiqi, R., eds. 1996. ElectricityTransmissionPric- IEEE (Institute of Electrical and Electronics Engineers) 1990. ing and Technology. New York: Kluwer Academic Publishers. Working Group on Voltage Stability. Voltage Stability of Power EPRI (Electric Power Research Institute).1999a. "Power Deliv- Systems: Concepts, Analytical Tools, and Industry Experience. ery in the 21st Century," EPRI Journal, Summer issue: 19­25. IEEE Special Publication 90TH0358-2-PWR. EPRI (Electric Power Research Institute). 1999b. "Utility Com- Ilic, M., F. Galiana, and L. Fink, eds. 1998. Power Systems Restruc- munications Go Into Orbit," EPRI Journal, Fall issue: 18­27. turing: Engineering and Economics. New York: Kluwer Acade- Graham, J. F. 1995. "Space Exploration From Talisman of the mic Publishers. Past to Gateway for the Future." NASA. Internet Bazaar. 2002. Available from . Mitigating the Vulnerability of Critical Infrastructure in Developing Countries 287 IPCC (The Intergovernmental Panel on Climate Change ) Work- Neumann, P.G. 1995. Computer-Related Risks. Boston: Addison- ing Group II. 2001. "Climate Change 2001: Impacts, Adap- Wesley Publishing Co. tation and Vulnerability." Available from . Infrastructures." In The Role of Computer Security in Protecting Janssens, N. and A. Kamagate. 2002. "Loop Flows in a Ring AC U.S. Infrastructures. Hearing, 105th Congress, 1st Session, No. Power System," Proceedings of the 14th Power Systems Compu- 33: 64­99. tation Conference. Paper 2 of Session 09, Seville, June 24­28. Phadke, A.G. and J.S. Thorp. 1988. Computer Relaying for Jardini, J.A., D. S. Ramos, J.S.C. Martini, L. B. Reis, and C.M.V. Power Systems. New Jersey: John Wiley and Sons. Tahan. 2002. "Brazilian Energy Crisis," IEEE Power Engineer- Phombeah, G. 2000. "Feeding Hungry Minds in Africa." BBC ing Review, April: 21­24. News, July 27. Available from. aging the Natural Disasters from Space Technology Inputs." Prates, J. P. 2001. "Brazil's Energy Crisis Complicates Progress Acta Astronautica 40 (2­8): 291­325. in Gas, Power Markets, but Outlook Brightening." Oil & Gas Koshcheev, L. A. 2001. "Basic Principles of Interstate Electrical Journal 99 (42): 77. Power Links Organization in Northeast Asia Region." Pro- Ravisankar, R. 2002. "Moving Up the Value Chain: From Ser- ceedings of the Workshop on Power Grid Interconnection in vices to Products," July 29, i-flex solutions Ltd. Available from Northeast Asia. May, Beijing, China. . Latin America Regional Reports: Brazil. 2002. "State Intervenes Rediff News. 2002. "Major Power Breakdown in Western India." in Troubled Sector." Latin American Newsletters, Ltd. Section July 30. Available from . Lu, C. N., K.T. Chen, M.C. Lin, and Y.T. Lin. 1997. "Perfor- Rediff News. 2002. "Indiscipline by Madhya Pradesh Caused mance Assessment of an Integrated Distribution SCADA- Grid Collapse: PGCIL." July 30. . 971­978. Rudnick, H. 1996. "Pioneering Electricity Reform in South Amer- Mason, C. 1990. "Software Problem Cripples AT&T Long-Distance ica." IEEE Spectrum August: 38­44. Network," Telephony January 22: 10. Rudnick, H., R. Varela, and W. Hogan. 1997. "Evaluation of Alter- McClellan, B. 2001. "In the Dark: Brazil's Auto Industry Strug- natives for Power System Coordination and Pooling in a Com- gles with Energy Crunch." ManufacturingSouth-America. Intertec petitive Environment." IEEE Trans. on Power Systems 12 (2): Publishing Corporation, August. 605­613. Middlekauff, S. W. and E. R. Collins. 1998. "System and Cus- Rudnick, H., N. de Franco, T.J. Hammons, L.V. Sbertoli, C. tomer Impact: Considerations for Series Custom Power Devices." Chahin, J.G.A. do Nascimento, J.W. Marangon Lima, and IEEE Transactions on Power Delivery 13 (1): 278­282, January. P. Gray. 1998. "Latin America Deregulation Processes." IEEE NASDA (The National Space Development Agency of Japan). Power Engineering Review December: 10­23. 2002. "Earth Observation satellite." Available from igation System." IEEE Spectrum (April). NASSCOM (National Association of Software and Service Com- Su, C.-L., C.-N. Lu, and M.-C. Lin. 2000. "Wide Area Network panies). 2002. July 29. Available from . Electric Power and Energy Systems 22: 9­14. Naushad. 2002. Available from . Sugarman, R. 1978. "New York City's Blackout: A $350 Million Nelson, R.J., N.H. Woodley, and E.M. Gulachenski. 1995. Drain." IEEE Spectrum 15 (11): 44­46. "Requirements for Dynamic Voltage Restoration to Relieve Dis- Tabors, R. T. 1996. "Lessons from the UK and Norway." IEEE tribution System Voltage Sags." Proceedings of the American Spectrum 33(8): 45­49. Power Conference, 57th Annual Meeting, Chicago, IL. Task Force on Harmonics Modeling and Simulation. 1996. "Mod- NERC (North American Electric Reliability Council). 1988. eling and Simulation of the Propagation of Harmonics in Elec- "NERC Disturbance Reports, 1984­1988." New Jersey. tric Power Networks." Part 1s and 2, IEEE Transactions on Power Available from . Delivery 11 (1): 452­474. 288 Building Safer Cities: The Future of Disaster Risk Thorp, J.S., A.G. Phadke, S.H. Horowitz, and S. Tamronglak. Wheatley, J. 2001. "What's Blacking out Brazil's Growth." Busi- 1998. "Anatomy of Power Systyem Disturbances: Impor- ness Week International Editions. June 11. The McGraw-Hill Co. tance Sampling." Electrical Power & Energy Systems 20(2): Available from . 147­152. Williams, B. R. and D. G. Walden. 1994. "Changing the Momen- Topfer, K. and V. Ramanathan. 2002. "Impact Study--The Asian tum: Distribution Automation Strategy for the Future." IEEE Brown Cloud: Climate and Other Environmental Impacts." Computer Applications in Power (July): 16­21. The Center for Clouds, Chemistry, and Climate. UNEP Wilson, G. L. and P. Zarakas. 1978. "Anatomy of a Blackout." report, April 8. Available from . Wood, H. 2002. "The Use of Earth Observing Satellites for Hazard Travis, P. 1990. "Why the AT&T Network Crashed." Telephony Support: Assessments and Scenarios." Final Report of the January 22: 11. CEOS Disaster Management Support Group. Available from USAID (United States Agency for International Development). . 