Cities, Seas, and Storms Managing Change in Pacific Island Economies Volume IV Adapting to Climate Change November 30, 2000 PAPUA NEW GUINEA AND PACIFIC ISLANDS COUNTRY UNIT ˇ THE WORLD BANK in collaboration with Environment and Conservation ˇIˇGˇCˇIˇ SPREP Division PICCAP Kiribati Copyright Š 2000 The International Bank for Reconstruction And Development/ THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing November 13, 2000 Second printing November 30, 2000 World Bank Country Study Reports are among the many reports originally prepared for internal use as part of the continuing analysis by the Bank of the economic and related conditions of its developing member countries and of its dialogues with the governments. Some of the reports are published in this series with the least possible delay of the use of the governments and the academic, business and financial, and development communities. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of this document and requests for copies or accompanying reports should be sent to: Mr. David Colbert Papua New Guinea and Pacific Islands Country Management Unit East Asia and Pacific Region The World Bank 1818 H Street, NW Washington, D.C., U.S.A. 20433 Fax: (1) 202-522-3393 E-Mail: Dcolbert1@worldbank.org Photo design and concept by Fatu Tauafiafi, SPREP Cover Photos by Fatu Tauafiafi and Jim Maragos E-mail: fatut@sprep.org.ws; Jim_Maragos@r1.fws.gov Inside Photos by SOPAC and Sofia Bettencourt THE WORLD BANK A partner in strengthening economies and expanding markets to improve the quality of life for people everywhere, especially the poorest. THE WORLD BANK HEADQUARTERS 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 Facsimile: (202) 477-6391 Telex: MCI64145 WORLDBANK MCI 248423 Cable Address: INTBAFRAD WASHINGTONDC World Wide Web: Http://www.worldbank.org E-mail: books@worldbank.org DRAFT NOT FOR CITATION OR CIRCULATION Cities, Sea, and Storms Managing Change in Pacific Island Economies Volume IV Adapting to Climate Change November 30, 2000 PAPUA NEW GUINEA AND PACIFIC ISLAND COUNTRY UNIT THE WORLD BANK ii Table of Contents Page No. Acknowledgements.................................................................................. v Acronyms and Abbreviations...................................................................... vii Executive Summary................................................................................ ix Chapter1. Key Challenges.................................................................. 1 Chapter 2. Climate Change Scenarios in the Pacific.................................. 5 Chapter 3. Impact of Climate Change on a High Island: Viti Levu, Fiji.......... 7 A. Impact on Coastal Areas............................................... 8 B. Impact on Water Resources........................................... 11 C. Impact on Agriculture.................................................... 13 D. Impact on Public Health ................................................ 15 Chapter 4. Impact of Climate Change on Low Islands: Tarawa Atoll, Kiribati.. 19 A. Impact on Coastal Areas................................................ 21 B. Impact on Water Resources............................................ 23 C. Impact on Agriculture................................................... 25 D. Impact on Public Health ............................................... 25 Chapter 5. Impact of Climate Change to Regional Tuna Fisheries................. 27 Chapter 6. Toward Adaptation: Moderating the Effects of Climate Change... 29 A. The Need for Immediate Action....................................... 29 B. Guidelines for Selecting Adaptation Measures...................... 30 C. Adaptation Options .................................................... 31 D. Implementing Adaptation ............................................. 34 E. Funding Adaptation ..................................................... 36 Chapter 7. Summary of Key Findings and Recommendations........................ 39 References Annex A. Assumptions Used in the Assessment of Climate Change Impacts Map iii iv Acknowledgments Volume IV of this report was prepared by a team managed by Sofia Bettencourt (World Bank) and Richard Warrick (IGCI). The report was the product of a partnership between the World Bank and the International Global Change Institute (IGCI), University of Waikato, New Zealand, the Pacific Islands Climate Change Assistance Programme (PICCAP) country teams of Fiji and Kiribati, the South Pacific Regional Environment Programme (SPREP), Stratus Consulting Inc. (U.S.), the Center for International Climate and Environmental Research (CICERO, Norway), and experts from numerous other institutions who participated in the research. Background studies to this report are included in References. The analysis of climate change impacts in Viti Levu, Fiji, was conducted by Jone Feresi (main editor, Ministry of Agriculture, Fisheries and Forestry, Fiji), Gavin Kenny (coordinator and agriculture impacts, IGCI), Neil de Wet (health impacts, IGCI), Leone Limalevu (coastal impacts, PICCAP national coordinator), Jagat Bhusan (agriculture impacts, Ministry of Land and Mineral Resources), Inoke Ratukalou (agriculture impacts, Ministry of Agriculture, Fisheries and Forestry), Russell Maharaj (coastal impacts, South Pacific Applied Geoscience Commission, SOPAC), Paul Kench (coastal impacts, IGCI), James Terry (water resources impact, University of South Pacific), Richard Ogoshi (agriculture impacts, University of Hawaii), and Simon Hales (health impacts, University of Otago, New Zealand). The analysis of climate change impacts in Tarawa, Kiribati was conducted by Tianuare Taeuea (main editor and health impacts, Ministry of Health and Family Planning, Kiribati), Ioane Ubaitoi (agriculture impacts, Ministry of Agriculture), Nakibae Teutabo (agriculture, PICCAP national coordinator at the Ministry of Environment and Social Development), Neil de Wet (health impacts, IGCI), Gavin Kenny (agriculture impacts, IGCI), Paul Kench (coastal impacts, IGCI), Tony Falkland (water impacts, Ecowise Environmental, Australia), and Simon Hales (health impacts, University of Otago). The analysis of climate change impacts on tuna fisheries was led by Patrick Lehodey and Peter Williams (Secretariat of Pacific Community). John Campbell (IGCI) summarized the results of the study and contributed to the review of adaptation strategies and climate variability impacts. Richard Jones and Peter Whetton (CSIRO, Australia) helped develop the scenarios for climate variability. W. Mitchell contributed to the analysis of sea level rise. The economic analysis of climate change impacts was carried out by Bob Raucher (Stratus Consulting), Sofia Bettencourt, Vivek Suri (World Bank), and Asbjorn Aaheim and Linda Sygnes (CICERO). Wayne King (SPREP) contributed a regional overview, and Maarten van Aalst provided an international perspective and reviewed the final work. Samuel Fankhauser (European Bank for Reconstruction and Development) and Mahesh Sharma (World Bank) provided advice on the design of the study and reviewed the draft results. Joel Smith (Stratus) contributed to the adaptation analysis. The authors are also grateful to Graham Sem and Gerald Miles (SPREP), Alf Simpson and Russell Howorth (SOPAC), Alipata Waqaicelua (Fiji Meteorological Services), Robin Broadfield and Noreen Beg (World Bank), Herman Cesar (Free University of Amsterdam), Clive Wilkinson (Australian Institute of Marine Sciences), and Ove Hoegh-Guldberg (University of Queensland) for their advice and support. This volume is part of a four-volume Regional Economic Report prepared by Laurence Dunn, Stuart Whitehead, Sofia Bettencourt, Bruce Harris, Anthony Hughes, Vivek Suri, John Virdin, Cecilia Belita, David Freestone, Philipp Müller, Gert Van Santen, and Cynthia Dharmajaya. Natalie Meyenn, Danielle Tronchet, Peter Osei and Maria MacDonald provided important advice and support. Barbara Karni helped edit the report. Fatu Tauafiafi (SPREP) contributed the cover photo design and concept. v Peer reviewers included Iosefa Maiava (Forum Secretariat), Sawenaca Siwatibau (ESCAP Pacific Operations Center), and Hilarian Codippily, Robert Watson, Mary Judd, Charles Kenny, and Richard Scheiner (World Bank). This report was funded by the World Bank Country Management Unit of Papua New Guinea and Pacific Islands, the Australian Trust Fund for the Pacific, the PICCAP program, the World Bank Climate Change group, the Norwegian Trust Fund, the New Zealand Trust Fund, and the Danish Trust Fund for Global Overlay. Many of the authors also contributed their own time and efforts to the study, for which the World Bank is grateful. vi Acronyms and Abbreviations A$ Australian Dollar A2 High Climate Sensitivity 4.5o C Greenhouse Gas Emissions Scenario ADB Asian Development Bank B2 Mid Climate Sensitivity 2.5 o C Greenhouse Gas Emissions Scenario CDM Clean Development Mechanism CICERO Center for International Climate and Environmental Research "CIMSIN" Container Inhabiting Mosquito Simulation Model CSIRO Commonwealth Scientific and Industrial Research Organization CSIRO9M2 9-Layer Global Circulation Model of Australia's Commonwealth Scientific and Industrial Research Organization (Mark 2 Version) DENSIM (Mosquito Population) Density Simulation Model DHF Dengue Hemorrhagic Fever DKRZ Deutsche Klimarechenzentrum (German Climate Monitoring Center) DSS Dengue Shock Syndrome EACNI East Asia and Pacific Country Management Unit for Papua New Guinea and Pacific Islands (World Bank) EASRD East Asia and Pacific Regional Development and Natural Resources Unit (World Bank) EAPVP East Asia and Pacific Vice Presidency (World Bank) ENB Earth Negotiations Bulletin ENSO El Niņo Southern Oscillation ESCAP Economic and Social Commission for Asia and the Pacific FAO Food and Agriculture Organization of the United Nations F$ Fijian Dollar FFD Fiji Fisheries Division GEF Global Environmental Facility GCM Global Circulation Model GDP Gross Domestic Product GHG Greenhouse Gases IGCI International Global Change Institute IPCC Intergovernmental Panel on Climate Change JICA Japan International Cooperation Agency vii M3 Cubic Meter MAGICC Model for the Assessment of Greenhouse Gas Induced Climate Change MHWS Mean High Water Spring (Level) MSL Mean Sea Level MT Metric Ton NGOs Nongovernmental Organizations PACCLIM Pacific Climate Change Impacts Model PICCAP Pacific Islands Climate Change Assistance Programme PLANTGRO Plant Growth Model from the Commonwealth Scientific and Industrial Research Organization PNG Papua New Guinea SEPODYM Spatial Environmental Population Dynamics Model SLR Sea Level Rise SOPAC South Pacific Applied Geoscience Commission SPC Secretariat of the Pacific Community SPCZ South Pacific Convergence Zone SPECTRUM Population Growth Model from Spectrum Human Resources Systems Corporation SPREP South Pacific Regional Environmental Programme SRES Special Report on Emission Scenarios STM Shoreline Translation Model SUTRA Saturated and Unsaturated Transport Model UNDAC United Nations Disaster Assessment and Coordination UNDP United Nations Development Programme UNESCO United Nations Educational, Scientific and Cultural Organization UNFCCC United Nations Framework Convention on Climate Change US$ United States Dollar US United States WHO World Health Organization WPWP Western Pacific Warm Pool WRI World Resources Institute Vice-President: Jemal-ud-din Kassum, EAPVP Country Director: Klaus Rohland, EACNF Acting Sector Director: Mark Wilson, EASRD Task Team Leader: Laurence Dunn, EACNF viii Executive Summary As the 21st century begins, the Pacific Island region. Chapter 3 examines the physical and people confront a future that will differ economic impact of these scenarios on a high drastically from the past. Their physical climate, island of the Pacific ­ Viti Levu, in Fiji. The access to resources, ways of life, external potential impacts of climate change on coastal relations and economic structures are areas, water resources, agriculture, and health undergoing simultaneous and interactive change. are discussed in turn. Chapter 4 examines the Pacific Island countries can actively engage in potential effect of climate change on a group of foreseeing and managing the process of low islands ­ the Tarawa atoll in Kiribati ­ adaptation to these changes, or they can have focusing primarily on coastal and water resource unplanned adaptation imposed on them by impacts. Chapter 5 discusses the potential forces outside their control. impact of climate change on the tuna fisheries of Managing these forces will be particularly the Central and Western Pacific. Chapter 6 critical in the area of climate change, a subject proposes a general adaptation strategy for that is very difficult for communities and Pacific Island countries. Key findings and governments to grasp, but of immense and recommendations are summarized on Chapter 7. immediate impact on Pacific Island countries. Annex A describes the methodology and Choosing a development path that decreases the assumptions used to assess climate change islands' vulnerability to climate events and impacts. Detailed background studies to this maintains the quality of the social and physical report are included in References. environment will not only be central to the future well being of the Pacific Island people, This volume is the last of a four-volume report but will also be a key factor in the countries' entitled "Cities, Seas and Storms: Managing ability to attract foreign investment in an Change in Pacific Island Economies" produced increasingly competitive global economy. by the World Bank as the Year 2000 Regional Economic Report for the Pacific Islands. In This volume examines the possible impacts of addition to this specialized volume, the series changes in climate on high and low islands of includes a summary report (Volume I), a volume the Pacific, and discusses key adaptation and dedicated to the management of Pacific towns financing strategies available to Pacific Island (Volume II), and a volume dedicated to the countries. The short-term outcome of the report management of the ocean (Volume III). is intended to be an improved understanding of the need and scope for adaptation policies in Impacts from Climate Change face of the challenges posed by climate change. Over the long term, it is hoped that the report can assist Pacific Island governments, The warming of the earth's atmosphere is likely businesses and communities to better adapt to to have substantial and widespread impacts on change by building on the strengths unique to Pacific Island economies, affecting sectors as their countries and their people. It is also hoped varied as health, coastal infrastructure, water that the findings of the report can contribute to resources, agriculture, forestry and fisheries. the on-going international dialogue on adaptation financing. Some policymakers dismiss the impacts of climate change as a problem of the future. But This volume is divided into seven chapters. there is evidence that similar impacts are already Chapter 1 outlines the nature of the challenges being felt: the Pacific Islands are becoming posed by climate change. Chapter 2 describes increasingly vulnerable to extreme weather climate change scenarios for the Pacific Island events as growing urbanization and squatter ix settlements, degradation of coastal ecosystems, Another possibility might be to assume the and rapidly developing infrastructure on coastal worst and embark upon major investments in areas intensify the islands' natural exposure to coastal protection ­ such as seawalls ­ and climate events. relocation of vulnerable infrastructure. The first According to climate change models, the sea approach is unwise in light of the increasing level may rise by 2343 centimeters and the evidence of climate change. The second is average temperature by 0.90­1.30C by 2050. impractical and unaffordable. Among the most substantial impacts of climate change would be losses of coastal infrastructure This report recommends that Pacific Island and land resulting from inundation, storm surge, countries follow a strategy of precautionary and shoreline erosion. Climate change could also adaptation. Since it is difficult to predict far in cause more intense cyclones and droughts, the advance how climate change will affect a failure of subsistence crops and coastal fisheries, particular site, Pacific Island countries should losses in coral reefs, and the spread of malaria avoid adaptation measures that could fail or have and dengue fever. unanticipated social or economic consequences if climate change impacts turned out to be In the absence of adaptation, a high island such different than anticipated (IPCC 1998). As a as Viti Levu could experience average annual first step, it is recommended that Pacific Islands economic losses (in 1998 dollars) of countries adopt `no regrets' adaptation US$23$52 million by 2050, equivalent to 24 measures that would be justified even in the percent of Fiji's GDP. A low group of islands absence of climate change. These include better such as the Tarawa atoll in Kiribati could face management of natural resources--particularly average annual damages of US$8$16 million of coastal habitats, land, and water--and by 2050, as compared to a current GDP of measures such as disease vector control and US$47 million. These costs could be improved spatial planning. considerably higher in years of extreme weather Acting now to reduce vulnerability to extreme events such as cyclones, droughts and large weather events would go a long way toward storm surges. preparing Pacific countries for the future, and reducing the magnitude of the damage. Taking How certain is climate change? A soon to be early action may require adjustments of released review by the Intergovernmental Panel development paths and the sacrifice of some on Climate Change (IPCC) concludes that short-term economic gains. But it would vastly mankind has contributed substantially to decrease the downsize costs should climate observed warming over the last 50 years. While change scenarios materialize. The challenge there is growing consensus that climate change will be to find an acceptable level of riskan is occurring, uncertainties remain on the timing intermediate solution between a policy of and magnitude of the changes. Most studies, inaction and investing in high cost however, consider the Pacific Islands to be at solutionsand start adapting long before the high risk from climate change and sea level rise. expected impacts occur. Under a `no regrets' adaptation policy, Pacific A Strategy for Adaptation Island governments would take adaptation goals into account in future expenditure planning, How should Pacific Island countries adapt to would support community-based adaptation, and climate change? One possibility is to do would require major infrastructure investments nothing, and by implication hope that climate to meet adaptation criteria. Adaptation would be change does not happen. This is the de facto viewed as a key feature in national policy in its present position of many governments, including own right, and would be taken into account in those of several Pacific Island countries. the development of policies in a wide range of sectors and activities. x The question of who will fund adaptation is a action using their own resources should not be difficult and sensitive issue. Insofar as `no penalized with lower future allocations. These regrets' measures help reduce the islands' and other disincentives against `no regrets' vulnerability to current climate events policies need to be urgently discussed in (independently of climate change) Pacific Island international forums. Of paramount importance, governments would be justified in funding however, will be for the international adaptation from reallocations in public community to move rapidly to develop a expenditures and development aid. Donors financing mechanism to assist countries such as could support this process directly, or as part of the Pacific Islands in taking early adaptive natural resources and environmental action. The urgency of this action for small management assistance. island states cannot be over-emphasized. The analysis of this report, however, clearly shows that the Pacific Islands are likely to Although uncertainties remain, it now seems experience significant incremental costs certain that climate change will affect many associated with climate change. It is urgent that facets of Pacific Island people's lives in ways the international community develop financing that are only now beginning to be understood. mechanisms to help countries in the receiving As such, climate change must be considered one end of climate change to fund `no regrets' of the most important challenges of the twenty- adaptation. Countries that have taken early first century and a priority for immediate action. xi Chapter 1 Key Challenges Across the Pacific, atoll dwellers speak of having to move their houses away from the Box 1. Climate Change and Climate Variability ocean because of coastal erosion; of having to change cropping patterns because of saltwater Climate change is the gradual warming of the earth's atmosphere caused by emissions of heat-absorbing "greenhouse intrusion; of changes in wind, rainfall, and ocean gases," such as carbon dioxide and methane. The term is currents. While these events may simply reflect generally used to reflect longer-term changes, such as higher climate variability, they illustrate the types of air and sea temperatures and a rising sea level. impacts likely to be felt under climate change. Climate variability reflects shorter-term extreme weather events, such as tropical cyclones and the El Niņo Southern Many policymakers dismiss climate change as a Oscillation (ENSO). While there is some evidence that climate problem of the future. But impacts similar to variability will increase as a result of climate change, many uncertainties remain. those resulting from climate change are already being felt, as the Pacific Islands become Mitigation and adaptation also have distinct meanings among increasingly vulnerable to extreme weather climate change experts. Mitigation refers to efforts to reduce greenhouse gas emissions. Adaptation refers to efforts to events and to climate variability. Cyclones Ofa protect against climate change impacts. and Val, which hit Samoa in 1990­91, caused losses of US$440 million--in excess of the country's annual gross domestic product (GDP). Fiji was hit by four cyclones, two droughts, and Table 1. Estimated Costs of Extreme Weather severe flooding in the past eight years. In the Events in the Pacific Island Region during the 1990s 1990s alone, the cost of extreme events in the (millions of US$) Pacific Island region exceeded US$1 billion (table 1). Estimated Event Year Country losses Cyclone Ofa 1990 Samoa 140 Rising Vulnerability to Extreme Cyclone Val 1991 Samoa 300 Weather Events Typhoon Omar 1992 Guam 300 Cyclone Nina 1993 Solomon Islands ­ Cyclone Prema 1993 Vanuatu ­ The impacts of extreme weather events are Cyclone Kina 1993 Fiji 140 Cyclone Martin 1997 Cook Island 7.5 becoming stronger as the islands' vulnerability Cyclone Hina 1997 Tonga 14.5 rises. Growing urbanization and squatter Drought 1997 Regional > 175a settlements, degradation of coastal ecosystems, Cyclone Cora 1998 Tonga 56 and rapidly developing infrastructure on coastal Cyclone Alan 1998 French ­ Polynesia areas are intensifying the islands' exposure to Cyclone Dani 1999 Fiji 3.5 extreme weather events. At the same time, traditional practices promoting adaptation, such ­ Not available. as multicrop agriculture, are gradually a. Includes losses of US$160 million in Fiji. Note: Minor events and disasters in Papua New Guinea not included. weakening (box 2). Costs are not adjusted for inflation. Source: Campbell 1999, and background studies to this report. 1 Box 2. Rising Vulnerability to Extreme Weather Events Tropical cyclones are regular occurrences in many Pacific Islands. Traditional societies adapted to these events by using a range of resilient food crops and food preservation techniques. Many communities used famine foods during times of scarcity and followed traditional obligations to provide for victims of disasters. This resilience is diminishing, however, leaving many Pacific Islands increasingly vulnerable to extreme weather events. An example is the increasing use of nontraditional crops, such as cassava. Cyclone Meli devastated much of the Southern Lau Island Group in Fiji in 1979 (see figure). Islands such as Nayau were subject to winds of more than 80 knots; other islands, such as Ogea, experienced only gale force winds. The effect of Cyclone Melia on crops depended on the distance Distance Percentage of root Percentage of tree crops from the storm (see table on left). While at the storm center from crops destroyed destroyed crop damage was nearly 100 percent, at distances of 30-100 Island storm Cassava Taro Yam Banana Coconut Breadfruit kilometers from the storm, traditional crops ­ such as taro and (kilometers) yam ­ suffered much less damage than nontraditional crops such as cassava. Nayau 0 100 100 80 100 100 100 Cicia 30 Cassava is becoming increasingly prevalent in the Pacific as a 100 96 54 100 91 100 subsistence crop because of its ability to grow on poor soils and Lakeba 30 94 55 48 82 75 50 the low labor inputs required. But its low resilience to cyclones Vanuavatu 45 75 75 60 50 increases the likelihood that food rations will be required during Oneata 67 60 10 50 40 40 the cyclone season. In most cases, the best strategy for food Komo 86 60 40 30 40 security in cyclone-prone areas is crop diversity and the Moce 88 60 10 50 40 40 maintenance of traditional crops. Namuka 99 50 50 15 30 If tropical cyclone intensity increases under climate change, it Kabara 99 60 10 50 40 is likely that the trend toward cultivation of cassava will result Fulaga 129 50 40 10 30 in greater food crop losses than would be the case if traditional Ogea 142 50 40 10 30 root crops were maintained. From this perspective, promoting Source: Campbell (1995) traditional multicrop agriculture may also be the best adaptation to climate change. Compounding Impacts of intense cyclones and droughts, the failure of Climate Change subsistence crops and coastal fisheries, losses in coral reefs, and the spread of malaria and dengue Arriving on top of this increased vulnerability, fever. These impacts could be felt soon: if climate change is only likely to exacerbate the climate change models are correct, the average current impacts, whether or not climate sea level could rise 11­21 centimeters and variability increases in the future--and there is average temperatures could rise 0.50­0.60C by some evidence that it may. In low islands, the 2025. most substantial damage would come from losses to coastal infrastructure as a result of The economic impact could be substantial. inundation, storm surge, or shoreline erosion. Estimates from this study indicate that if climate But climate change could also cause more 2 change scenarios materialize, a high island such Climate change would have the greatest impact as Viti Levu in Fiji could suffer economic on the poorest and most vulnerable segments of damages of more than US$23$52 million a the population--those most likely to live in year by 2050 (in 1998 dollars), equivalent to squatter settlements exposed to storm surges and 24 percent of Fiji's gross domestic product disease (where safety nets have weakened), and (GDP). The Tarawa atoll in Kiribati could face those most dependent on subsistence fisheries average annual economic damages of and crops destroyed by cyclones and droughts. US$8$16 million by 2050 (as compared with a Nevertheless, the impacts of climate change are GDP of about US$47 million). In years of strong likely to be pervasive and affect the lives of storm surge, up to 54 percent of South Tarawa most Pacific Islanders. could be inundated, with capital losses of up to US$430 million. 3 4 Chapter 2 Climate Change Scenarios in the Pacific In 1999­2000 the World Bank helped sponsor a would be critical for low-lying atolls in the study of vulnerability, adaptation, and economic Pacific, which rarely rise 5 meters above sea impact of climate change in the Pacific Island level. It could also have widespread region.1 The analysis used an integrated model implications for the estimated 90 percent of of climate change developed for the region, the Pacific Islanders who live on or near the Pacific Climate Change Impacts Model coast (Kaluwin and Smith 1997). (PACCLIM), complemented by sectoral impact models, population projections, and baseline data such as historical climate records. Based on ˇ Increase in surface air temperature. Air the best scientific information available for the temperature could increase 0.90-1.30 C by region, the following scenarios were used by the 2050 and 1.60-3.40C by 2100.2 study (table 2): ˇ Changes in rainfall. Rainfall could either ˇ Rise in sea level. Sea level could rise 0.2 rise or fall--most models predict an meters (in the best-guess scenario) to 0.4 increase--by 8-10 percent in 2050 and by meters (in the worst-case scenario) by 2050. about 20 percent in 2100, leading to more By 2100, the sea could rise by 0.5-1.0 intense floods or droughts. meters relative to present levels. The impact Table 2. Climate Change and Variability Scenarios in the Pacific Island Region Impact 2025 2050 2100 Level of Certainty Sea level rise (centimeters) 1121 2343 50103 Moderate Air temperature increase (degrees Centigrade) Fiji 0.50.6 0.91.3 1.63.3 High Kiribati 0.50.6 0.91.3 1.63.4 High Change in rainfall (percent) Fiji -3.7+3.7 -8.2+8.2 -20.3+20.3 Low Kiribati -4.8+3.2 -10.7+7.1 -26.9+17.7 Low Cyclones Frequency Models produce conflicting results Very Low Intensity (percentage increase in wind speed) 020 Moderate Region of formation No change Low Region of occurrence No change or increase to north and south Low El Niņo Southern Oscillation (ENSO) A more El Niņo­like mean state Moderate Note: Ranges given reflect a best-guess scenario (lower value) and a worst-case scenario (higher value). Sea level rise is derived from global projections, as regional models have not yet been developed. Temperature and rainfall projections are based on the CSIRO9M2 and the DKRZ Global Circulation Models. ENSO and cyclone scenarios are based on a comprehensive review of climate variability in the South Pacific (Jones and others 1999). For details, see Annex A. 2 A new report by the Intergovernmental Panel on Climate Change, scheduled to be finalized in early 2001, raises the 1 See Acknowledgments for a list of the experts and worst case scenario for surface air temperature to 6o C by organizations that participated in the study. Background 2100. This means that if the worst case scenario studies to this report are cited in References. The materializes, the impacts may be considerably higher than assumptions used by the study are detailed in Annex A. estimated here. 5 ˇ Increased frequency of El Niņo-like uncertainties are magnified because the area of conditions. The balance of evidence the countries usually falls below the levels of indicates that El Niņo conditions may occur resolution of the general circulation models more frequently, leading to higher average used. rainfall in the central Pacific and northern Polynesia. The impact of El Niņo Southern Some changes are more certain than others: Oscillation (ENSO) on rainfall in Melanesia, there is emerging consensus that global average Micronesia, and South Polynesia is less well temperature and sea levels will increase. understood (Jones and others 1999). Rainfall changes remain highly uncertain, however, as does the relationship between long- ˇ Increased intensity of cyclones. Cyclones term climate change and extreme events. may become more intense in the future (with Uncertainties also increase with spatial wind speeds rising by as much as 20 resolution: there is greater confidence in model percent); it is unknown, however, whether projections of global average changes than in they will become more frequent. A rise in projections of regional or local level changes. sea surface temperature and a shift to El Impacts on coastal areas and water resources are Niņo conditions could expand the cyclone generally more certain than impacts on path poleward, and expand cyclone agriculture and health. And uncertainty occurrence east of the dateline. The increases with time: projections for 2100 are less combination of more intense cyclones and a certain than projections for 2050. Despite these higher sea level may also lead to higher uncertainties, most studies consider the Pacific storm surges (Jones and others 1999). Islands to be at high risk from climate change and sea level rise (Kench and Cowell 1999). How certain is climate change? The Intergovernmental Panel on Climate Change Based on the results of the study, the physical (IPCC) stated in 1995 that "the balance of and economic impacts of climate change in the evidence suggests a discernible human Pacific Island region are illustrated here by the influence on global climate change" (IPCC example of a high island--Viti Levu in Fiji-- 1995). In a report scheduled to be finalized in and a group of low islands--the Tarawa atoll in early 2001, however, IPCC concluded that Kiribati. To give perspective to the analysis, all human influence had contributed substantially to economic damages were estimated for 2050 as if observed warming over the past 50 years. the impacts had occurred under today's socio- economic conditions. Ranges provided represent While there is growing consensus among experts a "best guess" scenario (the lower bound) and a that climate change is occurring, uncertainties "worst case" scenario (the upper bound). All remain about the magnitude and timing of the economic costs reflect 1998 US dollars, and changes. For small island states, these assume no adaptation. 6 Chapter 3 Impact of Climate Change on a High Island Viti Levu, Fiji With an area of 10,389 square kilometers, Viti By 2050, under the climate change scenarios Levu is the largest island in Fiji. Seventy-seven used by the study, Viti Levu could experience percent of Fiji's population--595,000 people in annual economic losses of US$23­$52 million 1996--reside there. It is also in Viti Levu that (table 3). Because the losses are annual Fiji's major cities, industries, and tourism averages, they dampen the potential impact of facilities are located (box 3). extreme weather events, which could be Box 3. Viti Levu (Fiji) at a Glance Climate: Oceanic tropical ˇ Average temperature: 23o 27o C. ˇ Average rainfall: East Viti Levu 3,0005,000 milimeters. West V. Levu 2,0003,000 milimeters. Major Influences ˇ Southeast trade winds. ˇ El Niņo Southern Oscillation (ENSO). ˇ South Pacific Convergence Zone. Extreme Events ˇ Tropical cyclones. ˇ Droughts (often associated with ENSO). Population ˇ Population: 772,655 in 1996 (595,000 or 77% in Viti Levu). ˇ Average annual growth rate: 0.8%. ˇ 60% rural, but growing urbanization. Economy ˇ GDP (1998): US$1,383 million. ˇ Narrow base, with sugar and tourism dominating. * ˇ Average annual growth: 2.7% (1993­96). ˇ Growth affected by tropical cyclones and drought. ˇ Main exports: sugar, gold, garments. Future Population Trends Future Economic Trends ˇ Increasing population density, especially in urban areas, ˇ Continued dependence on natural resources. coastal areas, flood plains, and marginal hill lands. ˇ Tourism sector likely to grow. Implications: ˇ Agriculture will continue to be important in subsistence ˇ Denser urban structures. production and as export earner. Role of sugar uncertain ˇ Spread of urban areas to coastal margins and inland. due to future removal of subsidies and land tenure. ˇ Increase in proportion of squatter housing. Possible increase in traditional crops, such as kava. ˇ Increase in coastal infrastructure (related to tourism). ˇ Growing importance of cash economy. ˇ Continued dependence on foreign aid. Future Environmental Trends Future Sociocultural Trends ˇ Environmental degradation in densely populated areas ˇ Reduction in importance of traditional kinship systems. (especially coastal and lowland) and in marginal farmland. ˇ Increase in preference for imported foods. ˇ Increase in deforestation. ˇ Increase in noncommunicable diseases associated with ˇ Increase in problems of waste disposal (sewage, solid waste, nutritional and lifestyle changes. and chemical pollution). ˇ Increase in poverty and social problems in urban areas. 7 Table 3. Estimated Annual Economic Impact of Climate Change on Viti Levu, Fiji, 2050 (millions of 1998 US$) Likely Impact Annual damagea Level of cost of an Extreme Certainty extreme eventb event Impact on coastal areas Loss of coastal land and infrastructure to erosion 3-6 Moderate Loss of coastal land and infrastructure to Large inundation and storm surge 0.3-0.5 Moderate 75-90 Storm Surge Loss of coral reefs and related services 5-14 Very low Loss of nonmonetized services from coral reefs, mangroves and seagrasses + Very low Impact on water resources Increase in cyclone severity 0­11 Moderate 40 Cyclone Increase in intensity of droughts (related to El Niņo) + Moderate 50-70 Drought Changes in annual rainfall (other than impacts on agriculture) + Low Impact on agriculture Loss of sugarcane, yams, taro, and cassava due to temperature or rainfall changes and ENSO 14 Moderate 70 Drought Loss of other crops + Very Low Impact on public health Increased incidence of dengue fever 1-6 Moderate 30 Large epidemic Increase in fatal dengue fever cases + Very Low Increased incidence of diarrhea 0-1 Low Infant mortality due to diarrhea + Very Low Impact of cyclones and droughts on public safety + Very Low Total estimated damages >23-52+ + Likely to have economic costs, but impact not quantified. Notavailable. aReflects the incremental average annual costs of climate change. bReflects the absolute (non-incremental) cost of a future extreme event. Numbers are rounded. Note: For assumptions, see annex A. Source: Background studies to this report. substantially higher in a particular year: an adaptation and are subject to large margins of average cyclone could cause damages of more uncertainty. But they probably underestimate the than US$40 million, while a drought comparable costs of actual damages, as many impacts (such to the 1997/98 event could cost Viti Levu some as nutrition and loss of lives) could only be US$70 million in lost crops. assessed qualitatively. Among the most significant incremental impacts of climate change would be damages to A. Impact on Coastal Areas infrastructure and ecosystems of coastal areas Viti Levu's coastal areas are naturally exposed (averaging about US$8­$20 million a year by to weather events. About 86 percent of the 750- 2050). But a higher intensity of cyclones could kilometer coast lies at elevations that are less also result in substantial damages, up to US$11 than 5 meters from sea level. Intensive urban million a year. Changes in rainfall could lead to development, growing poverty, deforestation of agricultural losses of US$14 million per year, watersheds, pollution, and increased exploitation and the combined effect of higher temperatures of coastal resources have exposed large areas of and stronger climate variability could result in the coast to erosion and inundation. Some public health costs of more than US$1­$6 villages have reported shoreline retreats of 15­ million a year. These estimates assume no 8 20 meters over the past few decades due to loss Figure 1. Likely Impact of Climate Change on Coastal of mangroves (Mimura and Nunn 1994). Areas in Viti Levu, Fiji Climate change is expected to affect the coast of Increased Increased Viti Levu through a rise in sea level (2343 Sea Level Sea Tropical centimeters by 2050), higher temperatures Rise Surface Cyclone (0.91.3o C by 2050), and more intense Intensity cyclones, resulting in further coastal erosion and Impacts on inundation as well as a decline in coral reefs Coral Reefs (figure 4.1). The resulting economic losses are conservatively estimated at US$8$20 million a Shoreward Increase wave Reduce sediment year by 2050. Retreat of Energy production Mangroves To assess the potential impact of sea level rise on coastal erosion and inundation, four sections Increased Increased Increased River of Viti Levu were surveyed (see map, box 3): exposure coastal Flooding to inundation erosion ˇ Suva Peninsula, representing major towns or about 5 percent of Viti Levu's coast. ˇ Korotogo on the southern coast, representing areas with major tourism settlements and coastal villages (28 percent of the coast). ˇ Tuvu, on the northwest coast, with intensive sugarcane fields and mangroves (about 47 percent of the coast). ˇ Western Rewa River Delta, Table 4. Potential Shoreline Retreat in Viti Levu, Fiji representing low-lying Resulting from Sea Level Rise, 2025, 2050, and 2100 mangrove and delta areas (10 percent of the coast). Impact 2025 2050 2100 The erosion analysis did not include Potential shoreline retreat (in meters) Suva because the city is already Korotogo (South coast) 1 1-3 4-9 heavily protected by seawalls. Tuvu (Northwest coast) 7-9 9-12 13-29 Western Rewa river delta 50-112 112-251 319-646 Coastal Erosion. The first-order Total land eroded (in hectares) Tourism areas (like Korotogo) 10-22 22-53 63-145 estimates of potential erosion Sugarcane areas (like Tuvu) 188-253 253-323 362-818 indicate that, by 2050, Viti Levu's Low-lying mangrove and delta (like Rewa) 390-875 875-1,955 2,485-5,036 shoreline could retreat by 13 Percentage coastal strip eroded (< 10 meters) 1-2 2-4 5-10 meters at Korotogo, 912 meters at Total land eroded (in hectares) 588-1,150 1,150-2,331 2,910-6,000 Tuvu, and 112251 meters at the Rewa river delta (table 4). Value of land lost to erosion (US$ million) Tourism areas (like Korotogo) 0.2-0.4 0.4-1.0 1.2-2.8 Extrapolating these results to other Sugarcane areas (like Tuvu) 4.3-5.8 5.8-7.3 8.2-18.6 areas of Viti Levu is difficult due to Low-lying mangrove and delta (like Rewa) 8.9-19.9 19.9-44.5 56.5-114.5 the variations in topography. Total capital value of land lost 13.3-26.1 26.1-52.8 66.0-136.0 Nevertheless, using estimates from Annualized lossesa 1.5-2.9 2.9-5.8 7.3-15.0 the three sites surveyed, 1,1502,300 hectares of coastline (2 Notes: Ranges reflect best guess (lower value) and worst case scenarios (higher value). Land eroded and value of land lost are extrapolated from the sites surveyed, and cover about to 4 percent of the land below 10 85 percent of the Viti Levu coast. meters altitude) could be lost by a. Reflects the annual value of the losses, or the capital recovery factor. 2050. By 2100, the proportion of Source: Background studies to this report. For assumptions, see annex A. 9 Table 5. Potential Inundation of the Coast of Viti Levu, Fiji, as a Result of Sea Level Rise Potential Inundation Physical Impact Inundation costs (in US$ million) Percentage of Sea Level Scenario Land total land Annualized Incremental capital value Rise (m) equivalent to inundated (ha) below 10 lossesa of lost assets during an meters altitude extreme event b 2025 worst-case 0.2 2050 best-guess 370 0.6 0.3 14.6 2050 worst-case 3,530 5.9 0.5 30.1 0.40.5 2100 best-guess a- Reflects the incremental annual value of the losses due to climate change, or the capital recovery factor. The costs take into consideration the impact of a 1 in 50 year storm event. b- Reflects the incremental cost of capital losses during a 1 in 50 year storm event. Note: For assumptions, see annex A. Source: Background studies to this report land eroded could be as high as 10 percent. ˇ Reduced efficacy of in-ground septic and Based on current values of land, the annualized sewer pumping systems. economic damages due to climate change would be in the order of US$2.9 to US$5.8 million a ˇ Increased sedimentation of channels, year by 2050 (table 4). This is likely to be an shoreward retreat of mangroves, and underestimate, as the sites surveyed were increased susceptibility to floods in the representative of just about 85 percent of Viti Rewa Delta. Levu's coast, and the Tuvu site under-represents low-lying sugarcane fields on the north shore. Mangroves and Coral Reefs. Viti Levu is estimated to have 23,500 hectares of mangroves Coastal Inundation. The analysis conducted in (Watling 1995) and about 150,000 hectares of Viti Levu indicates that sea level rise would coral reefs.3 Mangroves play key roles in result in relatively modest levels of inundation trapping sediments and protecting coastal areas affecting about 0.6 to 5.9 percent of coastal land against erosion. They are also vital nursery below 10 meters altitude by 2050. However, in grounds for coastal fisheries. The impact of sea years of strong storm surge such as the 1 in 50 level rise and storm surges on mangroves is year event shown on table 5 Viti Levu could expected to be mixed: some expansion might be experience losses in capital assets of US$75-$90 observed due to the increased sedimentation of million, some US$15-$30 million higher than the coastal zone; the net impact of erosion, what is experienced today (table 5). however, is expected to be negative. Past research also suggests the following likely Coral reefs are likely to be particularly affected impacts of climate change (Solomon and Kruger by climate change. A rise in sea surface 1996): temperature of more than 1oC could lead to ˇ extensive coral bleaching and, if conditions Overtopping of shore protection in persist, to coral mortality. Such bleaching events downtown Suva during extreme wave were observed during the 1997­98 El Niņo impacts (if sea level rises 25 centimeters). episode (Wilkinson and others 1999) and more ˇ Serious flooding in large parts of Suva Point recently in Fiji, Tonga and the Solomon Islands and downtown Suva even during moderate tropical cyclones (if sea level rises 100 3 centimeters). This estimate is derived based on the total area of coral reefs in Fiji (an estimated 1 million hectares, according to ˇ Raised water tables in low-lying areas. WRI 1999), and Viti Levu's share of the total coastal area of Fiji (15 percent). 