2001. "Hurricanes in Central America and the Caribbean." World Bank. 1998. "Meeting India's Future Power Needs: Planning USAID Status Reports (February 6). Available from . D.C. Available from . USGS (United States Geological Survey). 1998. "Central World Energy Council. 1999. "The Challenge of Rural Energy America Disaster Overview." USGS Center for Integration of Poverty in Developing Countries." Available from http://www. Natural Disaster Information (November 17). Available worldenergy.org/wec-geis/publications/default/launches/ from . World Markets Research Center. 2001. "World Infrastructure Van Cutsem, T. and C. Vournas. 1998. Voltage Stability of and Development." May. Available from . Chapter 19 Damage to and Vulnerability of Industry in the 1999 Kocaeli, Turkey, Earthquake Mustafa Erdik and Eser Durukal Over the past 10 years, Turkey has been hit by several business units, collapsed or were heavily damaged. moderate to large earthquakes that caused significant Another 220,000 housing units and 21,000 small loss of life and property. These took place in: Erzincan, business units experienced lesser degrees of damage. 1992; Dinar, 1995; Adana-Ceyhan, 1998; and Kocaeli Widespread building collapse in the two earthquakes and Duzce, 1999. Adana-Ceyhan and Kocaeli are the resulted in a substantial number of casualties. Deaths most industrialized regions in Turkey, and the earth- numbering 18,373 were recorded, and 48,901 people quakes of 1998 and 1999 resulted in extensive losses were hospitalized for injuries; 40 percent of them will to industry in these regions. be left permanently disabled. On August 17, 1999 a magnitude MW 7.4 (MW is The two major earthquakes that took place in 1999, moment magnitude) earthquake struck Turkey's north- namely the MW 7.4 Kocaeli, Turkey and the MW 7.6 western Kocaeli and Sakarya provinces, a densely pop- Chi-Chi, Taiwan (China), earthquakes, caused compa- ulated region in the industrial heartland. The earthquake rable economic and insured losses. In the Kocaeli earth- nucleated at a depth of about 15 kilometers, some 10 quake, the economic losses were estimated to have been kilometers east of the town of Gölcük. The earthquake $10­40 billion, while insured losses were estimated at was associated with a 120-kilometer rupture involving $550­750 million. The Chi-Chi earthquake resulted four distinct fault segments on the northernmost strand in economic losses of $8­14 billion and insured losses of the western extension of the 1300-kilometer-long of $500­$850 million (Johnson 2000). It is worth noting, North Anatolian fault system. Predominantly right- however, that there is an order of magnitude difference lateral strike-slip offsets were in the range of three to in the human losses in these earthquakes. The Kocaeli four meters over a significant length of the fault. Another earthquake resulted in more than 18,000 deaths and segment at the eastern end of the fault break ruptured nearly 50,000 injuries, while in Taiwan there were 2,405 on November 12, producing the MW 7.2 Duzce earth- deaths and 10,718 injuries. quake. The August 17, 1999 (MW 7.4) Kocaeli and The Kocaeli earthquake is considered the largest event November 12, 1999 (MW 7.2) Duzce earthquakes to have damaged an industrialized area since the 1906 resulted when a wedge of continental crust known as San Francisco and 1923 Tokyo earthquakes. In the earth- the Anatolian Block was squeezed between the Arabian quake, 70 percent of total insured losses related to direct and Eurasian plates. This motion was accommodated damage and 30 percent was due to business interrup- by two major strike-slip faults: the North and East tion. Estimations in U.S. dollars (all dollar amounts in Anatolian faults. this paper are U.S. dollars) by the insurance industry The 1999 Kocaeli and Duzce earthquakes caused of total insured losses were in the order of $1.5­$3.5 considerable damage to residential and commercial billion (RMS 1999) as compared to the $550­$750 buildings, public facilities, and infrastructure, and caused million estimated to have been paid by the industry significant casualties in an area 20 kilometers by 200 (Johnson 2000). kilometers. The number of condemned buildings after The epicenter of the 1999 Kocaeli earthquake was the earthquakes totaled 23,400. Some 16,400 of these, the main site of Turkey's heavy industry. Major indus- encompassing 93,000 housing units and 15,000 small tries located there include: automobile manufacturing; 289 290 Building Safer Cities: The Future of Disaster Risk petrochemicals; motor and railway vehicle manufacture losses were covered by insurance. Payments of claims and repair; basic metal works; production and weaving were reported to have totaled $600 to $800 million. The of synthetic fibers and yarns; paint and lacquer pro- State Planning Organization estimates a loss of $880 duction; tire manufacturing; paper mills; steel pipe pro- million just for the 19 state-owned enterprises located duction; pharmaceuticals; sugar processing; cement in the region (mainly in Tupras, Tuvasas, Igsas, Petkim, production; power plants; and tourism. Seka, and Asil Celik). The State Planning Organization The region affected by the earthquake is geographi- also estimates that the loss of business in these indus- cally extensive, economically dynamic, and the indus- tries may have amounted to $632 million. The tourism trial heartland of Turkey. The four districts most severely industry (based in Yalova) was virtually destroyed and affected (Kocaeli, Sakarya, Bolu, and Yalova) contribute has yet to pick up even three years after the earthquake. more than 7 percent of the country's GDP and 14 percent A fundamental regional restructuring in the tourism of industrial value-added. Per capita income is almost industry may be needed. double the national average. With only 4 percent of Rahnama and Morrow (2000) note that older, heavy the nation's population, the region contributes more than industrial facilities, especially