10 in April 2000. The deeper water resulting from Table 6. Summary of Estimated Annual Economic an increase in sea level could stimulate the Impact of Climate Change on the Coast of vertical growth of corals, but reef response is Viti Levu, Fiji, 2050 (millions of 1998 US$) likely to lag the rise in the sea level by at least 40 years (Hopley and Kinsey 1988; Harmelin- Vivien 1994). As a result, many coral species Category Annual damages may not be able to adapt sufficiently rapidly to a succession of bleaching events triggered by Impact on coastal assets: higher sea surface temperatures. Loss of land to erosion 2.95.8 Inundation of land and infrastructure 0.30.5 The climate change impact on coral reefs in Viti Levu is projected to cost an estimated Impact on coral reefs Loss of: US$5$14 million a year by 2050 in lost Subsistence fisheries 0.1­2.0 fisheries, habitat and tourism value (see annex Commercial coastal fisheries 0.0.50.8 A for detailed assumptions). Tourism 4.810.8 Habitat 0.20.5 Biodiversity + B. Impact on Water Resources Nonuse values + Impact on mangrovesa Average rainfall could either increase or Impact on seagrasses + decrease by 2050 (see table 2). The impact will Total estimated damages >8.420.4 depend to a large extent on the South Pacific Convergence Zone (SPCZ). If the SPCZ moves + Likely to have economic costs, but impact not quantified. aAccounted for in the erosion analysis. away from Fiji and the region shifts to a more Note: For detailed assumptions, see annex A. El Niņo­like state, Viti Levu could experience Source: Background reports to this study. more pronounced droughts. If the SPCZ intensifies near Fiji, average rainfall could Figure 2. Cyclone Wind Speed and Impact in Fiji increase. Relationship Between Maximum Wind Speed and Damages It is also possible that Viti Levu would Caused by Major Cyclones in Fiji (1983-1997) experience greater climate variability, with 100 Kina (1993) alternating floods and droughts brought on by more intense cyclones and ENSO fluctuations. 80 illion) Oscar (1983) The sequence of four cyclones and two droughts M experienced in 1992­99 could reflect the future $SU 60 pattern of climate variability. 8991( Eric (1985) s 40 Cyclones. With an average of 1.1 cyclones a gea Sina (1990) Gavin (1997) year (Pahalad and Gawander 1999), Fiji has the maD 20 Storm Gavin highest incidence of cyclones in the South (1985) Joni (1992) Pacific. The four tropical cyclones that hit Fiji 0 0 10 20 30 40 50 60 70 80 between 1992 and 1999 killed 26 people and Maximum Wind Speed (knots) caused damages estimated at US$115 million Source: Fiji M eteorological Services.. Damages are converted to 1998 costs (Fiji Meteorological Services undated). Most of the damage occurred on Viti Levu. speed could result in a 44­100 percent increase Regional studies indicate that cyclone intensity in cyclone damage (figure 2).4 Taking Fiji's may increase by 0­20 percent in the Pacific average annual cyclone damage for the 1992­99 Island region as a result of climate change period (US$14.4 million) as a baseline and (Jones and others 1999). Based on historical adjusting the figure to reflect the relative share records of cyclone damage in Fiji and scientific theory, a 20 percent increase in maximum wind 4 See Annex A, pages 32-34 for detailed assumptions. 11 of Viti Levu in Fiji's population, the likely Figure 3. Estimated Supply and Demand of Water in change in cyclone intensity could cost Viti Levu Western Viti Levu, under a Decreasing Rainfall Scenario as much as US$11 million a year by 2050. Droughts. Fiji experienced four El Niņo­related 160 droughts between 1983 and 1998. The 1997/98 Demand under Population growth: event--one of the worst on record--caused High damages of US$140­$165 million, equivalent to ) 120 Sustainable Supply about 10 percent of Fiji's GDP. The drought (ML Mid e Most likely scenario affected food supplies, commercial crops, m 80 livestock, and the water supply of schools and luoV Demand Low Worst case scenario Projections communities. Droughts of this severity could tera well become the norm in the future. However, W 40 due to the scarcity of economic data on past droughts in Fiji, it is not possible to separate the effects of climate variability from those of 0 1990 2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100 climate change. The incremental impact of Year climate change on drought intensity could only be computed for crop losses (see section C). Note: Assumes future demand to be 300 l/capita/day, 25% loss, yield 98 million l/day. Non-agricultural impacts related to water Sources: JICA 1998 and background studies to this report. shortages and nutrition are believed to be substantial, but could not be quantified. Table 7. Estimated Annual Economic Impact of Climate Change on Water Resources in Viti Levu, Fiji, 2050 Water Supply. Among the most important effects of climate change are the impacts of changes in rainfall on water supply. Models of Category Annual damage two streams in Viti Levu--the Teidamu and (millions of 1998 US$) Nakauvadra creeks--indicate that rainfall variations could cause a 10 percent change in water flow by 2050 and a 20 percent change by Changes in average rainfall + 2100. The direction of the change would Increased cyclone intensity 011.1 depend on whether rainfall increases or decreases. For larger rivers, an increase in Increased severity or frequency of El Niņo ++ rainfall could lead to extensive flood damage. related drought Total 011.1 Figure 3 shows the projected supply and + Likely to have significant economic costs, but impact could not be demand for water in Nadi-Lautoka, a prime quantified. tourism and urban area of Viti Levu which Source: Background studies to this report. serves 123,000 people. Provided the distribution system is fully efficient, the impact of a Table 7 summarizes the quantifiable economic decreasing rainfall scenario would not become impacts of climate change on the water substantial until the second part of the century. resources of Viti Levu. The estimate of US$0- Under a worst-case scenario and moderate $11 million reflects only the average population growth, demand would exceed incremental annual costs of more intense supply by 38 percent by 2100as compared to cyclones; absolute costs in disaster years could an 18 percent shortfall in the absence of climate be much higher, up to US$44 million for an change. The deficit caused by climate change is average cyclone. Given the uncertainties smaller than the amount currently lost to leakage surrounding extreme events--and the difficulty and water losses (29 percent), suggesting that associated with quantifying certain types of more aggressive leak repair would be a logical damages--these estimates should be viewed adaptation strategy. primarily as illustrations of what may happen. 12 C. Impact on Agriculture Figure 4. Likely Impacts of Climate Change on Agriculture in Viti Levu, Fiji Changes in rainfall, temperature, and climate variability will affect agricultural production in Increased Increased Viti Levu. An 8 percent increase in rainfall (as Mean drought magnitude expected in 2050) would benefit most crops rainfall frequency of and tropical except yams, while a drier climate (an 8 percent magnitude cyclones decrease in rainfall) would hurt most crops, Increase Decrease particularly sugarcane (figure 4). The impact of climate change on agriculture in Viti Levu is estimated to cost about US$14 Improved Decreased million a year by 2050 (table 8). This estimate production production of most crops ˇ Sugar reflects annual average costs; damages in an El (but yam Cane Niņo year could be much greater as indicated by production ˇ Taro may decline) ˇ Cassava the 1997/98 drought, which cost Viti Levu some US$70 million in lost crops (UNDAC 1998). The most significant economic damage would Figure 5. Sugarcane Production in Fiji, 19831998 be on sugarcane, which accounts for 45 percent of Fiji's exports and is cultivated primarily in 5000 Viti Levu. But losses of traditional crops, such El Nino-Related Droughts 4500 as yams and taro, could have a substantial effect on subsistence economies in Viti Levu. 4000 3500 Sugarcane. Sugarcane is particularly sensitive to droughts: the 1983 and 1997/98 events, for )T M 3000 example, resulted in a 50 percent loss in 000'( production (figure 5).5 In the future, increases in onit 2500 rainfall during good years may offset the oducrP 2000 impacts of warmer temperatures, with little 1500 change in sugarcane production. However, a warmer--and possibly drier--climate could lead 1000 to more intense droughts during El Niņo years. 500 Using the impact of the 1997/98 drought as representative of the intensity of future events 0 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 and assuming a drought frequency similar to that 981 981 981 981 981 981 981 991 991 991 991 991 991 991 991 991 Year observed in 198398 (one drought every four Source: Official Fiji Statistics. years), the following projections can be made time (5 out of 16 years) as a result of the for the next 25­50 years: residual effects of cyclones and droughts. ˇ Sugarcane production is likely to total 2 million metric tons--just half of output in a ˇ Sugarcane production might total 4 million metric tons--the normal level of output--44 normal year--every four years, or 25 percent of the time (7 out of 16 years). percent of the time (4 out of 16 years). ˇ Sugarcane production is likely to total 3 Under this scenario, the future production of million metric tons--three-quarters of sugarcane could average 3.2 million metric tons output in a normal year--31 percent of the a year, a drop of 9 percent from the 1983­98 average level of 3.5 million metric tons. The resulting economic losses would be about 5The agricultural climate change model used by the study US$14 million a year by 2025­50, assuming (Plantgro) did not provide reliable scenarios for sugarcane. The impacts were thus estimated based on historical data. constant sugar prices. 13 Table 8. Estimated Economic Impact of Climate on Change on Agriculture in Viti Levu, Fiji, 2050 Impact of change in average rainfall Impact of change in rainfall, temperature, and temperature and climate variability (ENSO) Current Change in Change in production Economic Impact average yield Economic Impact average yield Crop (US$ thousands) (US$ thousands) (percent) (US$ thousands) (percent) Sugarcane 147,200 -13,700 -9 Dalo (Taro) 800 -40 ­ +9 -5 ­ +1 -111 ­ +6 -15 ­ +1 Yam 1,600 -76 ­ +63 -5 ­ +4 -164 ­ +54 -11 ­ +4 Cassava 2,100 -189 ­ -105 -9 ­ -5 -242 ­ -128 -12 ­ -6 Total -13,80014,200 -- Not available. Minus signs indicate an economic cost. Plus signs indicate an economic benefit (from rainfall increases). Note: Ranges reflect best-guess and worst­case scenarios under two different climate change models. See annex A for assumptions used. Source: Background studies to this report. Figure 6. Effect of El Niņoinduced Droughts on The impact on the Fijian economy is expected Taro Cultivation Area and Yields in Viti Levu to be substantial, but localized. In 1997/98, for example, a 26 percent decline in sugarcane production value led to a decline in GDP of at least 1.3 percent (Ministry of Finance 1999). More importantly, Viti Levu could suffer a 50 percent drop in sugarcane production every fourth year due to stronger El Niņos. These periodic droughts could well prove to be the most disruptive to the Fijian economy once preferential trade agreements are phased out. 1990 Food Crops. By 2050 climate change may cost Viti Levu some US$70520,000 in lost food crop production (table 8). Projected changes in average climate conditions (temperature and rainfall) would have little effect on dalo production. In El Niņo years, however, the dalo yield could be reduced by 30­40 percent of current levels (figure 6). Yam production would also remain relatively unaffected by changes in average conditions. Current El Nino The response to climate variability is the opposite of dalo. During El Niņo events, production might be expected to remain the same or even increase. Production could decline by nearly 50 percent, however, as a result of wetter or La Niņa­type conditions. This response is consistent with the traditional use of yam and dalo as dry and wet season crops. Cassava output is expected to decline as a result of changes in average climate conditions, with 2050 El Nino yields falling 59 percent by 2050 (table 8). Productivity could also worsen with future Note: Shaded areas show land suitable for cultivation. Source: Background studies to this report. 14 climate variability, particularly under an Quantifying these impacts is difficult, yet it is intensified La Niņa. also vital for the development of appropriate Yaqona (kava) showed little response to climate public health policies. Based solely on the likely change or El Niņo/La Niņa anomalies. Yaqona increase in dengue fever and diarrheal disease, harvests were affected by the 1997/98 drought, the public health impacts of climate change in however. The crop is best suited to upland areas Viti Levu are estimated at US$1­$6 million a in central and southwestern Viti Levu, which are year by 2050. This figure is almost certainly an least affected by drought. This suggests that if underestimate, as it does not take into account production expands into nontraditional areas, the costs of fatalities, injuries, or illnesses from yaqona could become increasingly susceptible to cyclones or droughts; the costs of nutrition- climate variation. related diseases; or the indirect impact of climate change on the poor and the most vulnerable, including infants. D. Impact on Public Health Public Safety. Fiji has lost more than 77 people to cyclones over the past 20 years (table 9). Climate change could have significant impacts Injuries and illnesses caused by extreme events on public health as a result of higher are also believed to be significant. Cyclone Kina temperatures (0.91.3o C by 2050), changes in alone caused 23 deaths in 1992/93, in addition to water supply, and decline in agriculture US$96 million in damages (Fiji Meteorological production. The impacts could include: Services undated). An increase in cyclone intensity, as envisaged, could increase the Direct impacts on public safety, including impact on public safety by as much as 100 injuries, illness, and loss of lives due to percent relative to what is observed today. An cyclones or droughts. average cyclone in the future might come to Indirect effects, such as increased incidence resemble the impacts of cyclone Oscar (1983) or of vectorborne diseases (dengue fever and Eric (1985). malaria), waterborne diseases (diarrhea), and Dengue Fever. Dengue fever is a growing toxic algae (ciguatera). public health problem in Fiji. The most recent Nutrition-related diseases, particularly epidemic--which coincided with the 1997/98 malnutrition and food shortages during drought--affected 24,000 people and left 13 extreme events. Public health impacts are likely to be Table 9. Loss of Lives and Damages from particularly severe for the 12-20 percent of Recent Cyclones in Fiji, 198397 households in Fiji that live below the poverty line (UNDP and Government of Fiji 1997). Poor Cyclone Number of Number of Damages households are more vulnerable to the impacts lives lost people (1998 US$ missing million) of climate change because of their greater propensity for infectious diseases, limited access Oscar 9 -- 76 to medical services, substandard housing, and Eric 25 -- 33 exposure to poor environmental conditions. Storm Gavin 7 3 1 Many of the poor are also landless and Sina -- -- 17 Joni 1 -- 1 (particularly in urban areas) may lack access to Kina 23 -- 96 traditional safety nets that assisted them in times Gavin 12 6 18 of disaster. Poverty is thus both a contributor to vulnerability as well as an outcome of climate- Total 77 9 242 related events. Source: Fiji Meteorological Services. 15 dead, at a cost of US$3­$6 million (Koroivueta, epidemics). personal communication; Basu and others 1999). ˇ A change in seasonality, so that dengue Climate change is expected to cause significant fever outbreaks could occur in any month. increases in the frequency, severity and spatial distribution of dengue fever epidemics. Higher ˇ The emergence of more severe forms of the temperatures would increase the biting rate of disease, such as dengue hemorrhagic fever mosquitoes and decrease the incubation period and dengue shock syndrome, resulting in of the dengue virus. higher fatality rates. In 1990, 53 percent of Viti Levu was at low risk Diarrheal Disease. Diarrheal disease is likely to of a dengue epidemic. By 2100 less than 21 become more common in a warmer and wetter percent of the island, all in the interior world. More intense droughts and cyclones highlands, may remain at low risk. Under the could also increase the incidence of diarrhea by worst-case scenario, up to 45 percent of the disrupting water supplies and sanitation systems. island could be at high or extreme risk of a Quantitative analysis indicates that a 1oC dengue fever epidemic (table 10). The increase in temperature could result in at least economic impact would average about US$1-$6 100 additional reports of infant diarrhea a million a year by 2050 (table 11). month. Since the actual incidence of diarrhea is Climate change could also result in: at least 10 times the incidence of reported cases, a 1oC rise in temperature, as expected by 2050, ˇ A 20­30 percent increase in the number of could lead to 1,000 additional cases of infant cases of dengue fever by 2050 and as much diarrhea a month. These results can be used, as a 100 percent increase by 2100 (under a with lower levels of confidence, to estimate the worst-case scenario). potential impacts of diarrhea in children and adults. The economic costs of climate change on ˇ Dengue fever becoming endemic (that is, diarrheal disease are estimated to average occurring all the time rather than in US$300,000$600,000 a year by 2050 (table Table 10. Potential Impact of Climate Change on Dengue Fever in Viti Levu, Fiji Likely changes Baseline 2025 2050 2100 (1990) Estimated change in number of cases (percentage change) 0% 10% 2030% 40100% Epidemic potential in Viti Levu (in percentage of land area) a Low risk 53% 3839% 2531% 721% Medium risk 47% 6162% 6972% 4872% High risk -- -- 03% 741% Extreme risk -- -- -- 04% Seasonality Nadi Seasonal Seasonal Seasonal to all year All year Suva Seasonal Seasonal Seasonal to extended Prolongued season season to all year Frequency of epidemics 1 in 10 years Likelyincrease Severity of strains Likelyincrease a - Epidemic potential is an index that reflects the efficiency of transmission in a particular area. Note: Ranges represent the most likely and worst-case scenarios in the CSIRO9M2 General Circulation Model. Source: Background studies to this report. For assumptions, see annex A. 16 11). Nutrition-Related Diseases. More intense cyclones and droughts are likely to increase the Other Public Health Impacts. Fiji is presently incidence of nutrition-related diseases, as malaria-free: the strict border controls and subsistence crops and fisheries are affected. The quarantine requirements have so far been impacts may be similar to those experienced successful in keeping the malaria vector during the 1987 and 1997/98 droughts, when (Anopheles) away. Climate change could milk production fell 50 percent and some US$18 increase the risk of malaria, though modeling million in food and water rations had to be results indicate that the epidemic risks in Fiji distributed (UNDAC 1998). Up to 90 percent of due to climate change are small. the population in western Viti Levu required emergency food and water rations. Loss of Climate change could also increase the risk of agriculture income and failure of household filariasis. However, Fiji has started an intensive gardens also caused protein, vitamin, and program to control filariasis and is expected to micro-nutrient deficiency, particularly among eradicate the disease in 510 years (Koroivueta, young children and the poor. personal communication). Table 11. Estimated Annual Economic Impact of Climate Change on Public Health in Viti Levu, Fiji, 2025­2100 (millions of 1998 US$) Event 2025 2050 2100 Cyclones and droughtsa --------------Likely to be substantial------------------ Dengue fever 0.32.3 0.55.5 0.715.9 Diarrheal diseases 0.2 0.30.6 0.62.2 Nutriton-related illnesses + + + Total estimated costs 0.52.5 0.86.1 1.318.1 Not available. + No quantifiable data available, but damages are likely to be substantial. a. The effect of cyclones and droughts on health could not be calculated, though the overall impact of cyclones was taken into account in section B. Note: For assumptions, see annex A. Source: Background studies to this report. 17 18 Chapter 4 Impact of Climate Change on Low Islands The Tarawa Atoll, Kiribati Like most atolls, Tarawa (30 km2) is very The population density of the atoll is unevenly vulnerable to sea level rise. Most of the land is distributed, with South Tarawa (the capital) less than 3 meters above sea level, with an approaching 5,500 people per square kilometer, average width of only 430­450 meters (Lands while North Tarawa remains sparsely populated, and Survey Division undated). While Tarawa with less than 50 people per square kilometer. lies outside the main cyclone belt, it is As available land in South Tarawa becomes susceptible to storm surges and to droughts, scarcer, development in North Tarawa is particularly during La Niņa events. expected to accelerate. Box 4. Tarawa at a Glance Tarawa, Kiribati Climate ˇ Temperature: 2632°C (average 31°C). ˇ Average rainfall in central Tarawa: 2,749 milimeters. Major influences: ˇ Intertropical Convergence Zone. ˇ South Pacific Convergence Zone. ˇ El Niņo Southern Oscillation (ENSO). Extreme events: ˇ Droughts (often associated with ENSO). ˇ Tropical cyclones rarely affect Kiribati (but storm surges do). Population: ˇ Population of Kiribati: 77,658 in 1995 (35,000, or 45% in Tarawa) ˇ Average annual growth rate: 1.4% (3% in South Tarawa). ˇ Population Density: 250/km2 (Gilbert Islands); 1,800/km2 (South Tarawa). ˇ Growing urbanization. Economy (Kiribati): ˇ GDP (1998): US$47 million. ˇ Average annual growth: 4.7% (199096). ˇ Dominated by public sector. ˇ Significant exports: copra, fish. ˇ Subsistence agriculture (coconut, taro, breafruit, pandanus) and fisheries vital. Future Population Trends Future Economic Trends Increasing population density: ˇ Continued dependence on natural resources. ˇ Very high population density on South Tarawa. ˇ Agriculture will continue to be important in subsistence production, ˇ Accelerated population growth in North Tarawa. and will remain a small export earner. Implications: ˇ Cash economy will become increasingly important, but ˇ Denser urban structures. subsistence economy will remain significant. ˇ Industrial and commercial development. ˇ Continued dependence on foreign aid. ˇ Change from traditional housing styles and materials. ˇ Increase in proportion of squatter housing. ˇ Increase in coastal infrastructure. Future Environmental Trends Future Sociocultural Trends ˇ Environmental degradation in densely populated areas. ˇ Increasing role of cash economy. ˇ Continued degradation and irreversible damage to ˇ Changes in food preferences toward imported foods. mangroves and coral reefs. ˇ Increased noncommunicable diseases associated with nutritional ˇ Increased problems of waste disposal (sewage, and lifestyle changes. chemical, and solid waste pollution). ˇ Increased poverty and social problems in urban areas. 