those with taller struc- 16 percent of budget revenues. The districts immediately tures, that partially to totally collapsed, were more surrounding the area (Bursa, Eskisehir, and Istanbul) affected by the earthquake than newer facilities. It was were indirectly affected because of their close economic observed that any type and quality of anchorage improved linkages, since industries and small businesses supply the performance of machines and equipment, except services and material inputs to each other's production very sensitive equipment such as assembly line sensors processes. The greater region now realizes that it shares in the automotive industry and rotary kilns in cement a seismic risk and faces magnified uncertainty for the plants. Losses associated with business interruption were future. Taking all seven cities together, the wider earth- more severe for these types of facilities. For light indus- quake region accounts for 35 percent of national GDP trial facilities, building damage turned out to be the and nearly half of the nation's industrial output. Build- primary reason for direct and indirect losses. As was the ing losses reportedly amounted to $5 billion. Damage case in the 1998 Adana-Ceyhan earthquake, the poor to lifelines is estimated to be some $1 billion. Indus- performance of precast concrete structures was observed. trial facilities and small business losses are $2 billion For refineries and other chemical processing facilities, and $1 billion, respectively. If we assume that indirect nonbuilding structures turned out to be the most vul- socioeconomic losses will be as much as direct physi- nerable, with tanks being the most susceptible to cal losses, the total loss figure is in the vicinity of $16 earthquake and fire damage. It was observed that damage billion (about 7 percent of Turkey's GDP). Most indus- to industrial facilities was more severe and extensive trial losses were covered by insurance. than that seen in earthquakes with similar peak ground Private and public sector estimates of damage to acceleration levels. This observation was attributed to industry as a whole range from $1.1 to $4.5 billion. The the duration and long-period ground motion of the earth- value-added loss in manufacturing is estimated at $600 quake (MCEER 2000). Most industrial facilities dam- to 700 million. The State Planning Industry estimated aged by this earthquake were within 10 kilometers of that value-added losses stemming from damage to indus- the fault rupture and in intensity zone IX. try were $700 million. The losses may have resulted in In general, the earthquake damage at industrial facil- a 1.6 percent decline in the growth of the production ities in Turkey was not significantly different than that sector in Turkey. Other sources put this loss figure as observed in other earthquakes worldwide. Large storage high as $2 billion. According to the Kocaeli Chamber tanks,pipelines,transmissionlines,andprecisionmachin- of Industry, 214 enterprises (about 19 percent of all enter- ery are generally susceptible to damage from earthquakes. prises in the province) reported significant damage total- Due to the high relative value of contents, their vulner- ing $2.5 billion in capital losses. Many major facilities ability and dependence on structural performance are facedextensivebusinessinterruptions;however,thebiggest key in assessing loss potential, especially for heavy loss was that of qualified manpower. Most industrial manufacturing facilities. Port and harbor facilities are Damage to and Vulnerability of Industry in the 1999 Kocaeli, Turkey, Earthquake 291 particularly susceptible to sub-marine landslides or compensation. Partial quantification of these indirect eco- ground settlement due to liquefaction that may occur nomic losses can be found in ATC-25 (1991). during earthquakes. In addition, all processes that involve An industrial facility consists of many integrated com- a substantial risk of explosion, such as those in the petro- ponents and processes. As such, operation of a facility chemical industry and processes involving molten metal, depends upon the performance of its critical components. should be examined carefully. The greatest risk from an earthquake is that to life safety. Earthquake vulnerability is the measure of damage Building code requirements in most counties, including a building or structure is likely to experience when sub- Turkey, are written with the objective of protecting lives. jected to ground shaking of a specified intensity. The A building is allowed to be damaged, but it should not dynamic response of a structure to ground shaking is collapseandpeopleshouldbeabletoevacuateevenunder a complex behavior that is dependent upon a number extremeconditions.However,inlargeearthquakes,indus- of inter-related parameters that are often difficult, if trial buildings and related machinery and equipment not impossible, to predict precisely. These include: the damaged may be costly to repair and there may be addi- exact character of the ground shaking that the build- tional damage from fire and chemical spills. Since most ing will experience; the extent to which the structure revenue generated by industrial facilities is related to the will be excited by and respond to the ground shaking; products and services they provide rather than the the strength of the materials in the structure; the qual- physical assets of the company, any significant interrup- ity of construction and condition of individual struc- tion in the production of these goods and services will tural elements; the interaction of the structural and have an adverse effect on business. The risk of business nonstructural elements of the industrial facility; the interruption is a critical economic reason for controlling weight of contents in the facility at the time of the earth- earthquake and post-earthquake damage. As such, the quake; and other factors. Most of these factors can be design (or seismic retrofit) of industrial facilities should estimated, but never precisely known. As a result, it is preferably be based on performance-based methodolo- typically necessary to define vulnerability functions for gies with the objective of controlling structural and non- buildings within levels of confidence. structural damage. In addition to physical vulnerabilities, the socioeco- nomic vulnerabilities of industrial facilities need to be Bibliography assessed in terms of casualties, social disruption, and ATC-25 (Applied Technology Council). 1991. "Seismic Vulner- economic losses. Casualties in earthquakes arise mostly ability and Impact of Disruption of Lifelines in the Cotermi- from structural collapse and other collateral hazards. nous United States, ATC-25." Redwood City, California. Lethality per collapsed building can be estimated by a Johnson, Laurie A. 2000. "Earthquake Loss Modeling Applica- combination of factors representing the number of people tions for Disaster Management: Lessons from the 1999 Turkey, per building, occupancy at the time of the earthquake, Greece and Taiwan Earthquakes." Proceedings of EuroCon- occupants trapped by collapse, mortality at collapse, ference 2000 on Global Change and Catastrophe Risk Manage- and mortality post-collapse. Lethality for collateral haz- ment: Earthquake Risk in Europe. Laxenburg, Austria. Available ards is difficult to generalize and may require facility- from . specific assessments. MCEER (Multidisciplinary Center for Earthquake Engineering It is generally known that losses due to collateral haz- Research). 2000. "The Marmara, Turkey Earthquake of August ards and indirect economic losses constitute a major por- 17, 1999: Reconnaissance Report." Charles Scawthorn, ed. tion of total earthquake losses in industrial systems. University of Buffalo, New York. Indirect economic losses arise from shutting down dam- Rahnama Mohsen, and Guy Morrow. 2000. "Performance of aged facilities and include: production and sales lost by Industrial Facilities in the August 17, 1999 Izmit Earthquake." firms in damaged buildings; production and sales lost Proceedings of the 12th World Conference on Earthquake Engi- by firms unable to get supplies from other damaged facil- neering. On CD. Paper No: 2851, Auckland, New Zealand. ities; production and sales lost by firms due to damaged RMS (Risk Management Solutions). 1999. "Event Report, Kocaeli, lifelines; lost tax revenue; and increased unemployment Turkey Earthquake." San Francisco, California. Chapter 20 The Behavior of Retrofitted Buildings During Earthquakes: New Technologies Mikayel Melkumyan Five years have passed since buildings were retrofitted consequences of earthquakes of different intensities, under the World Bank­financed Armenia Earthquake and to justify the possibility of the practical applica- Zone Reconstruction Project. During this period sev- tion of new technologies to critical facilities. More specif- eral earthquakes of varying intensities occurred in north- ically, the study required the following: ern Armenia. Intense usage of these buildings has been 1. Visual inspections of three isolated buildings in taking place as all are occupied. Considering these fac- Vanadzor (one building with base isolation and two tors, there are lessons that can be drawn from the retro- buildings with isolated upper floors). Possible damage fitting, which uses new technologies, to learn whether in isolation systems and to structures below and above these technologies can be applied to critical facilities. was to be revealed and assessed. Different types of laminated rubber bearings were 2. Instrumental investigation of dynamic character- used to improve earthquake resistance in numerous istics of the buildings. The values of vibration periods buildings. Following the earthquakes, these bearings and damping factors were to be obtained and compared were examined and their performance was evaluated. with the design values at small deformations. Lessons learned from the experience should now be 3. Based on visual inspection and instrumental passed along to decisionmakers responsible for con- investigations, assess the technical conditions of the struction policies. Also, if the new technologies are buildings. deemed appropriate, they should be applied to critical 4. Survey the tenants regarding the behavior of build- facilities. The study of these retrofitting measures also ings during the various earthquakes that occurred in included analysis of the sustainability and effective- the last five years. ness of the new technologies for critical facilities such 5. Based on the findings, assess the behavior of build- as schools and hospitals. ings during earthquakes and draw lessons from retro- On the basis of the study, a report was prepared on fitting using new technologies. the findings outlined below. The conclusions reached 6. Justify the possibility of the practical applica- and recommendations made can be supported by the tion and effectiveness of new technologies to critical data received and working papers submitted that are facilities. kept in a permanent file. 7. Prepare recommendations to influence decision- makers in construction policy for further implementa- tion of new technologies in Armenia and other countries Project Scope where similar projects are financed by the World Bank and other international financial institutions. The objectives of the assignment were to assess the behav- In addition to the above scope, a similar study was ior of retrofitted buildings, to learn lessons based on the carried out for one building in Spitak. This paper was written with the participation of A. Azaryan, T. Nersesyan, S. Saroyan, A. Avagyan, G. Goulanyan, A. Ghevondyan, A. Karapetyan, G. Voskanyan, K. Avanesova, and M. Begumyan. 