19 Tarawa is already becoming Table 12. Estimated Annual Economic Impacts of Climate Change increasingly vulnerable to climate on Tarawa, Kiribati, 2050 change due to high population (millions of 1998 US$) growth rates and in-migration from outer islands, accelerated development, shoreline erosion, Impact Annual Level of Likely cost of and rising environmental damagesa Certainty an extreme eventb degradation. In such a fragile and crowded environment, even Impact on coastal areas small changes can have a large Loss of land to erosion 0.10.3 Low impact. Socioeconomic trends Loss of land and infrastructure to inundation 712 Low 210-430 point to a continuing rise in the Loss of coral reefs and related services 0.20.5 Very Low (storm surge) atoll's vulnerability in the future (box 4). Impact on water resources Replacement of potable water supply due to 13 Low By 2050, under the climate change in precipitation, sea level rise, and change scenarios shown in table inundation 2, Tarawa could experience Impact on agriculture annual damages of about Agriculture Output Loss + Low US$8$16 million (table 12). This estimate takes into account Impact on public health only the potential impacts of Increased incidence of diarrheal disease ++ Low climate change on coastal areas Increased incidence of dengue fever + Low (US$7­$13 million a year) and Increased incidence of ciguatera Low water resources (US$1­$3 Impact of climate change on public safety + Very Low million a year). The cost of and on the poor Potential increase in fatalities due to inundation, + Low several other important and water-borne or vector-borne diseases impacts--such as loss of agriculture crops and effects on Total estimated damages >816+ public health--could not be + Likely to have economic costs, but impact not quantified. Not available. estimated because of aReflects incremental average annual costs of climate change, equivalent here to the capital insufficient data. Indications recovery cost factor of land and infrastructure damaged by inundation, using a discount rate of 10 percent and a 10 year period. suggest that these damages may b Reflects financial damages to land and infrastructure caused by sea level rise and storm surge be substantial. during a 1 in 14 year storm event. For detailed assumptions, see annex A. Source: Background reports to this study. These costs reflect annual average losses due to climate change. In years of strong storm surge, Tarawa only US$47 million. These losses, however, could face capital losses of up to US$430 assume no adaptation. Communities would million in land and infrastructure assets likely adapt to sea level rise by elevating their destroyed by inundation. Relocation of houses or moving further inland, particularly if communities might be needed if the loss of land the changes were gradual. Nevertheless, sea and freshwater supplies become critical. level rise could profoundly affect the economy of Kiribati by inundating the causeways that Climate change is thus likely to place a now link the islets of Tarawa, thus disrupting substantial burden on the people and economy of socio-economic links. Much of the impact of Kiribati. The projected losses could be climate change will ultimately depend on the catastrophic for a country with a 1998 GDP of extent to which proactive adaptation measures are adopted. 20 A. Impact on Coastal Areas Higher rates of erosion could arise if sediment supply decreases, which may happen if coral Climate change is likely to affect Tarawa's coast reefs are weakened by climate change. Even through shoreline displacement resulting from these small changes, however, could cause the rise in sea level (by 0.20.4 meters by significant impacts given the atoll's narrow 2050), through inundation and storm surge, and width and population concentration. through coastal erosion due to the effect of increases in sea surface temperature and sea The islands are expected to become narrower level on coral reefs. and higher, with Buariki facing a shoreline displacement of 30 percent of the island width. To model the impact of coastal erosion and The frequent overwash would result in a build- inundation, two representative sections of the up of sediments in the center of the islands. Tarawa coast were selected: the islands of These sediments would have to be removed, or Buariki and Naa in North Tarawa and Bikenibeu infrastructure would have to be displaced. in South Tarawa (see map in box 4). These areas represent about 20 percent of the area of North Tarawa and about 7 percent of the area of Coastal Inundation. As a result of ENSO South Tarawa. events, Tarawa already experiences significant natural fluctuations of about 0.5 meters in sea Coastal Erosion. Models of shoreline level. These fluctuations will affect the displacement indicate that while all of the atoll's inundation potential of the atoll, particularly islands are undergoing coastal erosion, the loss when combined with storm surges and the of land due to sea level rise is likely to be projected increase in sea level. relatively small--a maximum of 3.2 percent of land by 2100 for Buariki and 3.9 percent for The coastal inundation impacts were modeled by Bikenibeu, leading to economic damages raising the mean high water spring level (the averaging US$0.1-$0.3 million a year by 2050. maximum water level reached during spring Table 13. Likely Impact on Buariki and Bikenibeu, Tarawa, Kiribati of Inundation Caused by Sea Level Rise, 2025, 2050 and 2100 Buariki Bikenibeu Projected losses Projected losses Scenario equivalent to Percentage Structures Roads Percentage Structures Roads Projected rise of land (number) (kilometers) of land (number) (kilometers) in sea level area area (meters) affected affected 0.2 2025 worst-case 18% 196 (59%) 6.55 (77%) 0% 0 0 2050 best-guess 0.4-0.5 2000 baseline + storm surge 30% 213 (64%) 6.55 (77%) 2% 34 ( 2%) 0 2025 best-guess + storm surge 2050 worst-case 2100 best-guess 2050 best-guess + storm surge 55% 229(69%) 7.5 (89%) 25% 423(27%) 1.3(29%) 1.0 2050 worst-case +storm surge 80% 245 (74%) 8.5 (100%) 54% 986 (63%) 2.83( 66%) 2100 best-guess + storm surge 2100 worst-case 1.5 2100 worst-case + storm surge 85% 316 (95%) 8.5 (100%) 80% 1302 (84%) 4.36 (100%) Notes: Storm surges are based on 1 in 14-year event (Solomon 1997). Source: Background studies to this report. 21 tides) by the projected increase in sea level. Figure 7. Projected Inundation of Bikenibeu Island The sea level rise scenarios were also Tarawa, Kiribati under Worst-Case Scenario combined with the effects of storm surges likely to occur once every 14 years (Solomon 1997). The results indicate that under a best- guess scenario, 18 percent of Buariki could be inundated by 2050 (table 13). By 2100 up to 30 percent of Buariki could be inundated. The impact on Bikenibeu would be relatively minor (2 percent inundation). Storm surges, however, could increase damages significantly, with up to 80 percent of Buariki and 54 percent of Bikenibeu inundated by 2050. Projected losses in infrastructure and roads could be substantial. Under a worst-case scenario, the village of Buariki could be inundated by 2050, as could 59 percent of the structures and 77 percent of the roads. In Bikenibeu, significant impact on infrastructure is not be expected to occur until 2100 under a worst-case scenario, but it could then become substantial, with 66­100 percent of all roads lost under the A: Present status combined effects of storm surge and sea B: Residual island under a worst case scenario, 2100; C: Residual island under worst case scenario and storm surge, 2100 level rise (figure 7). Source: Background studies to this report. The projected loss of land and structures is based on existing infrastructure. Actual the rise in sea level will amount to a condition of losses in 2025 and 2050 could be substantially "permanent inundation" which, combined with higher, particularly if Tarawa were affected by the existing standard of housing in Tarawa, is larger storm surges (such as one in 50-100 year likely to result in the loss of structures unless events). The damage could be moderated if adaptive action is undertaken. Other factors in population growth was contained and the analysis--such as more severe storms and redistributed to less urbanized areas. increases in their frequency--are likely to have Extrapolating the losses for Buairiki and been underestimated. Bikenibeu to the rest of North and South Tarawa, the projected financial losses of land Mangroves and Coral Reefs. Tarawa has lost and structure assets in Tarawa could average some 70 percent of its mangroves since the US$7$12 million a year by 2050 (table 14). 1940s, and only 57 hectares remain (Metz 1996). During years of actual storm surges, Tarawa Hence the impact of sea level rise is expected to could experience losses of capital assets be relatively minor. approaching US$210 to $430 million. Healthy coral reefs may respond to increases in The analysis assumes the loss of all land and sea level by growing vertically. In fact, the structures affected by inundation and storm historical accretion rate of the Tarawa reef flat surge, which may overly pessimistic. However, (8 milimeters a year) exceeds the rate of 22 Table 14. Estimated Economic Impact of Inundation Caused by Sea Level Rise in Tarawa, Kiribati, 2050 and 2100 (millions of 1998 US$) Potential Inundation Costs (in US$ million) Scenario equivalent to Annualized lossesa Capital value of lost assets during a storm surge eventb 2050 best-guess 6.6 158.0 2050 worst-case 12.4 374.6 2100 best-guess 2100 worst-case 69.7 497.3 a- Reflects the annual value of the losses, or the capital recovery factor. The costs take into consideration the incremental impact of a 1 in 14 year storm event. b- Reflects the incremental cost of capital losses during a 1 in 14 year storm event. Note: For assumptions, see annex A. Source: Background studies to this report projected sea level rise (Marshall and Jacobsen B. Impact on Water Resources 1985). However, many corals may not be able to adapt to warmer sea surface temperatures and Climate change is likely to affect the water increased concentration of carbon dioxide in the resources of Tarawa through variations in atmosphere, both of which inhibit coral growth. rainfall, evapo-transpiration (caused by a rise in As stated previously, vertical coral growth is temperatures), increases in the sea level, and likely to lag the increase in the sea level by at extreme events (figure 8). Average rainfall is least 40 years. This could create a high-energy expected to either decrease by 811 percent, or window, allowing waves of increasing strength increase by 57 percent by 2050 (most models to reach the shore. In addition, while the reef is predict an increase). The economic losses growing vertically the amount of sediment for resulting from these changes are estimated at island building could decline. about US$1­3 million a year (in 1998 dollars) by 2050. For heavily damaged reefs affected by increased bleaching events a number of Figure 8. Likely Impact of Climate Change on consequences are likely. These include Water Resources in Tarawa, Kiribati further depletion of reef fisheries, failure of the reef to act as an effective buffer of wave energy, and increased island Increased Drought Coastal Effects Frequency & of instability as sediment resources decline. Mean Rainfall Magnitude Climate Change The economic losses of coral reef degradation attributed to climate change Increase Decrease Loss of Increased Inundation would be in the order of US$200,000­ Land $500,000 a year primarily as a result of lost fish production (table 12). This estimate does not include loss of coastal Decreased protection, which was reflected in the Improved Rain Reduced Reduced Water Water Quantity Quality inundation analysis, or other important Resources Collection and reef functions that could not be Storage Groundwater quantified. 23 Tarawa receives less rainfall than surrounding ˇ Changes in sea level. A rise in sea level of atolls. The population relies primarily on 0.4 meters (the worst-case scenario in 2050) groundwater resources, complemented by would have little effect on the groundwater rainfall collected from roofs and desalinated supply and could even raise its volume, as water (a desalination plant has been operational the groundwater table (the top of the since 1999). Population growth and economic freshwater lens) would tend to rise while its development are likely to place considerable base remained relatively unaffected. pressure on these resources over the next However, if the width of the islands were century. reduced by inundation--which is likelythe thickness of the groundwater could decline Water Supply. Models of climate change impact 29 percent. on Tarawa's main groundwater supply (Bonriki) indicate the following: The combined effect of these impacts is shown in table 15. By 2050 if rainfall declines by 10 ˇ Changes in rainfall. By 2050 a 10 percent percent, the sea level rises by 0.4 meters, and the decline in rainfall could cause a 14 percent islands' width is reduced by inundation, the reduction in groundwater recharge. If, by thickness of the groundwater could decline as contrast, average rainfall increased by 7 much as 38 percent. The resulting economic percent, the groundwater recharge would impacts, extrapolating for the Tarawa atoll, increase 5.5 percent. would be on the order of US$1.4$2.7 million a year. If, by contrast, average rainfall increased ˇ Changes in evapo-transpiration. Evapo- in the future, the annual economic costs would transpiration would increase if the climate amount to US$0.7$1.3 million. warmed, but its effects on groundwater recharge would be much milder than the The value of the groundwater was estimated effect of changes in rainfall. A 10 percent based on what it would cost to replace it by increase in annual evapo-transpiration could alternative sources, either through expansion result in a 6 percent decline in groundwater into alternative groundwater sources or through recharge. desalination. A third alternative source-- Table 15. Estimated Annual Economic Costs of Climate Change on Water Resources in Tarawa, Kiribati, 2050 (millions of 1998 US$) Percentage Economic costs of substitution by change in alternative sources groundwater Expansion into Desalination Climate change scenario thicknessa new groundwater sources Current sea level, 7% increase in rainfall +5.5 +0.2 +0.4 Current sea level, 10% reduction in rainfall -14.0 -0.5 -1.0 0.2 meter sea level rise, current rainfall -0.9 -0.0 -0.1 0.4 meter sea level rise, current rainfall +2.0 +0.1 +0.1 0.4 meter sea level rise, 10% reduction in rainfall -12.0 -0.4 -0.8 0.4 meter sea level rise, current rainfall, reduced island width -29.0 -1.0 -2.0 0.4 meter sea level rise, 7% increase in rainfall, reduced island width -19.0 -0.7 -1.3 0.4 meter sea level rise, 10% reduction in rainfall, reduced island width -38.0 -1.4 -2.7 Total costs (under a reduced rainfall scenario) -1.4 to -2.7 Total costs (under an increased rainfall scenario) -0.7 to -1.3 -0.7 to -2.7 Total costs a. Reduction in thickness was modeled for the Bonriki freshwater lens in Tarawa and may not apply to other groundwater sources. Note: Shaded areas indicate the most relevant scenarios of those shown in the table. See annex A for detailed assumptions. Source: ADB 1996 and background studies to this report. 24 rainfall collectors--could be less costly but may Sea level rise could affect agriculture crops in not be able to compensate for the shortages. two major ways: first, through saltwater intrusion, which would affect te babai Extreme Events. The effects of extreme events production in particular. Second, through loss of on the water supply of Tarawa could be coastal land due to inundation, which could significant. Currently, high sea levels during El reduce production of copra, breadfruit, and Niņo years can lead to seawater contamination pandanus. Estimates of the cost of damage could of freshwater lenses. Recovery is generally rapid not be made due to data and time constraints. due to the accompanying high rainfall.6 The higher overtopping and inundation that may D. Impact on Public Health occur in the future, however, could considerably increase the risks of saline contamination. Climate change could exacerbate public health problems in Tarawa. The incidence of ciguatera The impacts of droughts could also be poisoning, diarrheal disease, malnutrition, and substantial, though difficult to quantify. The vectorborne diseases, such as dengue fever, is 1998­99 drought, for example, dried rainwater likely to rise as a result of increased tanks in South Tarawa and caused shallow temperatures and changes in rainfall. groundwater reserves to become brackish (White and others 1999). When rain arrived in March Dengue Fever. There have been four known and April 1999, it contaminated groundwater outbreaks of dengue fever in Kiribati, two wells, causing a high incidence of diarrhea. during the 1970s and two during the 1980s. South Tarawa is at a relatively high risk of C. Impact on Agriculture dengue fever epidemics due to a combination of crowded urban areas, ideal climate conditions Many of the crops grown in Kiribati are affected for the vector (average temperatures of 31oC and by changes in climate. Production of copra--the rainfall of 500 millimeters a month), the main cash crop for about 55 percent of Kiribati's presence of an international airport, and the population--is sensitive to rainfall, as coconuts proliferation of discarded empty bottles and used require annual rainfall of at least 1,000­1,500 tires. millimeters. Te babai (giant taro) is extremely sensitive to reductions in groundwater. Te babai A simple model suggests that the risk of dengue pits are also prone to saltwater intrusion as a fever will increase in the future as a result of result of storm surges and overwash. climate change, with the epidemic potential--an index measuring the efficiency of disease Climate change is most likely to affect transmission--expected to increase 2233 agricultural crops through changes in rainfall. If percent by 2050 (table 16). Most of South wetter conditions prevail, production of water- Tarawa's population would be exposed in the sensitive crops--coconut, breadfruit, and te event of an epidemic. However, while future babai--is likely to increase. If rainfall epidemics could expand faster, the number of decreases, coconut and te babai production will cases would probably not increase from current likely decline. levels. The increased prevalence of all dengue- virus serotypes worldwide could also lead to a Climate variability may also affect agricultural higher incidence of severe forms of dengue production, especially during La Niņa years, fever--in particular dengue hemorrhagic fever when droughts are most likely to occur. and dengue shock syndrome, which can be fatal. Ciguatera Poisoning. Kiribati has one of the highest rates of ciguatera poisoning in the Pacific (Lewis and Ruff 1993). The disease is 6El Niņo increases rainfall in Kiribati. La Niņa events are generally accompanied by droughts. contracted by consuming reef fish that have been contaminated by ciguatoxins. 25 A recent study found a statistically significant in droughts--would increase the incidence of relation between sea surface temperatures and diarrhea, as water shortages exacerbate the reported incidence of ciguatera fish sanitation problems. The projected rise in poisoning in Kiribati (Hales and others 1999). temperature may increase the incidence of This relation was used to model the projected diarrhea, primarily by increasing the likelihood increases in ciguatera poisoning (table 16). The of spoiled or contaminated food. Sea level rise model shows that a rise in temperatures is could also increase the incidence of diarrhea by expected to increase the incidence of ciguatera decreasing the size of the freshwater lens, poisoning from 35­70 per thousand people in exacerbating overcrowding conditions, and 1990 to about 160­430 per thousand by 2050. disrupting sanitation and water supply. These results should be interpreted cautiously, as Tarawa has experienced cholera outbreaks in the the model is based on many uncertainties and past. It is possible that increased temperatures limited data. The overall impact of climate may enhance the pathway of cholera change on ciguatera should perhaps be measured transmission through the high level sewage not in terms of incidence rates but in terms of contamination in Tarawa's coastal waters. how people respond to the increased risk (Ruff and Lewis 1997). This may include changes in Indirect Public Health Effects. The indirect diets, decreased protein intake, increased public health effects of climate change could be household expenditures to obtain substitute far-reaching. They could include increases in proteins, and loss of revenue from reef fisheries. malnutrition due to losses of subsistence In addition, reef disturbance has been linked to agriculture and fisheries; deterioration in ciguatera outbreaks (Ruff, 1989; Lewis 1992), standards of living due to impacts on primary suggesting that improved management of coastal sectors; loss of land and infrastructure, leading areas would be an important adaptation strategy. to increased crowding and land shortages; and the immense economic, social, and cultural Diarrheal Disease. Increased rainfall would impacts associated with population relocation if likely result in a reduction in the overall rate of it was required as a result of inundation or water diarrhea due to improved water quality and shortages. These diffuse effects could well prove availability (though flooding may also lead to to be the most important impacts of climate groundwater contamination). Decreased change on the public health of the atoll. rainfall--particularly if it resulted in an increase Table 16. Estimated Increases in Dengue Fever Epidemic Potential and Incidence of Ciguatera Poisoning in Kiribati as a Result of Climate Change, 2025, 2050, 2100 Impact Baseline 2025 2050 2100 1990 Dengue fever Projected epidemic potentiala 0.18 0.20 0.220.24 0.250.36 Percentage change from 1990 n.a. 11 2233 39100 Ciguatera poisoning incidence (per thousand population) 3570 105240 160430 2451,010 a- The epidemic potential index measures the efficiency of disease transmission. A value of 0.2 or above indicates a high epidemic potential. n.a. ­ Not applicable. Note: Ranges indicate best-guess and worst-case scenarios. Changes in atmospheric temperatures were used as a surrogate for sea surface temperature in forecasting the incidence of ciguatera. The model assumed a reporting case rate of 10­20 percent. Source: Background studies to this report; Hales and others (1999). 