293 294 Building Safer Cities: The Future of Disaster Risk Visual and Instrumental Inspection of Three is going on with the purpose of establishing a museum Buildings in Vanadzor and One in Spitak: dedicated to the 1988 Spitak earthquake. This project Assessment of the Technical Conditions was initiated by the Armenian Association for Earthquake Engineering. The three buildings in Vanadzor were strengthened A visual inspection was carried out for three build- within the framework of the "strengthening" subcom- ings in Vanadzor (one building with base isolation and ponent of the World Bank-financed Earthquake Zone two buildings with Additional Isolated Upper Floor Reconstruction Project. "This subcomponent funded (AIUF)) in addition to the building with base isolation the strengthening of structures damaged and weakened in Spitak. Hairline cracks were found in some reinforced- by the 1988 earthquake that could be made habitable concrete elements below and above the isolation system through retrofitted structural reinforcing or seismic iso- and some violations of the rules of exploitation were lation systems. Also this subcomponent has established revealed in the building Yerevanyan #149. While the new technologies for the strengthening of existing struc- cracks are not dangerous for the building, the violation tures, notably new techniques for installing seismic iso- of rules of exploitation in the form of filled seismic gap lation systems in new and existing occupied buildings. by stones, mortar, soil, and wood is absolutely inadmis- The project is the worldwide pioneer for several tech- sible. All filled materials that restrict the displacement of niques, and two project pilot structures have attracted the isolation system must be removed immediately. international professional attention by establishing appro- The roof layers and some parts of exterior nonbear- priate low technology methods for the strengthening ing structures in the building Isahakyan #4 were sig- of existing structures."1 nificantly damaged due to violation of the rules of In the proceedings of the 7th International Seminar exploitation of the building by tenants. As a result, the on Seismic Isolation, Passive Energy Dissipation and seismic isolators of AIUF were immediately influenced Active Control of Vibration of Structures (Assisi, Italy, by atmospheric precipitations, seriously deteriorating 2­5 October, 2001) it is mentioned: the surface of six isolators. Also in buildings Isahakyan #4 and Cherkassi #1, the seismic isolators are not appro- As far as the ordinary apartment buildings are con- cerned, the number of new applications of the Innova- priately protected from low temperatures. The roof tive Antiseismic (IAS) techniques ... is particularly large and nonbearing exterior structures of the Isahakyan in Japan and the P.R. China and Armenia..." (page xxiii); #4 building must be repaired and six seismic isolators "...with regard to the interest in the IAS techniques ...it must be replaced. All isolators in both Isahakyan #4 and is now extending to new countries like, for instance, Turkey, Cherkassi #1 buildings must be taken into special Iran and India. Some other countries are beginning to jackets due to light insulation materials. The estimated follow the excellent example of Armenia..., where seis- mic isolators are locally manufactured also for foreign mar- cost for jacketing, replacement of isolators, and repair kets, like Syria" (page xxv); "Very interesting on-site tests work for both buildings is about US$15,000. of ... an existing bank building at Irkutsk-City in Russia, No damage was revealed in reinforced-concrete struc- retrofitted by applying the technology invented by Prof. tures below and above the isolation system of Shahumyan M. Melkumyan in Armenia..." (page xxv). building. Seismic isolators were also in a normal tech- The building inspected in Spitak was constructed nical state. under the terms of the World Bank project. It is a four- Instrumental investigations of all four buildings were story apartment building on Shahumyan Street with carried out. The rubber hardness and dynamic charac- reinforced-concrete bearing walls. This is the first newly teristics of buildings were measured. The increase in constructed building in Armenia using the seismic (base) rubber hardness due to low outside temperatures was isolation system. Following the 1988 Spitak earthquake, observed in the Isahakyan #4 and Cherkassi #1 build- no buildings over two stories were constructed. With ings. The dynamic characteristics of the Yerevanyan #149 base isolation, however, a four-story building was con- and Shahumyan buildings fully correspond to the design structed for the first time after the earthquake. Presently, values of total initial stiffness of isolation systems. The where the isolation system is located, renovation work dynamiccharacteristicsoftheIsahakyan#4andCherkassi#1 The Behavior of Retrofitted Buildings During Earthquakes: New Technologies 295 buildings are 1.4 times less on average than the design Table 20.1 Short form of the European Microseismic values as a result of increased rubber hardness. The pro- Scale EMS-98 posedjacketingofseismicisolatorsofAIUFinbothbuild- EMS Definition of Description of typical observed effects ings will help to avoid disturbing their tuning in order intensity earthquake (abstracted) to keep effective seismic protection of the buildings. III Weak Felt indoors by a few people. People at Based on visual and instrumental investigations, the rest feel a swaying or light trembling. IV Largely Felt indoors by many people, outdoors technical conditions of the buildings were assessed. It observed by very few. A few people are awakened. is stated that in three buildings, Yerevanyan #149, Windows, doors, and dishes rattle. Cherkassi #1, and Shahumyan, all systems that pro- V Strong Felt indoors by most, outdoors by few. Many sleeping people awake. A few are vide increased earthquake resistance are in satisfactory frightened. Buildings tremble througout. condition from a technical point of view. Nonbearing Hanging objects swing considerably. structures of AIUF and some seismic isolators in Isa- Small objects are shifted. Doors and hakyan #4 building are in bad technical condition. windows swing open and shut. Data on the Earthquakes Recorded