26 Chapter 5 Impact of Climate Change on Regional Tuna Fisheries Climate change is likely to affect regional tuna fisheries in two major Box 5. Tuna Fisheries and Climate Variability ways: by raising average ocean The distribution of tuna fisheries is affected by the location of the Western Pacific surface temperatures to levels Warm Pool (WPWP), an area of warm surface waters (more than 28oC) that produces currently experienced during virtually all of the tuna caught by purse seine (a fishing method used to collect medium-intensity El Niņos surface tuna for canning), while catch of tuna by longline (a method used to collect (Timmermann and others 1999) and deep water tuna for the sashimi market) is more widely distributed over the whole by increasing year-to-year climate tropical and sub-tropical Ocean. By itself the WPWP is nutrient poor. By contrast, the colder waters of the central equatorial Pacific generate an upwelling of colder, variability. Such change may not nutrient-rich waters. These two ocean areas meet in a zonal band called the "cold have an equivalent today. The tongue," the primary productivity of which is strongly affected by ENSO variability. impacts are likely to be pervasive, During El Niņo years the WPWP can extend eastward into the central Pacific by affecting the distribution, abundance, nearly 4,000 kilometers. and catchability of tuna fisheries Tuna fisheries, particularly skipjack fisheries, move with the WPWP. During El Niņo (box 5): years, countries in the Central Pacific, such as Kiribati and Samoa, experience higher purse seine catches. Countries in the Western Pacific, such as the Solomon Islands ˇ Decline in primary productivity. and Marshall Islands, enjoy higher catches during La Niņa years. Primary productivity in the In addition to this geographical displacement, El Niņo also influences the abundance central and eastern Pacific would of tuna. El Niņo years tend to result in higher than average abundance of skipjack a decline due to the increased few months later, while La Niņa years generally result in higher abundance of adult stratification between warmer albacore in the subsequent years. Yellowfin and bigeye abundance are also likely surface waters and colder deeper influenced by the ENSO variability. However, as these species are more widely distributed and have extended spawning grounds in both east and west tropical water (and consequent reduction Pacific, the relationship with ENSO is more complex. in upwelling). Productivity in the western Pacific could rise. Movement of Tagged Skipjack Tuna in the Central and Western Pacific ˇ Decline in tuna abundance. The decrease in upwelling would lead to a decline in the bigeye and adult yellowfin population (the species targeted by the longline fleet). The abundance of purse seine­caught skipjack and juvenile yellowfin is not expected to be affected. ˇ Increased pressure on longline fishing. Given the continued high demand for sashimi and the possibility that prices may rise with a decline in catches, it is likely that longline fishing pressure on adult yellowfin tuna will increase to compensate for Source: Lehodey et al. (1997); Lehodey (in preparation) 27 the decline in adult bigeye abundance, leading to unsustainable exploitation if the Box 6. The Likely Future Climate fishery is not well managed. Correspondence with ˇ Spatial redistribution of tuna resources. Likely future climate present climate The warming of surface waters and the decline in primary productivity in the central Mean state Moderate El Niņo and eastern Pacific would result in a Moderate El Niņo event Strong El Niņo event redistribution of tuna resources to higher Strong El Niņo event Unknown, extremely latitudes (such as Japan) and toward the warm event western equatorial Pacific. Moderate La Niņa Current mean state Strong La Niņa Moderate La Niņa ˇ Increase in climate variability. Climate change could increase the intensity and frequency of annual climate variability ˇ Higher impact on domestic fleets. Distant (Jones and others 1999). The likely impact water fishing fleets should be able to adapt would be an increase in the annual to changes in the spatial distribution and fluctuations of the spatial distribution and abundance in tuna stocks. But domestic abundance of tuna. It is possible that more fleets would be vulnerable to fluctuations of frequent cold events (such as strong La Niņa tuna fisheries in their Exclusive Economic episodes) could compensate for the decrease Zones. Countries in the central Pacific, such in productivity under an El Niņo mean state. as Kiribati, are likely to be more adversely In addition, even though it is difficult to affected than those in the west. Kiribati's know what a strong El Niņo would mean in high dependence on tuna fisheries renders it the future (box 6), it is likely that such an the more vulnerable to these changes, and extreme event could lead to a dramatic points to the need to closely collaborate with decline in productivity in the eastern Pacific. other coastal states in minimizing the impact of year-to-year fluctuations. 28 Chapter 6 Toward Adaptation: Moderating the Impact of Climate Change The economic costs of climate change estimated in chapters 3 and 4 assume no adaptation. In Box 7. Can Climate Change Be Stopped? practice, Pacific Island governments and communities could help offset these costs by Carbon dioxide in the earth's atmosphere is expected to undertaking adaptation measures. The question double by 2050-2100, leading to changes in temperature, rainfall, and sea level rise. is determining which adaptation measures are best in the face of uncertain future impacts. Could the climate then stabilize? It does not appear so. Even if all the major countries signed the Kyoto Protocol There is little Pacific Islands can do to prevent and succeeded in stabilizing emissions by 2010, the climate change (box 7). At the same time, doubling of carbon dioxide concentration in the Pacific Island governments cannot afford to atmosphere may only be delayed by a decade or so. ignore the problem. Adapting to climate change Stabilizing emissions does not yet mean stabilizing the concentration of greenhouse gases in the atmosphere. may soon become an economic and political Furthermore, after the concentrations stabilize, the rise in imperative. sea level could continue for several centuries (Church and Gregory 2000). Adapting to these changes will A. The Need for Immediate therefore be of paramount importance to countries on the receiving end of climate change. Action The development choices made by Pacific Island change, acting now to reduce current governments today will have a profound impact vulnerability will also prepare the Pacific Islands on the future vulnerability of the islands and on for the long-term effects of climate change. the magnitude of climate change impacts. Another reason for acting now is that failure to One of the most compelling arguments for do so may result in a loss of opportunities that acting now is the rising impact of extreme may not exist in the future. Coral reefs, for weather events in the Pacific. Even those who example, may not be able to recover from argue that climate change may never happen bleaching events if they are weakened by cannot dispute the urgency of reducing the threats such as pollution and mining. islands' vulnerability against severe climate events. The recent drought and the sequence of Finally, adaptation strategies may require cyclones which affected many Pacific Islands several decades to be discussed and during the 1990s attest to an increasing exposure implemented. Communities living in low-lying that will, sooner or later, put mounting public areas, for example, may need to relocate further pressure on governments and politicians to act. inland into other communities' customary land. No less compelling is the fact that under an This will require extensive public debates on increasing globalized economy, those countries how to place the common good of all above the which invest early on adaptationand, in the good of the clan or immediate family, a process process improve the quality of life and reduce that cannotand should notbe rushed. investment risksare likely to hold a competitive advantage for foreign investment. Since it is difficult to predict far in advance how As measures to reduce vulnerability are also climate change will affect a particular site, among the most effective in adapting to climate Pacific Island countries should avoid adaptation 29 measures that could fail or have unanticipated different than expected, investments in these social or economic consequences if climate measures could have been wasted. change impacts turn out to be different than 2. Level of implementation. Adopt general anticipated (IPCC 1998). More appropriate will be 'no regrets' adaptation measures that would be rather than site-specific measures, at least justified even in the absence of climate change. until there is more certainty about localized impacts. These include, for example, sound management of coastal areas and water supplies, control of 3. Bottom up or top-down. Use community- pollution, and investment in preventive health. based (bottom-up) rather than top-down As it will be shown, a 'no-regrets' adaptation interventions. Many traditional adaptation strategy need not involve large investments of measures have been tested and adjusted over public resources but it will require strong the years in response to extreme events. political will, as adaptation measures may face These measures are likely to be more stiff competition from other development effective than top-down solutions. At the activities for scarce funds. Yet it is important to same time, communities will need external understand that the short-term economic gains of help to handle threats--such as a 'do nothing' strategy could be easily dissipated pollutionthat are beyond their control. A by the impact of future climate events. collaborative partnership between the government and communities may prove to A development path that takes adaptation into be the most effective (see volume III of this account might sacrifice some potential short- report). term gains in favor of more diversification and a reduction in vulnerability. But it would vastly 4. Environmental impacts. Select adaptation decrease the downside costs should climate measures based on their impact on the change scenarios materialize. The challenge overall vulnerability of the islands, not only will be to find an acceptable level of risk an on their impact at a particular site (de Wet intermediate solution between investing in high 1999). A sea wall, for example, may solve cost solutions and doing nothing and start the problems of a particular site but increase adapting long before the expected impacts occur. erosion downstream. 5. Cultural acceptability. Ensure that measures are compatible with the socio- B. Guidelines for Selecting cultural traditions of local communities and Adaptation Measures do not cause social disruption. Pacific Island countries have a vast array of 6. Timing. Time measures appropriately. adaptation measures at their disposal. The Some adaptation measures--such as following criteria may help guide their selection: expansion of rainwater collectors in Tarawa--may need to be implemented immediately. Others could wait while 1. No regrets. Give priority to `no regrets' appropriate responses are developed. As a measures, such as water resource general rule, the most urgent measures are management, which would be beneficial those needed to protect against current even in the absence of climate change. climate events and those on which it may no Structural measuressuch as sea walls and longer be possible to act in the future. groynes, which provide few benefits other than protection--require a high degree of 7. Cost-benefit. The potential benefits of certainty about the impact at a particular adaptation measures should clearly exceed site. If climate change impacts turn out to be their costs. 30 Two key principles should be kept in Figure 9. A Seawall in Qoma, Fiji mind when selecting adaptation options. First, adaptation is not necessarily limited to interventions that reduce climate change impacts. Measures that increase the resilience of natural systems--by controlling pollution's effects on coral reefs, for example--should also be considered, as should policies that facilitate action on adaptation, such as a legislation empowering communities to manage their own coastal resources. Second, it is vital to consider the sociocultural conditions of the Pacific Islands. To an external observer, it may seem appropriate to reinforce Sea walls are built throughout the Pacific to protect settlements against coastal traditional Samoan houses to protect erosion and storms. However, sea walls do not solve the underlying cause of against cyclones. From the local erosion and may cause further problems downstream. In Qoma, Fiji (photo above) the community reported experiencing frequent inundation, which might communities' point of view, however, have been exacerbated by their sea wall. Strategic replanting of mangroves might a `do nothing' strategy may well be well have been a more efficient solution to guard against periodic inundation. justified, because labor and materials might be readily available from within the extended family and the houses might be ˇ Management of coral reefs and mangroves. easily rebuilt following cyclones. The adaptation Adaptation strategies should involve process thus needs to be highly participatory and community leaders in enforcing penalties for allow for adjustments as new knowledge about reef and mangrove destruction, controlling climate change impacts is obtained. pollutants, promoting sources of construction materials other than coral, and replanting mangroves. Structural adaptation C. Adaptation Options measures-- such as groynes or seawalls-- should be screened for their compatibility Table 17 lists possible adaptation options for with coral reef management. Pacific Island countries in accordance with the guidelines outlined above. The options include ˇ Protection of towns. Construction of the following: seawalls is likely to be the measure of choice to prevent erosion in densely Adaptation Options for Coastal Areas populated coastal areas. However, seawalls do not resolve the underlying cause of A coastal zone management framework that is erosion, and they can promote offshore tailored to the sociocultural conditions of each movement of beach sediments (figure 9). island should be used for adaptation planning. They are also costly to build and maintain, This framework should have three major goals: and they will need to be extended as the sea preventing loss of lives and property, avoiding level rises. Seawalls should be used only to development in inundation-prone areas, and protect valuable property and buildings that ensuring that critical coastal ecosystems, such as cannot be relocated. For new infrastructure, coral reefs, are protected and remain functional. the use of setbacks and relocation could be Specific adaptation options could include: considered. 31 ˇ Land use policies. Land use policies should ˇ Population relocation. If all other measures encourage settlements away from low-lying fail, population relocation may need to be and high-risk coastal areas through, for considered. While some communities may example, the use of coastal hazard mapping opt to move on their own, population (as currently developed in Samoa). relocation would pose immense social and political risks for Pacific Island ˇ Prevention of erosion. Depending on the governments, as nearly all inhabitable land infrastructure and population density, is under some form of customary ownership. adaptation options to prevent coastal erosion may include (i) no response, where there is little habitation or infrastructure; (ii) Adaptation Options for accommodation, where property is replaced Water Resources as it is damaged; and (iii) shoreline protection, in areas with large populations The uncertain impacts of climate change on and significant infrastructure. In low islands rainfall call for adaptation measures that take or atolls, where it is essential to retain into account both drought and flood control. In overwash sediments, options might include arid islands in particular, it will be vital to replantation of mangroves, pandanus, and improve the management of existing water other coastal vegetation to promote resources and to develop supplementary sources shoreline accretion, closing or narrowing of supply. Interventions could include: selected passages between the lagoon and the ocean, and the strategic use of groynes to ˇ Leakage control. Current rates of water help minimize the transfer of sediments leakage--29 percent in Western Viti Levu from the ocean side to the lagoons. Groynes, and 50 percent in Tarawa--could be however, should be used only in key considerably reduced through improved locations--such as the passage edges of plumbing. Spring-loaded taps and islands--as they tend to cause downstream communal tanks and stand pipes may also erosion and require continuing maintenance. help reduce wastage. In less developed areas the use of setbacks to control future development, beach ˇ Water conservation incentives. The nourishment and relocation of infrastructure introduction of water fees and metered might be preferable. consumption--as done in Tonga--could help discourage high levels of water use. ˇ Protection against inundation. On islands Licenses issued to large water users should with little infrastructure, the costs of require that water be conserved during protection are likely to be prohibitive, and droughts and should impose strict penalties relocation or modification of structures to for unauthorized connections. accommodate surface flooding could be considered. On the more populated atoll ˇ Watershed management. In high islands islands--such as South Tarawa in Kiribati, such as Viti Levu, management of water Majuro and Ebeye in the Marshall Islands, resources should be combined with land and Funafuti in Tuvalu--strategies to allow management in the form of reforestation, overwash sediment to naturally increase the protection of wetlands, and soil elevation of the island may help offset the conservation. This could be facilitated by impacts of inundation. Where land consolidating water and catchment ownership disputes are not an issue, new management responsibilities under a single structures should be set back from the authority. shoreline and elevated to allow for periodic flooding. 32 ˇ Development of alternative sources of water. ˇ Promotion of land use planning. Wider On arid islands, particularly on atolls, promotion of land use planning and alternative water sources may need to be improved seasonal forecasting, needs to be developed. Rainwater collection could be part of a wider `adaptation package'. promoted by fitting new buildings with Mapping of soil and climate zones, underground cisterns and encouraging all particularly in high islands, would improve new houses to be fitted with rainwater the matching of crops and land use practices. storage. Desalination should be considered only when rainwater or groundwater sources ˇ Importation of food may be increasingly are insufficient, as the cost--about US$4 per required to handle the effects of droughts cubic meter--remains high. Future and cyclones. technological breakthroughs may help make desalination more affordable. Water Adaptation Options for Public Health importation is not considered a viable alternative due to the high costs--about Adaptation strategies to minimize public health US$19 per cubic meter--and shipping risks impacts do not require extensive new (ADB 1996; Shalev 1992). interventions. Rather, existing initiatives that reduce the vulnerability of the population, and ˇ Flood control. In islands with extensive particularly the poor, should be enhanced. rivers (such as Viti Levu) flood control Actions should include not only improving measures might include widening and public health but also strengthening the diverting channels, retarding basins, and resilience of the ecosystems on which the building weirs (JICA 1998). The risk of population depends for food and income. flood damage could also be reduced by Specific measures could include: regulating development on flood plains and promoting flood-proof housing. ˇ Integrated adaptation strategies. Adaptation strategies should include a range Adaptation Options for Agriculture of interventions to reduce the vulnerability of the population, such as improved Adaptation strategies for the agricultural sector sanitation and water supply, management of should focus on `no regrets' measures that also solid and liquid waste, protection of help reduce the adverse impacts of extreme groundwater, reduction of poverty weather events. These include the following: (particularly among urban squatter settlements), increased access to primary ˇ Climate-proofing farming systems. These health care, and protection of subsistence could be promoted through research, food supplies. Many of these measures enhancement, and promotion of traditional would also help control the incidence of land management practices, including diarrheal disease. dry/wet season crop rotations and breeding for drought tolerance. ˇ Control of dengue fever. Adaptation strategies should include further support to ˇ Promotion of sustainable production vector control programs that collaborate systems. Sustainable production systems with communities to reduce mosquito include agroforestry and cover crops to breeding sites. They should also improve improve soil fertility, conserve moisture, epidemic preparedness through vector and prevent soil erosion (FAO 1999). This is monitoring, early warning systems, and especially recommended in high islands better preparation of primary health care such as Viti Levu. facilities to treat dengue hemorrhagic fever and dengue shock syndrome. 