in Northern buildings but also to schools, hospitals, kindergartens, Armenia from 1996 to 2002 and other critical facilities. Appropriate questionnaires were elaborated to carry out the survey (a sample question- Practically all of the territory of Armenia is situated in a naire is given below). seismically active zone. The size of earthquakes ranges An analysis of the results showed that 94 percent of up to M = 7.1 (according to historical and paleoseismic all respondents were aware of the advantages of their estimations). Focal depth is, on average, 10 kilometers. building over others, 77 percent felt safe during earth- All sources are located on active faults with an average quakes, and 93 percent would like to apply seismic slip rate of about one centimeter per year. During the protection systems if constructing their own houses. One last seven years, more than 20 earthquakes were recorded hundred percent of tenants considered the application in northern Armenia. By geographical location of the of seismic protection systems to critical facilities to be a epicenters of the earthquakes, the distances between the priority and 100 percent wished to see an increase in epicenters and the cities of Vanadzor and Spitak were the number of buildings with seismic protection systems calculated and the maximum possible intensity values to reliably protect people from earthquakes in Armenia. in Vanadzor and Spitak were in the range of III to V. Such Other comments from respondents included: seismic impacts are not a hazard for the buildings · Usually we learn from tenants of neighboring build- inspected, but they cause distress in the population. ings that an earthquake has occurred. From this point of view, the opinions of the buildings' · We have felt few earthquakes in this building during tenants are important for assessing the behavior of build- the past several years. ings during earthquakes. The behavior of people for each · We are glad to live in this type of building. level of intensity from III to V is described in table 20.1. · We feel safe and protected. · If I don't experience any vibration during earthquakes, Survey of the Tenants of the Four Buildings it means I am safe. Examined · We bought this apartment because we knew the build- ing was improved with a seismic protection system. A survey among tenants of apartment buildings with · We believe that the existence of the seismic protec- base isolation or AIUF was carried out on February 6, tion system makes the building safer. 2002 in the cities of Vanadzor and Spitak. The objective · We invite our neighbors from close apartment build- of the survey was to find out how safe the tenants feel in ings to take refuge at our place when an earthquake the buildings that were newly constructed or retrofitted occurs. by new technologies, and to ascertain the advisability of · I have learned about earthquake occurrences only applying AIUF and base isolation not only to apartment from the radio. 296 Building Safer Cities: The Future of Disaster Risk · When there is noise and lights are switched on in the · We would like all of the buildings in our city to be front building, I assume that an earthquake occurred. safe, and all people to feel safe. · We believe that schools, hospitals, and kindergartens · I am safer here in this house than in the temporary should be seismically isolated first of all. dwelling-wagon. Sample Questionnaire for Tenants of the Nine-Story Apartment Building #4 with AIUF in Isahakyan Street, Vanadzor Several buildings in Vanadzor were provided with base isolation or other seismic protection systems from 1995­97. We would like to know your opinion on the advisability of applying such seismic protection systems. This research is being carried out under a World Bank contract. Your sincere answers will help to make our research effective. Thank you in advance. 1. Where were you when the earthquake of 1988 occurred? Out of Armenia In Armenia In the Earthquake Zone 2. For how long have you been living in this building? a) more than 14 b) 6-14 years c) up to 6 years 3. Do you know the advantage of your building over the others? Yes No Find difficulty in responding 4. How many times have you experienced earthquakes or have you heard that an earthquake has occurred? a) have experienced b) have not c) have heard d) have not e) find difficulty .........times experienced ..........times heard in responding 5. If 4a). Did you feel safe at home during the earthquakes? Yes No Find difficulty in responding 6. What kind of apartment (building) did you live in before the earthquake of 1988, and what is the difference between it and the one you live in now? 7. If you would construct your own house, would you like to apply seismic protection systems? Yes No Find difficulty in responding 8. Do you consider seismic protection systems advisable to be applied not only to apartment buildings but also to schools, hospi- tals and other critical facilities? Yes No Find difficulty in responding 9. What can you say about your neighboring apartment buildings tenants' behavior during the earthquakes? 10. Based on your experience would you like to see increase of the number of buildings with seismic protection systems in Armenia? Yes No Find difficulty in responding Floor apartment # Age Sex Male Female Education Secondary Technical college Incomplete higher Higher The Behavior of Retrofitted Buildings During Earthquakes: New Technologies 297 Thus, in the course of the last seven years, the 5. Seismic (base) isolation provides high reliability buildings studied were struck by some 10 earthquakes for all structures and the buildings with seismic isola- of different intensities in the range of III­V by EMS-98. tion are virtually safe from earthquakes. AIUF in com- No serious damage was evident in bearing structures parison with base isolation provides less reliability, but and isolators. Therefore, the behavior of buildings was earthquake resistance is still increased 1.6­1.8 times. assessed based on the tenant