33 Control of ciguatera. In countries affected Governments cannot do it alone. Adaptation by ciguatera, adaptation measures should measures are and will continue to be include control of non climate-related implemented primarily by communities, the threats to coral reefs (such as pollution and private sector, and individuals. But the role of blast fishing), monitoring of ciguatoxic Pacific Island governments will be essential in areas, and public awareness of the risks of mainstreaming adaptation into policy and consuming the heads, roe and viscera of reef development planning, in creating partnerships fish. with communities, nongovernmental organizations (NGOs) and the private sector, and Adaptation Options for in dealing with problems only the government Tuna Fisheries can handle (such as disaster management). Mainstreaming Adaptation In the short-term, Pacific island nations need to reduce their vulnerability to fluctuations in the Adaptation goals need to be identified as a clear tuna catch of their Exclusive Economic Zones. priority in national policies and development This could involve: plans. Of particular importance will be the role ˇ Stronger regional collaboration in the of the Departments of Health, Environment, negotiation of multilateral agreements with Agriculture, Public Works, and Fisheries. distant water fishing nations (see Volume Conflicts among these agencies' development III, Chapter 3). and adaptation goals--such as the impact of sand mining licensing on coastal management ˇ Income smoothing mechanisms for license programs--need to be addressed. The objective fees. would be to transform climate change from "something that may happen in the future" to a ˇ Better use of ENSO forecasting, to help priority feature of current development planning. prepare countries for spatial and temporal changes in tuna distribution. In the short to medium term, all major new development projects--such as coastal mining ˇ Diversification of domestic fleets, and and dredging--should undergo adaptation eventual reduction of the fishing effort to screening. This process should assess both the adjust to increased fluctuations in tuna likely impact of climate change on the project, resources. as well as the project's impact on the islands' vulnerability (de Wet 1999). Adaptation In the long term it will be essential to strengthen screening need not require extensive new the management of bigeye and yellowfin tuna legislation but rather a revision of environmental stocks, which appear most threatened by future impact assessments to take adaptation into climate change. Since declines in tuna fisheries account. The Coastal Hazard Mapping program are likely to shift the domestic fleet's fishing in Samoa is a step in this direction. pressure to overexploited coastal resources, measures to improve coastal management are Building Partnerships also urgently needed. In building partnerships with communities, D. Implementing Adaptation individuals, and the private sector, the government may need to play a pivotal role in the following areas: The previous sections argued for Pacific Island governments to encourage `no regrets' ˇ Creating an Enabling Policy and Legal adaptation. But how should this be implemented Framework. This may include prioritizing in practice? adaptation into national planning, 34 Table 17. Selected Examples of Adaptation Measures Bottom up Negative Goal Adaptation measure No Level of implementation or top Environmental Culturally Timing Cost- regrets? down impacts? acceptable? benefit Moderate impacts on coastal areas Protection of critical ecosystems Increase Public awareness Generic Both No Yes Immediate Positive Prohibit extraction of reef and sand Yes Sector specific Both No May increase Immediate Positive building costs Prevent mangrove removal Yes Sector specific Both No Unknown Immediate Positive Control pollution Yes Generic Top down No Unknown Immediate Unknown Control overfishing Yes Sector specific Both No Loss of food Immediate Positive Protection of towns and property Engineered structures (such as seawalls) No Site specific Top down Probably Unknown Unknown Unknown Set back development from shoreline No Site specific Both Unknown Land tenure? Can wait Unknown Raise structures No Site specific Both Unknown Unknown Can wait Unknown Land use policies Coastal hazard mapping Yes Site specific Top down No Yes Immediate Unknown Control of erosion Mangrove replantation Yes Sector specific? Both No Yes Immediate Positive Engineering works in passages No Site specific Top down Probably Unknown Can wait Unknown Groynes No Site specific Top down Probably Unknown Immediate Positive(?) Moderate impacts on water resources Water resource management Leakage control Yes Sector specific Both No Yes Immediate Positive Pricing policies (fees, levies, surcharges) Yes (?) Sector specific Top down No Problematic Immediate Positive Conservation plumbing Yes Sector specific Both No Unknown Immediate Positive Stricter penalties to prevent waste Yes (?) Generic Top down No Resistance? Immediate Positive Catchment management Reforestation, soil conservation Yes Generic and site specific Both No Yes Immediate Positive Establishment of a Water Authority Yes Sector specific Top down No Unknown Immediate Positive Alternative water supply Expansion of rainwater collection Yes Sector and site specific Both Unknown Maybe Immediate Unknown Alternative groundwater use Yes Sector and site specific Top down Unknown Land tenure? Can wait Unknown Desalination No (?) Sector and site specific Top down Unknown High costs Can wait Unknown Importation No (?) Sector specific Top down No High costs Can wait Negative Flood control Diversion channels, weirs, etc. No Site specific Top down Probably Unknown Immediate Unknown Land use controls, flood proof housing No (?) Site specific Both No Land tenure? Immediate Unknown Moderate impacts on agriculture Community sustainability programs Traditional weather-resistant practices Yes Sector specific Bottom up No Yes Immediate Positive Sustainable production systems Agroforestry, water conservation Yes Sector specific Both No Unknown Immediate Positive Research Flexible farming systems Yes Sector specific Top down No Unknown Immediate Positive(?) Land use policies Mapping of suitable cropping areas Yes Generic Top down No Unknown Immediate Positive Avoid cultivation on marginal lands Yes Site specific Top down No Disruptive ? Positive Moderate impacts on public health Integrated adaptation strategies Poverty reduction programs Yes Generic and site specific Top down Unknown Yes Immediate Positive? and control of diarrheal disease Improved sanitation and water supply Yes Sector and site specific Both No Yes Immediate Positive Waste management Yes Sector and site specific Both No Unknown Immediate Positive Protection of groundwater Yes Sector and site specific Both No Unknown Immediate Positive Squatter settlement management Yes Site specific Both Unknown Yes ? Immediate Positive Control of dengue fever Community-based vector control Yes Sector and site specific Bottom up No Unknown Immediate Positive Improved preparedness (monitoring) Yes Sector specific Top down No Yes Immediate Positive Prevention of exposure Yes Sector specific Bottom up Unknown Difficult? Unknown Unknown Control of ciguatera poisoning Reduce destructive practices to Yes Sector specific Both No Food, income? Immediate Positive coral reefs Monitoring and public awareness Yes Sector specific Both No Yes Immediate Positive Moderate impacts on tuna fisheries Stronger regional collaboration Multilateral agreements Yes Sector specific Top down Unknown Distrust? Immediate Positive Research Better ENSO forecasting Yes Generic Top down No Yes Immediate Positive Improved tuna management Yes Sector specific Top down No Yes Immediate Positive Fleet management Diversification of domestic fleets No Sector and site specific Top down Unknown Problematic Can wait Positive 35 harmonizing conflicting sectoral policies, E. Funding Adaptation and providing the necessary legal and technical support for community-based Much of the costs and success of adaptation will adaptation measures such as co-management depend on the extent to which communities, in coastal areas. individuals, and the private sector own and implement the strategies. This requires ˇ Strengthening Institutions. Government government support for community-based planning in Pacific Island countries is often efforts, and may require working through sector-oriented, with little capacity to traditional decision making processes to ensure respond to local level needs and conditions. `buy-in' at the local level. By asking new Where this is the case, there is a need to development projects to follow adaptation strengthen the links between local standards, Pacific Island governments could also communities and national and regional shift part of the costs of adaptation to private governments so that the communities investors. increasingly gain a voice in planning and budgetary decisions. Local communities `No regrets' adaptation measures do not involve should also be encouraged to work across significant costs if initiated sufficiently early. village boundaries to reach consensus on the Samoa's environmental health program, for adaptive strategies that need to be applied to example, operates with a budget of US$113,000 larger areasparticularly if relocation is a year. The Coastal Zone Management Project likely to be needed. in Majuro, financed by UNDP, cost US$367,000 for four years of operation. By contrast, sea ˇ Supporting Collaborative Programs. walls surrounding the Tarawa atoll would Community-based programs such as vector require capital investments of about US$1.5- control, water conservation, coastal $1.8 million (table 18). management, or mangrove replantation will need the support of external partners such as In this context, it is recommended that Pacific the government or NGOs. At first, external Island countries adopt urgently a `no regrets' support should focus on galvanizing policy aimed at decreasing their present community action. Later, it should shift to vulnerability to extreme weather events (which technical advice and assistance in areas may exist independently of climate change). As communities cannot handle on their own. a first step, Pacific Island governments should assess how public expenditures could be ˇ Mobilizing Public Action. Public adjusted to support this strategy, and how other awareness and discussion forums involving partners in the processin particular community representatives could help communities and the private sectormay help convey information about the impacts of defray the costs. As a second step, Pacific climate change and gain consensus on the Island governments and donors should study adaptation options. Of special importance how to reallocate or attract new development aid would be awareness efforts aimed at to fund `no regrets' activities that cannot be community leaders. adequately funded by public expenditures. The recently agreed "Pacific Islands Framework for ˇ Handling Disaster Mitigation and Action on Climate Change, Climate Variability Providing Public Services. Some adaptation and Sea Level Rise" (SPREP 2000) could be measures will need to rely on government used as a basis to prioritize donor assistance. interventions. These include early warning Many `no regrets' interventionssuch as systems and disaster mitigation programs, improved sanitation or coastal improvements in primary health care, and managementcould be justified as part of coastal protection in town areas. regular environmental assistance. 36 Even though `no regrets' Table 18. Indicative Adaptation Costs (US$) measures have the double benefit of reducing short-term exposure Measure Cost to climate variability as well as Annual Operational Costsa: long-term vulnerability to climate Land use planning 33,700 change, it is important that the Waste management 181,900 two aspects be kept separate in Biodiversity protection and natural parks 167,000 international negotiations. Environmental education and information 102,500 Adoption of an early `no regrets' National disaster council 30,700 Reforestation 297,800 strategy by a country should not Watershed projection and management 113,800 diminish its chances of accessing Support to community-based fisheries management 81,400 climate change adaptation funds Community disease control 205,800 in the future. Environmental health 112,600 Nutrition 83,400 Similarly, donors should not be Investment Costs: led to believe that because `no Human waste management (composting toilets)b 800,000 regrets' adaptation benefits the Elevating houses b 1,700,000-3,200,000 countries independently of Seawalls c 1,540,000-1,830,000 climate change, the justification Coastal Zone Management Project for Majuro Atolld 367,300 for incremental financing is weak. To do so would be to tip the scale a Costs reflect Samoa 1999-00 public expenditure allocations. GDP Samoa US$205 million. b Covering North Tarawa (population 6,000, area 1,500 ha). GDP Kiribati US$47.9 million in favor of structural solutions c Covering Tarawa atoll (population 35,000, area 3,200 ha). The cost per linear (such as seawalls), which are meter is about US$155, excluding maintenance costs. d clearly incremental. One of the Costs represent allocation for four years for Majuro (population 86,110). Source: Legislative Assembly of Samoa 1999; UNDP 1996; background studies to this report. reasons communities like sea walls is that they can receive government support for their construction. Pacific Island funds from the Global Environmental Facility government officials have often expressed the (GEF), the main financing mechanism for view that it is easier to obtain international aid climate change, have been available only for for structural measures than for `no regrets' mitigation of greenhouse gas emissions and for solutions. These disincentives need to be studies and capacity building done in the context urgently addressed in future international of national communications.7 International climate change discussions, in order to maintain negotiations under the Conference of Parties of `no regrets' strategies at the forefront of the UNFCCC have not yet agreed to the adaptation financing, and benefit rather than financing of actual adaptation (Stage III) penalizethe countries most willing to take measures. early action The international debate on financing of Pacific Island countries are understandably adaptation has not progressed far. Globally, the concerned about the slow pace of these United Nations Framework Convention on negotiations. Since they contribute only a Climate Change (UNFCCC) provides the negligible amount to the world's greenhouse gas umbrella agreement for mitigation of greenhouse emissions, they view the stalling of Phase III as gas emissions. The Convention also includes a way for emission-producing countries to avoid provisions to begin work on adaptation to recognizing their responsibilities toward climate change. To date, however, progress on adaptation has been slow. Many observers feel 7 that the perceived high costs of adaptation may National assessments of vulnerability and adaptation. have curbed enthusiasm to assist those countries National communication strategies have been supported by the Pacific Islands Climate Change Programme (PICCAP), most in need of support. As a consequence, funded by UNDP through the South Pacific Regional Environmental Programme (SPREP). 37 countries on the receiving end of climate community to move urgently with a financing change. mechanism to help coastal states defray these costs. The urgency of this action for small island Other funding mechanisms may be available states such as the Pacific Islands cannot be over- sooner. One of the most promising sources is the emphasized. Clean Development Mechanism (CDM) under the Kyoto Protocol.8 A share of the CDM At the same time, Pacific Island countries should proceeds is envisaged to help vulnerable continue to speak with one voice at international countries meet the costs of adaptation. The climate change forums. Much has been timing of this `CDM tax' will depend to a large accomplished already under the support of the extent on the entering into force of the Kyoto Pacific Island Climate Change Programme Protocol, however, and it is unlikely to be (PICCAP). A strengthened focus on optimal available in the short term. adaptation strategies and economic analysisparticularly on the costs and benefits The findings of this report clearly show that the of adaptation measurescould strengthen their Pacific Island countries are likely to experience case in international negotiations, broaden the significant incremental costs associated with climate change constituency, and mainstream global climate change in the future. The climate change into the economic and responsibility is now on the international development planning of the Pacific Islands. 8The Kyoto Protocol, launched in 1997, is a commitment to decrease world emissions of major greenhouse gases by at least 5 percent below 1990 levels by 2008­12. The Clean Development Mechanism is a process to promote joint reduction of greenhouse emissions by developing and industrial countries (ENB 1999). 38 Chapter 7 Summary of Key Findings and Recommendations The following conclusions can be derived from Based on these conclusions, a number of key the analysis: recommendations can be derived. The Pacific Islands are already experiencing severe impacts from climate events. This is evidenced by cyclone damage of more than Pacific Island Governments US$1 billion during the 1990s and by the impact of recent droughts in Federated States § Adopt a `No Regrets' Adaptation Policy. of Micronesia, Fiji, Kiribati, Marshall Pacific Island governments should put in Islands, and Palau (SPREP 2000). place an urgent policy of 'no regrets' adaptation, aimed at increasing the natural The islands' vulnerability to climate events is resilience of the islands and reducing their growing, independently of climate change. vulnerability to present-day weather events. Current trends point to a continuing rise in 'No regrets' measures could include, for vulnerability in the future which will be example, the management of critical coastal exacerbated by climate change. ecosystems (such as coral reefs), control of urban pollution, water conservation, culture Climate change is likely to impose major of weather-resistant crops and disease vector incremental social and economic costs on control. Under such a policy, Pacific Island Pacific Island countries. In disaster years the governments would take adaptation goals impact could be particularly high, causing into account in future expenditure and significant economic and social problems. development planning. Insofar as these measures helped reduce existing vulnerability Climate change may affect all Pacific (independently of climate change), Pacific Islanders, particularly the poor and most Island governments would be justified in vulnerable. Climate change may also using reallocations of public expenditures exacerbate poverty by reducing coastal and development aid to fund these activities. settlement areas and affecting the crops and fisheries on which many communities § Develop a Broad Consultative Process for depend. Implementing Adaptation. Pacific Island governments should start a process of Failure to adapt now could not only lead to consultation with community representatives, major damages, but also result in a loss of the private sector, and other civil society opportunities to act in the future. Some coral institutions (such as churches and NGOs), on reef areas, for example, may no longer be a national strategy for adaptation. The able to recover in the future if degradation strategies should build upon the National continues at the present rates. Communications developed by the PICCAP By acting now to reduce their present-day country teams. The objective would be vulnerability to extreme weather events, mainstream adaptation into national policies Pacific Island countries could go a long way and development plans, to gain consensus on toward diminishing the effects of climate priority adaptation measures, and to build change in the future. partnerships for their implementation. 39 § Require Adaptation Screening for Major International Community Development Projects. To help defray future costs, Pacific Island governments should § Operationalize Adaptation Financing. Given require all major infrastructure projects to the importance of taking early action on undergo adaptation screening as part of an adaptation, the international community expanded environmental impact assessment. needs to urgently agree on the mechanism and size of adaptation financingbe it in the § Strengthen Socio-Economic Analysis of form of the Global Environmental Facility, a Adaptation Options. Further work on the tax on the Clean Development Mechanism as specific socio-economic impacts of climate currently discussed, or others. The findings change and adaptationsuch as done under from this study support the argument that this studycould help strengthen the Pacific Pacific Island countries will likely experience Island countries' position in international significant incremental costs from climate discussions on adaptation financing. A better change, and will need access to global understanding of the physical and economic adaptation funding. impacts would also help mainstream climate change into broader development planning. § Remove Incentives against Immediate Action on `No Regrets' Adaptation. Countries that have taken early action on adaptation using their own public expenditures or Donors development aid should not be penalized with a lower allocation of global adaptation § Support 'No Regrets' Adaptation. Donors funds, once these become available. have an important role to play in discussing Similarly, the justification for international with Pacific Island countries how to best financing of `no regrets' adaptation needs to orient development assistance in support of be recognized and promoted in its own right. national adaptation strategies. This could be Failure to do so could tilt the balance towards done either through stand alone interventions a `wait and see' attitude, in favor of more or as part of natural resources and expensive, but clearly incremental, structural environmental management programs. solutions (such as seawalls). § Support Adaptation Screening. To the extent Although many uncertainties remain, it now possible, donors should adopt adaptation seems clear that climate change will affect many screening as part of their policy requirements facets of Pacific Island people's lives and on environmental impact assessments. economies in ways that are just now beginning to be understood. Climate change therefore must be considered one of the most important challenges of the twenty-first century and a priority for immediate action. 40 References Background Studies to this Volume Aaheim, H. A., L. Sygna (2000). An Economic Assessment of Impacts of Climate Chang on Fisheries in the Pacific. Report Prepared for the World Bank by the Center for International Climate and Environmental Research , University of Oslo, Norway. Campbell, J. (2000). Climate Change Vulnerability and Adaptation Assessment for Fiji. Technical Summary and Synthesis. Report Prepared for the World Bank by the Center for International Global Change Institute, Waikato University, Hamilton, New Zealand. Campbell, J. (2000). Climate Change Vulnerability and Adaptation Assessment for Kiribati. Technical Summary and Synthesis. Report Prepared for the World Bank by International Global Change Institute, Waikato University, Hamilton, New Zealand. Falkland, T. (2000). Additional Groundwater Modelling of the Bonriki Freshwater Lens, Tarawa, Kiribati, using the SULTRA Groundwater Model. Prepared for the World Bank and International Climate Change Institute by Ecowise, Australia. International Global Change Institute, Waikato University, Hamilton, New Zealand, and the World Bank. Washington. D.C. Feresi, J., G. Kenny, N. De Wet, L. Limalevu, J. Bhusan and I. Ratukalou, editors (2000), with contributions from S. Hales, R. Maharaj, R. Ogoshi, and J. Terry. Climate Change Vulnerability and Adaptation Assessment for Fiji. Prepared and published by Fiji Pacific Island Climate Change Assistance Programme (PICCAP). Copyright: 2000 Government of Fiji and IGCI, University of Waikato, Hamilton, New Zealand. International Global Change Institute (IGCI) and South Pacific Regional Environment Programme (SPREP) (1999). PACCLIM WORKSHOP - Modelling Climate and Sea-level Change Effects in Pacific Island Countries, August 23-27, 1999. International Global Change Institute. Hamilton, New Zealand. International Global Change Institute (IGCI) in collaboration with the Pacific Islands Climate Change Assistance Programme (PICCAP) Fiji Country Team (2000). Climate Change Vulnerability and Adaptation Assessment for Fiji. Supplemental Fiji Coastal Impacts Study. Report Prepared for the World Bank. University of Waikato, New Zealand. King, Wayne (2000). Climate Change Overview and Background Related to Adaptation Options and Evaluation. Contribution for the World Bank Regional Economic Report for Pacific Island. South Pacific Regional Environmental Programme. Apia, Samoa. Lehodey, P. (2000). Impacts of Climate Change on Tuna Fisheries in the Tropical Pacific Ocean. Prepared by the Secretariat of the Pacific Community as edited by IGCI in partnership with South Pacific Regional Environment Programme (SPREP) and Pacific Islands Climate Change Assistance Programmed (PICCAP). Noumea, New Caledonia. Lehodey, P. and P. Williams (2000). Data on Tuna Catch in Central and Western Pacific. Electronic files, Secretariat of Pacific Community, Noumea, New Caledonia. 41 Stratus Consulting (2000). Economic Implications of Climate Change in Two Pacific Island Country Locations. Case Illustration of Tarawa, Kiribati and Viti Levu. Prepared under sub-contract to CICERO (Oslo, Norway). Boulder, Colorado. Taeuea, T., I. Ubaitoi, N. de Wet and G. Kenny, editors (2000), with contributions from N. Teuatabo, P. Kench, T. Falkland, and S. Hales. Climate Change, Vulnerability and Adaptation Assessment for Kiribati. Prepared and published by Kiribati Pacific Island Climate Change Assistance Programme (PICCAP). Copyright: 2000, Government of Kiribati and IGCI, University of Waikato, New Zealand. Van Aalst, Maarten. Contribution to the Climate Change Chapter of the Regional Economic Report. Washington, D.C. References9 ADB (1996). Findings and Recommendations, Sanitation and Public Health Project. Final Report. Prepared for the Asian Development Bank by Royds Consulting Party Ltd. in association with KPMG and Kab Beriki and Associates. December. Manila, Philippines. Alam, K. and A. C. Falkland (1997). Vulnerability to Climate Change of the Bonriki Freshwater Lens , Tarawa. Report No. HWR97/11, ECOWISE Environmental, ACTEW Corporation, prepared for the Ministry of Environment and Social Development, Republic of Kiribati, April 1997. Basu and Others (1999). Rising to the Challenge! A resource document for the Dengue Action Planning Forum at the Forum Secretariat. Suva, May 1999. Fiji Ministry of Health, Suva. Beagley, (1998). Bertignac, M. P. Lehodey, J. Hampton (1998). A Spatial Population Dynamics Simulation Model of Tropical Tunas Using a Habitat Index Based on Environmental Parameters. Fisheries Oceanography, 7(3/4): 326-335. Campbell (1995). Dealing with Disaster. Government of Fiji, Suva; East-West Center, Honolulu, Hawaii. Campbell, J. (1999). Vulnerability and Social Impacts of Extreme Events. Presented during the PACCLIM Workshop, August 23-27, 1999. In International Global Change Institute (IGCI) and South Pacific Regional Environment Programme (SPREP) (1999). PACCLIM WORKSHOP - Modelling Climate and Sea-level Change Effects in Pacific Island Countries, August 23-27, 1999. International Global Change Institute. Hamilton, New Zealand. Cesar, H. (1996). Economic Analysis of Indonesian Coral Reefs. Environmental Department. The World Bank, Washington, D.C. Christensen, L. (1982). Management and Utilization of Mangroves in Asia and the Pacific. FAO Government Paper # 3. Food and Agriculture Organization of the United Nations. Rome, Italy. 9Many of the references were consulted by the contributing authors to this volume, and are included in the bibliography of their background studies. Only references that were directly quoted in this volume are included here. 42 Clark, K. M. (1997). Current and Potential Impact of Hurricane Variability on the Insurance Industry. In H. F. Diaz and R.S. Pulwarty (editors). "Hurricanes, Climate and Socioeconomics." Springer. Constanza, R., R. d'Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R.V. O'Neill, J. Paruelo, R. G. Raskin, P. Sutton, and M. van den Belt (1997). The Value of the World's Ecosystem Services and Natural Capital. Nature, 387:253-260. Cubash, U., K. Hasselmann, H. Hock, E. Maier-Reimer, U. Mikolajewicz, B. Santer and R. Sausen (1992). Time Dependent Greenhouse Warming Computations with a Coupled Ocean- Atmosphere Model. Climate Dynamics, 8:55-69. De Groot, R. S. (1992). Functions of Nature: Evaluation of Nature in Environmental Planning, Management and Decision Making. Wolters-Noordhoff, Groningen. 315 pp. De Wet, Neil (1999). A Conceptual Framework for Adaptation to Climate and Sea-Level Change in Pacific Island Countries. In International Global Change Institute (IGCI) and South Pacific Regional Environment Programme (SPREP) (1999). PACCLIM WORKSHOP - Modelling Climate and Sea-level Change Effects in Pacific Island Countries, August 23-27, 1999. International Global Change Institute. Hamilton, New Zealand. Earth Negotiations Bulletin (ENB) (1999). Summary of the Fifth Conference of the Parties to the Framework Convention on Climate Change. Earth Negotiations Bulletin. Published by the International Institute for Sustainable Development (IISD). Http:// www.iisda.ca accessed November (1999). Eastman, J. R. (1995). IDRISI for Windows User's Guide Version 1.0. Clark Labs for Cartographic Technology and Geographical Analysis, Clark University, Worcester. Environmental Protection Agency (EPA). 1999. Health Risk Reduction and Cost Analysis for Radon in Drinking Water. United States Environmental Protection Agency, EPA-815-2-99-002. Washington, D.C. Food and Agriculture Organization of the United Nations (FAO). 1996. FAOSTAT Database. http://apps.fao.org/cgi-bin/nph-db.pl?subset = agriculture. Accessed January 19, 2000. Feresi, J., G. Keeny, N. De Wet, L. Limalevu, J. Bhusan and I. Ratukalou, editors (2000), with contributions from S. Hales, R. Maharaj, R. Ogoshi, and J. Terry. Climate Change Vulnerability and Adaptation Assessment for Fiji. Prepared and published by Fiji Pacific Island Climate Change Assistance Programme (PICCAP). Copyright: 2000 Government of Fiji and IGCI, University of Waikato, Hamilton, New Zealand. Fiji Fisheries Division (FFD) 1999. Annual Report 1998. Prepared by the Ministry of Agriculture, Fisheries and Forests. Suva, Fiji. Fiji Real Estate (2000). http://www.fijirealestate.com/waidroka/property-map.html. Consulted May 2000. Food and Agriculture Organizations of the United Nations (FAO) (1999). Environment and Natural Resources in Small Island Developing States. Rome, 12 March, 1999. 43 Gittinger, J.P. (1982). Economic Analysis of Agricultural Projects. EDI Series in Economic Development. Economic Development Institute, the World Bank. Washington, D.C. Gordon, H.B. and S.P.O. O'Farrell (1997). Transient Climate Change in the CSIRO Coupled Model with Dynamic Sea Ice. Monthly Weather Review, 125(5): 875-907. Graham, T., N. Idechong, and K. Sherwood (2000). The Value of Diving and the Impacts of Coral Bleaching in Palau. Presented at the 9th International Coral Reef Symposium, October 23-27, Bali, Indonesia. Hackett, C. (1988). Matching Plants and Land: Development of a General Broadscale System from a Crop Project for Papua New Guinea. Natural Resources Series No. 11, CSIRO Division of Water and Land Resources. Commonwealth Scientific and Industrial Research Organization, Australia. Hackett, C. (1991). PLANTGRO: A Software Package for Coarse Prediction of Plant Growth. Australia Commonwealth Scientific and Industrial Research Organization. Clayton, Australia. Hales, S., P. Weinstein, Y. Souares, A. Wood (1999). El Nino and the Dynamic of Vector Borne Disease Transmission. Environmental Helth Perspectives 107: 99-102. Harmelin-Vvien, M.L. (1994). The Effects of Storms and Cyclones on Coral Reefs: A Review. Journal of Coastal Research. Special Issues 12, 211-232. Holland, G. J. (1997). The Maximum Potential Intensity of Tropical Cyclones. Journal of Atmospheric Science, 54 : 2519-2541. Hopley, D. and D.W. Kinsey (1988). The Effects of a Rapid Short-term Sea-Level Rise on the Great Barrier Reef in Taueva, Tianuare, Ioane Ubaitoi. In I.G. Pearman (ed.). "Greenhouse: Planning for Climate Change". Commonwealth Scientific and Industrial Research Organization (CSIRO), Melbourne, Australia, p. 189-201. International Global Change Institute (IGCI) and South Pacific Regional Environment Programme (SPREP) (1999). PACCLIM WORKSHOP - Modelling Climate and Sea-level Change Effects in Pacific Island Countries, August 23-27, 1999. International Global Change Institute. Hamilton, New Zealand. Intergovernmental Panel on Climate Change (IPCC)(1996). Climate Change 1995: The IPPC Second Assessment Report. Watson, Rt., M.C. Zinyowera and R.H. Moss (eds). Cambridge University Press, Cambridge and New York. Intergrovernmental Panel on Climate Change (IPPC) (1998). Summary Report: IPCC Workshop on Adaptation to Climate Variability and Change. March 29-April 1, 1998. San Jose, Costa Rica. JICA (1998). The Study on Watershed Management and Flood Control for the Four Major Viti Levu River in the Republic of Fiji. Draft Final Report, Main Report prepared by the Yachiyo Engineering Co., Ltd. for the Japan International Cooperation Agency, Ministry of Agriculture, Fisheries and Forests, Fiji. July. Jones, R.N., P.H. Whetton, K.J.E. Walsh, R. Suppiah and K.J. Hennessy (1999). Scenarios of Climate Variability for the South Pacific. In International Global Change Institute (IGCI) and South 44 Pacific Regional Environment Programme (SPREP) (1999). PACCLIM WORKSHOP - Modelling Climate and Sea-level Change Effects in Pacific Island Countries, August 23-27, 1999. International Global Change Institute. Hamilton, New Zealand. Kench, P. and P. Cowell (1999). Impacts of Sea Level Rise and Climate Change on Pacific Coasts. In International Global Change Institute (IGCI) and South Pacific Regional Environment Programme (SPREP) (1999). PACCLIM WORKSHOP - Modelling Climate and Sea-level Change Effects in Pacific Island Countries, August 23-27, 1999. International Global Change Institute. Hamilton, New Zealand. Koroivueta, personal communications (1999). As quoted In Feresi, J., G. Keeny, N. De Wet, L. Limalevu, J. Bhusan and I. Ratukalou, editors (2000), with contributions from S. Hales, R. Maharaj, R. Ogoshi, and J. Terry. Climate Change Vulnerability and Adaptation Assessment for Fiji. Prepared and published by Fiji Pacific Island Climate Change Assistance Programme (PICCAP). Copyright: 2000 Government of Fiji and IGCI, University of Waikato, Hamilton, New Zealand. Legislative Assembly of Samoa (1999). Approved Estimates of Receipts and Payment of the Government for the Financial Year Ending 30th June 2000. Parliamentary Paper 1999, No. 9. Apia, Samoa. Lehodey, P. (Submitted). The Pelagic Ecosystem of the Tropical Pacific Ocean: Dynamic Spatial Modelling and Biological Consequences of ENSO. Submitted to Progress in Oceanography, Special Issue of the "Beyond El Niņo" Conference, March 23-26, 2000. La Jolla, California, USA. Lehodey, P. (2000). Impacts of Climate Change on Tuna Fisheries in the Tropical Pacific Ocean. Prepared by the Secretariat of the Pacific Community as edited by IGCI in partnership with South Pacific Regional Environment Programme (SPREP) and Pacific Islands Climate Change Assistance Programmed (PICCAP). Noumea, New Caledonia. Lehodey, P., J. M. Andre, M. Bertignac, J. Hampton, A. Stoens, C. Menkes, L. Memery, N. Grima (1998). Predicting Skipjack Tuna Forage Distributions in the Equatorial Pacific Using a Coupled Dynamical Bio-Geochemical Model. Fisheries Oceanography, 7(3/4): 317-325. Lehodey, P., M. Bertignac, J. Hampton, A. Lewis and J. Picaut (1997). El Nino Southern Oscillation and Tuna Western Pacific. Nature. 389: 715-718. Lewis, R. (1992). Ciguatera in the Pacific. Bulletin Societe Pathologie Exotique 85: 427-434. Lewis, R. J. and T. A. Ruff (1993). Ciguatera: Ecological, Clinical, and Socio-Economic Perspectives. Critical Reviews in Environmental Science and Technology 23: 137-156. McLean, R. F. (1989). Kiribati and Sea Level Rise. Report to the Commonwealth Secretariat. Marshall, J.F. and G. Jacobsen (1985). Holocene Growth of a mid-Pacific Atoll: Tarawa, Kiribati: Coral Reefs, 4:11-17. Metz, (1996). Mimura, N., and Nunn, P.D. (1998). Journal of Coastal Research, 14, 37-46. Ministry of Finance of Fiji (MNP). 1999. Economic and Fiscal Update. Supplement to the 2000 Budget Address. Ministry of Finance and National Planning. November 5, 1999. Suva, Fiji. 45 Pahalal, Janita and Jai Shree Gawander (1999). The Impact of Climate Change on Sugar Cane in Fiji. Fiji Meteorological Service and Sugarcane Research Center and Fiji Sugar Corporation, Suva and Lautoka, Fiji. Patz, J.A., W.J. M. Martens, D.A. Focks, and T.H. Jetten (1998). Dengue Fever Epidemic Potential as Projected by General Circulation Models of Climate Change. Environmental Health Perspectives, 106:3:147-153. PNG Consultants (1996). Ruff, T. A. (1989). Ciguatera in the Pacific: A Link with Military Activities. Lancet 1L 201-205. Ruff, T.A. and R.J. Lewis (1997). Clinical Aspects of Ciguatera: An Overview. Pacific Health Dialog Vol. 4. No.2, pp. 119-127. Shalev, Z. (1992). Draft 10 Year National Water Master Plan. United Nations Department of Technical Cooperation for Development. Project KIR/87/006. Sistro, N. (1997). The Economic Value of Fiji's Ecosystems. Technical Group 5 Report for Fiji Biodiversity Strategy and Action Plan. Government of Fiji, Suva. Solomon, S. (1997). Assessment of the Vulnerability of Betio (South Tarawa, Kiribati) to Accelerated Sea-level Rise. SOPAC Technical Report 251. South Pacific Geoscience Commission, Suva, Fiji. Solomon, S. and J. Kruger (1996). Vulnerability and Adaptation Assessment: Coastal Impact of Sea Level Change, Viti Levu, Fiji. SOPAC Technical Report No. 242. South Pacific Geoscience Commission, Suva, Fiji South Pacific Applied Geoscience Commission (SOPAC), 1998. Demand Management and Conservation Project, Field Investigation in South Tarawa, Kiribati. Prepared by Ed Burke, Project Manager, Water Resources Unit, SOPAC. Miscellaneous Report 302. Suva, Fiji. South Pacific Regional Environment (SPREP) 2000. Draft Pacific Island's Framework for Action on Climate Change, Climate Variability and Sea Level Rise. Apia, Samoa. SPECTRUM (1999). Spectrum Population Software. Human Resource Systems Corporation. http://www.spectrumhr.com. South Pacific Regional Environmental Programme (SPREP), 2000. Draft Pacific Island's Framework for Action on Climate Change, Climate Variability and Sea Level Rise. Apia, Samoa. Stratus Consulting (2000). Economic Implications of Climate Change in Two Pacific Island Country Locations. Case Illustration of Tarawa, Kiribati and Viti Levu. Prepared under sub-contract to CICERO (Oslo, Norway). Boulder, Colorado. Taeuea, T., I. Ubaitoi, N. de Wet and G. Kenny, editors (2000), with contributions from N. Teuatabo, P. Kench, T. Falkland, and S. Hales. Climate Change, Vulnerability and Adaptation Assessment for Kiribati. Prepared and published by Kiribati Pacific Island Climate Change Assistance Programme (PICCAP). Copyright: 2000, Government of Kiribati and IGCI, University of Waikato, New Zealand. 46 Timmermann, A., Oberhuber, J., Bacher, A., esch, M., Latif, M., Roeckner, E. (1999). Increased El Niņo Frequency in a Climate Model Forced by Future Greenhouse Warming. Nature 398, 694-697. United Nations Development Programme (UNDP) and Government of Fiji ( 1997). Fiji Poverty Report. United Nations Development Program. Quality Press, Suva, Fiji. United Nations Development Programme (UNDP) (1996). Establishing a Coastal Management Program for Majuro Atoll. Proposal of the Government of the Marshall Islands. United Nations Development Programme, Suva, Fiji. United Nations Disaster Assessment and Coordination (UNDAC) 1998. UNDAC Mission Reports on Fiji Drought. http://www.reliefweb.int/w/rwb.nsf/S/A3238C8D6E14D385C12566C9004C0D9C. Accessed September 1999. In Stratus Consulting (2000). Economic Implications of Climate Change in Two Pacific Island Country Locations. Case Illustration of Tarawa, Kiribati and Viti Levu. Prepared under sub-contract to CICERO (Oslo, Norway). Boulder, Colorado. United Nations Educational, Scientific and Cultural Organization (UNESCO). 1991. Hydrology and Water Resources of Small Islands. A Practical Guide. Studies and Reports on Hydrology No. 49. Prepared by A. Falkland (ed.) and E. Custodio with contributions from A. Diaz Arenas and L. Simler and case studies submitted by others. Paris, France, 435 pp. Voss, C. I. (1984). SUTRA: A Finite Element Simulation Model for Saturated, Unsaturated, Fluid- Density Dependent, Groundwater Flow with Energy Transport or Chemically-Reactive Single- Species Solute Transport. US Geological Survey, Water Resources Investigation Report, 84- 4389, 409 pp. Watling, D. (1985). A Mangrove Management Place for Fiji (Phase 1). South Pacific Commission and Fiji Government. Suva, Fiji. Westmacott, S., H. Cesar and L. Pet-Soede (2000). Socio-Economic Assessment of the Impacts of the 1998 Coral Reef Bleaching in the Indian Ocean. In Westmacott, S. , H. Cesar, L. Pet-Soede, J. De Schutter. "Assessing the Socio-Economic Impacts of Coral Reef Bleaching in the Indian Ocean. CORDIO Report Submitted to the World Bank African Environment Department. The World Bank, Washington, D.C. White, I, A. Falkland, B. Etuati, E. Merdi, and T. Metuatera. (1999). Recharge of Fresh Groundwater Lenses: Field Study, Tarawa Atoll, Kiribati. Second International Colloquium on Hydrology and Water Management i n the Humid Tropic, Panama City, 21-26 Larch. Wigley, T.M.L. (1994). MAGGIC: Model for the Assessment of Greenhouse-gas Induced Climate Change. Users Guide and Scientific Reference Manual, University Corporation for Atmospheric Research, Boulder, Colorado. Wilkinson, Clive, Olof Linden, Herman Cesar, Gregor Hodgson, Jason Rubens and Alan E. Strong (1999). Ecological and Socioeconomic Impacts of 1998 Coral Mortality in the Indian Ocean: An ENSO Impact and a Warning of Future Change? Ambio Vol 28 ,No. 2, March: 188-196. World Bank (2000a). Cities, Seas, and Storms: Managing Change in Pacific Island Economies. Volume I: Summary Report (draft). Papua New Guinea and Pacific Islands Country Unit. The World Bank. Washington, D.C. November 13, 2000. 47 World Bank (2000b). Cities, Seas, and Storms: Managing Change in Pacific Island Economies. Volume II: Managing Pacific Towns (draft). Papua New Guinea and Pacific Islands Country Unit. The World Bank. Washington, D.C. November 30, 2000. World Bank (2000c). Cities, Seas, and Storms: Managing Change in Pacific Island Economies. Volume III: Managing the Use of the Ocean (draft). Papua New Guinea and Pacific Islands Country Unit. The World Bank. Washington, D.C. November 30, 2000. World Health Organization (WHO), 1996. Climate change and Human Health. [McMichael, A.J., Haines, A., Sloof, R., and Kovats, S. (eds.)]. World Health Organization, Geneva. World Health Organization (WHO). 1998. Weekly Epidemiological Record, No. 36, 4. September 1998. Published by the World Health Organization, Geneva. World Resources Institute (1999). Status of Coral Reefs Classified by Potential Threat from Human Activities. Status of the World's Coral Reefs: Pacific Ocean Page. World Resources Institute. http://www.wri.org/wri/indictrs/reefocea.htm. Accessed January 2000. 48 Annex A Page 1 Annex A Assumptions Used in the Assessment of Climate Change Impacts The analysis of climate impacts was conducted by a multi-disciplinary team from more than 20 different institutions (see Acknowledgments). To allow for a more in-depth assessment of climate change impacts on key economic sectors, the study team focused on Viti Levu (Fiji) as an example of a high island, and on the Tarawa atoll (Kiribati) as an example of low islands in the Pacific. The two study sites were also selected based on the availability of data. The assessment of climate change impacts relied on an integrated assessment model, the Pacific Climate Change Impacts Model (PACCLIM), which was developed for the Pacific Island region by the International Global Change Institute (IGCI). PACCLIM was originally developed as a regional scenario generator under the Pacific Islands Climate Change Assessment Program (PICCAP). For the purposes of this study, PACCLIM was enhanced to provide projections of the effects of climate change and sea-level rise on four major sectors (figure A.1): ˇ Coastal areas ˇ Water resources ˇ Agriculture ˇ Health Where possible, historical data, qualitative observations, expert judgments and existing literature were also used to verify or reject projected impacts, or to provide further information on possible effects. Figure A.1 . The PACCLIM Model System and Main Components Global Climate Model (MAGICC)1 Greenhouse Gas Emissions to Concentration Models Global Mean Temperature and Sea- Emissions Scenarios Level Change Model and User-defined Model Parameters GCM Patterns Regional Climate Change Interpolated Scenario Generator Climate Data Sectoral Impact Models Coastal Resources Water Resources Agriculture Health Environmental Effects 1- Model for the Assessment of Greenhouse Gas Induced Climate Change Annex A Page 2 The impact of climate change on regional tuna fisheries was assessed separately using present knowledge on tuna biology and fisheries, as well as the recent findings on the environmental impacts of ENSO on tuna stocks in the western central Pacific Ocean (Lehodey and others, 1997; Lehodey 2000). The analysis was also based on simulation results from the Spatial Environmental Population Dynamics Model (SEPODYM) developed at the Secretariat of Pacific Community Oceanic Fishery Programme (Lehodey and others 1998; Bertignac and others 1998; Lehodey, submitted). This model takes into consideration the movement of tuna and the effects of environmental variability. Sea surface temperature, oceanic currents, and primary production are used in the model to delineate tuna spawning areas, the transport of larvae and juveniles, and simulations of tuna forage distribution. A. General Scenarios Rainfall and Temperature Changes The scenarios developed for Viti Levu and Tarawa were based on general circulation models (GCMs). While GCMs do not have the resolution to yield accurate results at the scale of the Pacific Island region or individual countries, there tends to be an agreement among the various models on changes in temperature. However, the models show inconsistencies in projections of rainfall, and capture poorly the effects of ENSO phenomena. These shortcomings were handled as described below. Temperature and rainfall scenarios for Viti Levu and and Tarawa were projected for the years 2025, 2050 and 2100 using the PACCLIM Scenario Generator. Two greenhouse gas (GHG) emission scenarios were used: the Special Report on Emission Scenarios (SRES) B2 mid (climate sensitivity of 2.5o C), and the SRES A2 high (climate sensitivity of 4.5o C) scenario. These correspond to the "best-guess" and "worst- case" scenarios described in this report. The study adopted the results from two different GCMs. The first of these is known as the 9 Layer Global Circulation Model of Australia's Commonwealth Scientific and Industrial Research Organization, or CSIROM2 (Gordon and O'Farrell 1997) which has been scrutinized and validated for the South Pacific region. The second GCM chosen was the Deutsche Klimarechenzentrum (German Climate Monitoring Center) DKRZ, developed by Cubasch and others (1992). PACCLIM also included two other models, the Canadian Climate Centre and the Hadley Center GCMs. However, these models agree with CSIROM2 in projecting rainfall increases. The DKRZ model, by contrast, projects a decrease in rainfall. Given the importance of droughts for Pacific Island countries, the DKRZ model was chosen along with the CSIROM2 to represent the range of possible impacts resulting from rainfall changes. The