survey. The results obtained 6. People feel very safe in the buildings protected by demonstrate that the seismic isolation and seismic seismic (base) isolation or by AIUF. In their opinion, the protection systems implemented behaved during the new technologies must be used extensively and not only earthquakes in the manner envisaged at the design stage. for apartment buildings but also for schools, hospitals, The results of the study represent a strong argument and other critical facilities. for the further use of seismic isolation, not only for apart- 7. During earthquakes of different intensities, build- ment buildings but also for critical facilities. In an arti- ings protected by seismic isolation or AIUF exhibit the cle published in the newspaper "AZG Armenian Daily behavior envisaged at the design stage. This is an objec- #219," November, 29 2001, Professor Nazaretyan, who tive reality showing that seismic isolation technologies conducted an independent study of the building, will provide high efficacy in retrofitting existing or con- wrote that, "During the 1999 Spitak earthquake with structing new critical facilities. an intensity of 5­6, the tenants of the Shahumyan Street 8. Maintaining buildings with the application of seis- building in Spitak and the Yerevanyan #149 building mic isolation technologies, as for any other type of build- in Vanadzor did not feel the earthquake, while people ing, is an important factor. Unsatisfactory maintenance from neighboring buildings ran outside." decreases the effectiveness of seismic protection. Lessons Drawn from Retrofitting Existing Justification for the Possibility of Practical Buildings and from Constructing New Buildings Application and Effectiveness of New Using Seismic Isolation and Protection Systems Technologies to Critical Facilities 1. Retrofitting existing buildings with seismic (base) After completion of the World Bank­financed Earth- isolationorconstructingnewseismic(base)isolatedbuild- quake Zone Reconstruction Project and successful ings can be realized in a relatively short time. Building implementation of new technologies in 1995­1997, the Yerevanyan #149 in Vanadzor was retrofitted in one year application of seismic isolation continued in Armenia. and the Shahumyan building in Spitak was constructed Moreover, technologies developed in Armenia were also in just over one year. For comparison, it should be implemented in Russia and Syria. Armenian factories, noted that strengthening a building similar to Yerevanyan namely "Nairit" and "Yerevan Factory of Rubber Tech- #149byconventionalmethodstakesmorethantwoyears. nical Articles" (YFRTA), are now able to manufacture 2. Retrofitting existing buildings using new tech- seismic isolators and bridge rubber bearings at a high nologies (seismic isolation or AIUF) does not require quality and to international standards. The life expectancy interruption of the use of the buildings. While per- of the bearings is guaranteed by the factories for 45 years. forming the retrofitting, tenants can stay in their apart- Aftertheguaranteeperiod,thebearingsshouldbeinspected. ments. On the contrary, strengthening by conventional If the elastic properties (hardness) of the rubber are methods requires evacuation. unchanged, they can continue to be used. Bearings should 3. The cost of retrofitting with new technologies is also be inspected following strong earthquakes. significantly cheaper (two-thirds less) than the cost of From 1998­2001, 14 seismic-isolated buildings were conventional strengthening. It is even cheaper when designed in Armenia. Included were 12 apartment build- funds for temporary housing are factored in. ings for the new Huntsman Village in Gyumri city, one 4. The cost of construction of new buildings using house for a single family in Proshyan, and one school in seismic isolation technologies is around 30 percent less Vanadzor, which will be retrofitted using seismic isola- than the cost of construction of conventional buildings. tion. The school project in Vanadzor is actually the 298 Building Safer Cities: The Future of Disaster Risk first application of new, advanced seismic isolation tech- Table 20.2 Results of comparative analysis of nology to a critical facility. Building construction in seismic-(base)-isolated and fixed-base Huntsman Village was financed by American John (conventionally designed) buildings Huntsman. The construction of the single-family house Fixed-base Seismic-isolated Name of parameter building building was financed by the owner, an Armenian citizen. This is the first application of base isolation in private con- Total shear force (kN) 40800 10200 Required reinforcement (ton) 360 104 struction in Armenia. The retrofitting of the school in Required reinforcement per 1 m2 110 32 Vanadzor was financed by Swiss CARITAS. of the area of the building (kg) An important factor for the application of such Distance between the reinforcing 20 × 20 40 × 40 technologies in Armenia is the presence of Armenia's bars in the walls (cm) Grade of the concrete B20 (M250) B10 (M150) chemical industry, which is able to locally manufacture Required cement (ton) 810 428 high-quality isolators. Another factor is the presence of Required cement per 1 m2 of the 250 132 scientific and engineering sources capable of designing, area of the building (kg) Cost of reinforcement $144,000 $41,600 investigating, testing, and improving the systems using Cost of cement $32,210 $17,550 new technologies. It is appropriate to note that retro- Cost of seismic isolators -- $24,700 fitting technologies were developed, designed, and imple- Total cost of bearing structure $270,000 $177,640 mented for the first time in Armenia. This fact is Cost of bearing structure per 1 m2 $83 $55 of the area of the building recognized by the international scientific and engineering community. Although the buildings studied were struck Seismic isolation will cost in U.S. dollars: ($144,000 + $32,210) - ($41,600 + $17,550 + $24,700) = $92,360 by medium-strength earthquakes, the experience showed that seismic isolation technology is extremely reliable. on this amount, it can be concluded that, with seismic For example during the destructive 1995 earthquake isolation, costs can be reduced by 30 percent. Almost in Kobe, Japan, the two seismic-isolated buildings the same results were obtained for the school project were not affected at all, while a very large number of in Vanadzor. In this case, seismic isolation reduced build- conventionally designed buildings collapsed. ing materials by 15 tons of steel and 250 cubic meters AIUF as a seismic protection method is very efficient of concrete. Generally, the Armenian experience has for tall, flexible structures where base isolation is inap- shown that savings due to seismic isolation for retro- propriate. Although AIUF increases the seismic resist- fitting of existing buildings could be much higher, some ance of flexible buildings less than base isolation increases 2 to 2.5 times, in comparison with conventional strength- the seismic resistance of rigid buildings, the application ening. Also, seismic isolation allows acceleration of the of AIUF provides resiliency to buildings during future whole construction process. These factors encouraged earthquakes. AIUF is the only method currently devel- Swiss CARITAS, donors for the school project, to under- oped in Armenia that allows increasing seismic resist- take seismic isolation for this critical facility. Presently ance of existing occupied flexible buildings. the Huntsman Village, as well as a school in Vanadzor, With such capabilities, the retrofitting or construc- are under construction. tion of ordinary (apartment) buildings and critical facil- Thus, the successful implementation of new tech- ities using seismic isolation costs much less than nologies during the last eight years; the presence of capa- conventionally designed buildings. For example, a com- ble industry that can locally manufacture seismic isolators; parative analysis was carried out for the Huntsman Vil- the presence of capable scientific and engineering sources lage Project. Some average results of that analysis are locally developing and designing seismic isolation sys- given in table 20.2 for a four-story building. Two situ- tems; the possibility of retrofitting by seismic isolation ations are considered: one building is designed with an without interruption of facilities use; the low cost of ordinary type of foundation (conventional design), retrofitting and new construction using seismic isola- and the other is seismic-isolated. tion; the possibility of accelerating the whole construction The cost of the bearing structure of this building process; and high reliability of seismic isolation fully (defined through a tender) is around $270,000. Based justify further practical application and effectiveness of The Behavior of Retrofitted Buildings During Earthquakes: New Technologies 299 the advanced seismic isolation technologies to critical 5. A Center of Anti-Seismic Technologies should be facilities. established, possibly at the American University of Armenia. Such a center could train people working in the construction industry to deal with the develop- Recommendations Influencing Decisionmakers ment of new designs and further improve new tech- for Construction Policies for Further nologies and their application to critical facilities. Implementation of New Technologies 6. Guidelines for seismic isolation and the protec- tion of buildings and structures should be developed 1. It is recommended that International Financial and published to promote further implementation of Institutions (IFIs) implement the new advanced tech- new technologies in the country (and in the region). nologies in relevant projects in Armenia and other coun- 7. The experience of Armenia with retrofitting and tries. Such technologies were born within the framework seismic isolation should be shared with authorities in of the World Bank Earthquake Zone Reconstruction other countries with similar earthquake risk. Project in Armenia, and it would be desirable to encour- 8. Local authorities and tenants of buildings where age the use of these technologies, especially for critical new technologies were implemented should be well facilities. trained in building maintenance to ensure that the tech- 2. It is recommended that IFIs encourage the Gov- nologies perform well during earthquakes. The Gov- ernment of Armenia and others involved in construc- ernment of Armenia and donors should provide funds tionfinancingtocontinuetheapplicationofnewadvanced for jacketing isolators and repairing nonstructural ele- technologies, not only for apartment buildings but for ments in some buildings. critical facilities as well. A booklet briefly describing all 9. The 8th World Seminar on Seismic Isolation, Pas- of the advantages of these technologies could be pre- sive Energy Dissipation and Active Vibration Control of pared and distributed among the interested parties. Structures will take place in October 2003 in Yerevan. 3. IFIs should be consistent in insisting that con- This decision was made in recognition of the work under- struction funded includes reliable earthquake-resistant taken in Armenia in the field of seismic isolation. The technologies. Special attention should be paid to con- participation of decisionmakers in construction policy struction companies, since many are reluctant to use from Armenia and other countries should be encour- new technologies and are not interested in obtaining aged in that high-level forum. smaller contracts. 4. A series of lectures should be organized to train those in construction companies, the Ministry of Urban Note Development, design institutes, and donor organiza- 1. See World Bank ICR No. 17255, Armenia Earthquake Zone tions about the new technologies. Reconstruction Project, December 1997. T H E W O R L D B A N K 1818 H Street, N.W. Washington, D.C. 20433 U.S.A. Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: books@worldbank.org ISBN 0-8213-5497-3