resulting scenarios of temperature and rainfall change for Fiji and Kiribati are shown on tables A.1 and A.2. Table A.1. Summary of Temperature and Rainfall Change Scenarios for Fiji General 2025 2050 2100 Circulation Emissions Temp Rainfall Temp Rainfall Temp Rainfall Model Scenario (°C) (%) (°C) (%) (°C) (%) B2 (mid) 0.5 3.3 0.9 5.7 1.6 9.7 CSIROM2 A2 (high) 0.6 3.7 1.3 8.2 3.3 20.3 DKRZ B2 (mid) 0.5 -3.3 0.9 -5.7 1.6 -9.7 A2 (high) 0.6 -3.7 1.3 -8.2 3.3 -20.3 Annex A Page 3 Table A.2. Summary of Temperature and Rainfall Change Scenarios for Kiribati General 2025 2050 2100 Circulation Emissions Temp Rainfall Temp Rainfall Temp Rainfall Model Scenario (°C) (%) (°C) (%) (°C) (%) CSIROM2 B2 (mid) 0.5 2.8 0.9 5.0 1.6 8.4 A2 (high) 0.6 3.2 1.3 7.1 3.4 17.7 DKRZ B2 (mid) 0.5 -4.3 0.9 -7.5 1.6 -12.8 A2 (high) 0.6 -4.8 1.3 -10.7 3.3 -26.9 Sea Level Rise Confidence in GCM projections of sea level rise at the regional level remains low. There is also limited long-term historical data at the country level. Accordingly, global mean projections of sea level rise were used as first order estimates for the analysis (table A.3). This was carried out by using a global climate model, the Model for the Assessment of Greenhouse Gas Induced Climate Change, or MAGICC (Wigley 1994) in the linked model system of PACCLIM. Table A.3. Summary of Global Sea Level Rise Projections Scenario 2025 2050 2100 B2 (mid-range, best guess) 11 cm 23 cm 50 cm A2 (high-range, worst case) 21 cm 43 cm 103 cm Climate Variability: Changes in Cyclones and ENSO Because the GCMs do not yet account for ENSO variability or changes in the frequency or magnitude of extreme climate events, analogues based on recent patterns of occurrence were used to analyze the potential impact of future events in Fiji. Kiribati lies outside the cyclone path, and there was insufficient information to quantify the impact of ENSO-related droughts or floods beyond qualitative statements. Cyclones. A recent review of climate variability in the South Pacific area (Jones and others 1999) projects an increase in cyclone intensity of 0 to 20 percent by mid-century. This increase in cyclone intensity was applied to baseline conditions derived from the sequence of actual cyclone events in Fiji from 1992 to 1999. No changes were assumed in cyclone frequency. ENSO. Existing regional research (Jones and others 1999) predicts that average conditions in the future will increasingly resemble a present-day El Niņo. For Fiji, the frequency of El Niņo-induced droughts in 1983-981 one drought every four yearsand an intensity comparable to that of the 1997/98 drought were used to represent future climate conditions. Socio-Economic Scenarios Given the high level of uncertainty associated with future socio-economic conditions, the economic costs of climate change were estimated based on the likely economic impacts of 2050 scenarios as applied to today's (1998) conditions. While this approach is likely to underestimate the magnitude of future 1The use of the 1983-98 period was largely based on the availability of sugarcane data for the period. Annex A Page 4 impacts, it provides policy makers with an estimate that is closer to the present-day reality that surrounds them. An exception to this principle was made for health impacts, which are closely related to population size. The economic impacts on health were computed by taking into account population projections for 2050 (tables A.4 and A.5). Table A.4. Population Projections for Fiji Growth Rate Projection 2026 2051 2096 Low 1,110,000 1,260,000 1,280,000 Medium 1,180,000 1,480,000 1,720,000 High 1,210,000 1,620,000 2,300,000 Note: Assumes an on-going reduction in total fertility rate, gradual increases in life expectancy and decreasing infant mortality rates. Migration patterns have not been included in the projections. Projections are based on the 1996 census and were developed using SPECTRUM Policy Modeling System (Version 1.33) demographic software (Futures Group International, USA). Numbers are rounded to the nearest 1000. Table A.5. Population Projections for Kiribati Growth Rate Projection 2025 2050 2100 Low 128,000 154,000 165,000 Mid-range 139,000 187,000 246,000 High 146,000 215,000 351,000 Note: Projections are based on the 1995 census data and have been developed using SPECTRUM demographic software. All costs were reported in 1998 US dollar values. Data for different years were converted into 1998 values by appropriate price indeces (such as producer or consumer price indeces). The costs represent annual average losses due to climate change and/or future climate variability. When ranges are given, they reflect a best-guess (the lowest value) and a worst-case scenario (the highest value). The economic costs assume no adaptation. In practice, communities are likely to undertake adaptation measures to protect against climate change impacts. Hence, the costs presented in this analysis should be interpreted as what could happen under a policy of inaction. The economic estimates are also partial and do not take into account secondary interactions. For example, climate change is not only likely to affect water resources, but could also decrease economic production. This could in turn lead to a lower water demand than what would be expected in the absence of climate change. The nature of these interactions, however, is too complex and uncertain to be included in the analysis. The use of annual average costs mask the actual impact of climate events in a given year. For example, if droughts of severity A happened every four years, the annual average cost would be one-fourth of the costs of A. But for a disaster year, the impact would be A , or four time as high as the annual average. In order to illustrate this, the summary tables for Viti Levu and Tarawa include both the average annual damages as well as the likely costs of an extreme event. For periodic events affecting major infrastructuresuch as the impacts of sea level rise on coastal areasthe annualized value of the losses was estimated using the capital recovery factor of infrastructure and land. This factor is estimated as follows (Gittinger 1982): [i(1+i)n] [(1+i)n -1] where i = rate of interest (assumed to be 10 percent) and n = number of years when losses can accrue. Annex A Page 5 Uncertainties Box A.1. Summary of Uncertainties The development of climate change scenarios is Relative Level fraught with uncertainties. The uncertainties of Certainty become even greater as one seeks to identify impacts resulting from the projected climate and Base scenarios sea-level changes. For small island states the uncertainties are further magnified as the areas of ˇ Temperature the countries usually fall below the levels of ˇ Rainfall resolution of global circulation models. ˇ Extreme events As one moves from the general to the more drought specific, the level of certainty decreases. A major ˇ problem with climate impact assessment is that ˇ tropical cyclones many impacts are site specific, depending on the shape of a beach, the conditions of a watershed, ˇ ENSO events the level of existing environmental degradation, the crops a farmer plants or the socio-economic ˇ Sea level conditions of a given locality. Accordingly, adaptive action, when implemented, will need to ˇ Socio-economic be carried out with the specific impacts in mind. For this reason, choosing specific adaptive Impact scenarios actions is problematic until greater certainty in impact assessment is achieved. Nevertheless, one ˇ Coastal can point to some generic sets of impacts and a ˇ Water Resources range of broad adaptive approaches to them. This was the approach taken in this report. ˇ Agriculture ˇ Health The levels of predictive certainty also recede as attention is focused further into the future. It is ˇ Adaptive Responses perhaps no exaggeration to state that developing ˇ Generic scenarios of social and economic conditions one century into the future is even more difficult than ˇ Sector specific projecting climate and sea-level conditions. ˇ Site specific It is not possible to place clear statistical Key: Level of Certainty boundaries to the uncertainties inherent in the High Medium Low vulnerability and adaptation studies. The accompanying box, however, summarizes the uncertainties in the study. Uncertainties associated with the economic analysis are Reduced Certainty As Predictive included in the summary tables of Viti Levu Horizon is Expanded (table 3, Chapter 3) and Tarawa (table 12, Chapter 4). The assumptions used are clearly delineated in this Annex to enable replication of Level the results, application to other studies, and Of corrections in the original analysis should Certainty improved assumptions become available in the future. Time Annex A Page 6 B. Impacts on Coastal Areas General The analysis of impacts on the Viti Levu and Tarawa coasts involved the following: Selection of case study sites, representative of the coastal types found in the islands. Transect surveys on each study site. Assessment of Island Shoreline Displacement (Erosion analysis) using the Shoreface Translation Model. Assessment of Island Inundation, using topographic maps. Assessment of Land and Infrastructure impacted by inundation. Extrapolation of case study sites to the remainder of the islands. Economic analysis of land and infrastructure losses, comparing 2050 conditions with the baseline. Selection of Case Study Sites In both Viti Levu and Tarawa, case study sites were selected to represent the broad physical and socio- economic conditions found in the islands. The case study sitestwo in Tarawa, and four in Viti Levuwere selected based on morphology, biological characteristics, population density, and land-use. The selection involved the PICCAP country teams of Fiji and Kiribati, as well as specialists from IGCI. Transect Surveys In each site, 2-3 representative transects were surveyed for the erosion analysis, from the reef edge and across the islands (in Tarawa) and across the coastal margin (in Viti Levu). Assessment of Island Shoreline Displacement (Erosion Analysis) The analysis of island shoreline displacement (erosion analysis) was based on the Shoreface Translation Model (STM), adapted to the conditions of Tarawa and Viti Levu under rising sea levels (Cowell et al. 1995, Kench and Cowell 1999). This model incorporates elements of both the Standard and Generalized Brunn Rules,2 as well as hybrids of the two, but goes further in allowing for time-varying morphological dimensions (such as shoreface) and sediment gains and losses (such as those due to littoral sand transport). Simulations of shoreline recession and changes to the coastal morphology were undertaken for 0.1 meter increments of sea level rise (figure A.2). The simulations assumed the following: ˇ A balanced sediment budget (no additional gains or losses) under present conditions and conditions of sea level rise. This assumption is unlikely to hold for the Rewa river delta site in Viti Levu or for significant rises in sea level (0.5 meters or more). However, no data were available on sediment discharge rates or possible changes in sediment transport with sea level rise. 2The Standard Brunn Rule is the most common method for assessing shoreline change (Solomon 1997). However, this rule applies to steep substrates and is not appropriate to low-lying atolls. The Generalized Brunn Rule, by contrast, allows for sub- aerial beach dune and lagoon to move upward as an equilibrium response to sea level rise, and is more appropriate to atoll conditions. Annex A Page 7 Figure A.2. Simulation of Shoreline Retreat After Varying Rates of Sea Level Rise in a transect of Buariki, North Tarawa, Kiribati 2 m 20 m A. 0.1m SLR Sea-Level Rise Shoreline Recession (m) 0.1 0.6 0.2 1.2 0.3 2.6 0.4 3.2 2 m 0.5 4.5 20 m 0.6 5.2 B. 0.2m SLR 0.7 6.5 0.8 7.2 0.9 8.4 1.0 9.5 2 m 20 m C. 0.4m SLR 2 m 20 m D. 0.5m SLR 2 m 20 m E. 1.0m SLR Annex A Page 8 ˇ For Tarawa, hard surfaces (either reef flat or conglomerate platform) were assumed to extend horizontally under sand island and truncate the shoreline profile. ˇ For Tarawa, the uppermost hard substrate (conglomerate platform or reef surface) was assumed to extend horizontally beneath the sand island from the point at which it is buried by the beach- sediment lens. ˇ For Viti Levu, based on anecdotal evidence of storm inundation, washover of the coastal margin was allowed in the model (up to 50 meters), allowing for deposition of sediment on the island surface. Assessment of Island Inundation (Inundation Analysis) The inundation analysis was done using a simple drowning concept. For Viti Levu, topographic maps had too coarse a scale (20 meter intervals) to be used. Therefore, field surveys of elevation and aerial photographs were used to construct maps with 1 meter contours along sections of the coast at each case study site. For Tarawa, the survey team used topographic maps supplied by the Land Management Division, which contained 1 meter contours with all surveyed structures and road systems clearly marked. To determine baseline conditions, the Mean High Water Spring tide level (MHWS) was determined using data from the Suva tide gauge in Viti Levu (MHWS=0.64 meters), and from the National Tidal Facility in Tarawa (MHWS=2.53 meters). A second baseline was constructed by adding the impact of periodic storm surges to the MHWS, based on historical records of water levels and the frequency of these storms. For Viti Levu, a 1 in 50 year storm surge level causing a surge of 0.98 meters (Solomon and Kruger 1996) was added to the MHWS, producing a baseline with storm surge of 1.62 meters. For Tarawa, a 1 in 14 year storm surge causing a surge of 0.88 meters (Solomon 1997) was used, producing a baseline with storm surge of 3.41 meters. The inundation analysis was performed by raising the MHWS level by different sea level rise increments, corresponding to the scenarios of sea level rise (table A.9), and sea level rise with storm surge. The analysis assumes no change in the level of MHWS. It was also assumed that the frequency of storms would not increase under sea level rise. This is a very conservative assumption, as an increase in cyclone intensity could lead to more storms surpassing the baseline storm surge conditions. However, the effects of more intense cyclones are already accounted for in the water resources analysis of Viti Levu. The estimate of climate change impact also does not take into account other potential major events, such as a 1 in 100 year storms. Finally, the analysis assumes that the islands' surface remains static, such as what would happen in Tarawa if sediment redistribution on top of the islands was prevented. Assessment of Land and Infrastructure Impacted by Erosion and Inundation In Viti Levu, the land types and infrastructure likely to be impacted by erosion and inundation at the survey sites were estimated from aerial photography and field surveys. In Tarawa, the land and infrastructure likely to be inundated was calculated from the maps provided by the Land Management Division. Due to the quality of the maps the type of structure could not be determined. Extrapolation of Case Study Sites to Viti Levu and Tarawa Atoll In order to determine the impact on the whole islands, the results of the case study sites had to be extrapolated. In Viti Levu, the extrapolation was based on the length of coast sampled relative to the total length of coast for that land type. Given the morphology of Viti Levu a high island an extrapolation by area (rather than length of shoreline) would have require a judgment of which altitude is considered "coastal". For Tarawa, an extrapolation by area was possible due to the low altitude of the atoll (where most of the land is coastal). Annex A Page 9 Economic Analysis of Land and Infrastructure Losses The analysis of land and infrastructure structures lost to erosion and inundation was based on the extrapolated values (above), and the difference between sea level rise conditions and the baseline. Thus, a sea level rise of 0.4 meters (the worst case scenario for 2050) was compared to current conditions of MHWS. A sea level rise of 0.4 meters with storm surge was compared to current conditions with storm surge. These two results were then weighted by the frequency of storm surge. For Kiribati, for example, the impact of sea level rise without storm surge was assumed to occur in 13 out of 14 years. The impact of sea level rise with storm surge was assumed to occur once every 14 years (reflecting the present frequency of storms). The economic value of land and infrastructure should reflect the value of their potential future use. Under perfect market conditions, the economic value of land should resemble its market price. This may not hold in Kiribati, however, as most land is not freely traded. For lack of economic estimates, the market value of structures and land as provided by Government agencies was used in the Tarawa analysis. For Viti Levu, where more economic data were available, the cost of land lost reflects its economic value. In the Tarawa analysis, it was assumed that all land and infrastructure affected by inundation would be lost. This may be overly pessimistic. However, it should be remembered that the analysis assumes that in 13 out of 14 years, inundation is caused solely by sea level rise (and not by storm surge). This progressive rise in sea level is a form of permanent inundation which is likely to destroy the existing structures in the absence of adaptation. In Viti Levu, only land was considered to be lost in its entirety, while structures were considered to be only partially lost due to construction standards that already account for periodic inundation. The economic value of land and infrastructure reflects the capital value of a stock. In order to compute the annualized losses, the capital value needs to be adjusted by a capital recovery factor that reflects losses to depreciation and the way the population values present-day assets. An interest rate of 10 percent was used for both Viti Levu and Tarawa. Viti Levu structures were assumed to last 25 years, while Tarawa structures were assumed to last 10 years based on current construction standards. Viti Levu Impact of Erosion Caused by Climate Change on Viti Levu's Coast A. Physical Impact of Coastal Erosion due to Sea Level Rise The coast of Viti Levu is approximately 750 km long. Using a 1:50,000 topographic map, the coast was divided into four broad categories of land types representing about 90 percent of the total coast (table A.6). The remaining area (not sampled) consists primarily of sand spits and small deltas. Table A.6. Division of Viti Levu Coast into Land Types Length of Coast % of Total Coast Case Study Length of Coast of Case Coastal Type (in km) Represented Site Study Site (in km) Southern coast: 170 28 Korotogo 0.44 Narrow coastal plain, high tourism area, overlying fringing reef 281 47 Tuvu 1.0 Northern coast: Mangrove fringed coast bordered by barrier reef. High density sugarcane plantations. 37 5 Suva Peninsula 5.575 Urban areas: Suva, Lautoka, Lami, Nadi Rewa river delta: 78 10 Western Rewa 1.0 Southeast Viti Levu. Low-lying delta with mangrove systems river delta 566 90 Total Annex A Page 10 The erosion analysis used three of the case study sites: Korotogo, Tuvu, and Rewa river delta. The urban center of Suva was not chosen for this analysis as the center is almost entirely protected by seawalls and the shoreline would not respond naturally to erosion processes. The analysis involved three key steps: Step 1. Estimate Potential Shoreline Retreat for Case Study Sites. The estimated potential erosion for all three case study sites is shown on table A.7, as obtained through the Shoreline Translation Model. Step 2. Extrapolate Erosion to the Rest of the Coast. To Table A.7. Potential Shoreline Retreat in Viti estimate the impact of sea level rise on whole coast of L Viti Levu, the potential shoreline retreat at the case study Sea Level Potential Shoreline Retreat sites was multiplied by the length of coast that each study Rise (meters) (in meters) Korotogo Tuvu Rewa site represented. 0.1 0.6 6.7 50.0 Example: For a sea level rise of 0.4 meters, the 0.2 1.3 9.0 112.2 extrapolated shoreline retreat for the southern coast 0.3 1.9 10.6 181.7 (represented by Korotongo) is: 0.4 3.1 11.5 250.7 0.5 3.7 12.9 318.6 0.6 4.9 15.9 385.6 0.7 5.5 18.2 451.8 170 km 0.8 6.7 21.4 517.1 0.9 7.3 25.5 581.9 3.1 m 1.0 8.5 29.1 645.6 3.1 meters x 170 kilometers (or 170,000 meters) = 527,000 square meters or 53 hectares. The extrapolated land lost to erosion for other rises in sea level and land types is as shown on table A.8 (the shaded cell corresponds to the example above). Adding all three coastal land types, the total land lost to erosion is estimated at 590-1,150 hectares in 2025, 1,150-2,330 hectares in 2050, and 2,910-6,000 hectares in 2100 (numbers are rounded). Step 3. Estimate Proportion of Coastal Land that is Lost to Erosion. The total area of coastal land in Viti Levu that is below 10 meters is 600 square kilometers, or 60,000 hectares. Dividing the area lost to erosion by this figure, the percentage of coastal land below 10 meters lost to erosion is 1-2 percent in 2025, 2-4 percent in 2050, and 5-10 percent in 2100 (see last shaded row on table A.8). Table A.8. Estimated Coastal Retreat, Land Eroded and Value of Eroded Land for Viti Levu Sites Estimate 2025 2050 2100 Best guess Worst case Best guess Worst case Best guess Worst case Sea level rise 0.1 0.2 0.2 0.4 0.5 1.0 Southern Potential retreat (in meters) at Korotogo 0.6 1.3 1.3 3.1 3.7 8.5 Coast Extrapolated land eroded (in hectares) 10 22 22 53 63 145 Value of eroded land (in US$ million) 0.2 0.4 0.4 1.0 1.2 2.8 Northern Potential retreat (in meters) at Tuvu 6.7 9.0 9.0 11.5 12.9 29.1 Coast Extrapolated land eroded (in hectares) 188 253 253 323 362 818 Value of eroded land (in US$ million) 4.3 5.8 5.8 7.3 8.2 18.6 Rewa Potential retreat (in meters) at Rewa 50.0 112.2 112.2 250.7 318.6 645.6 Delta Extrapolated land eroded (in hectares) 390 875 875 1,955 2,485 5,036 Value of eroded land (in US$ million) 8.9 19.9 19.9 44.5 56.5 114.5 Total Extrapolated area eroded (in hectares) 588 1,150 1,150 2,331 2,910 5,999 Percentage of coastal land (below 10 meters) 1 2 2 4 5 10 Value of eroded land (in US$ million) 13.3 26.1 26.1 52.8 66.0 136.0 Annex A Page 11 Example: Total area lost to erosion in 2050 (1,150 to 2,331 hectares) / 60,000 hectares = 1.9-3.9 or 2-4%. B. Economic Impact of Coastal Erosion due to Sea Level Rise The economic losses due to erosion can be estimated in three key steps: Step 1: Estimate the Economic Value of a Hectare of Land Lost. This analysis starts by assessing the land types likely to be lost to erosion, and estimating their economic value per hectare. The economic value of land is equal to the expected value of its future use. The two major types of land use found in the study sites are mangroves (Tuvu and Rewa) and tourism and habitation (Korotogo): Korotogo high value, tourism and habitation use (southern coast) Tuvu mangrove land Rewa delta mangrove land 1.1. Economic Value of Mangrove Land: the economic value of mangrove land is the present value of its use and ecological functions (both present and future). In Viti Levu, mangroves are important as a source of subsistence and commercial fisheries, medicinal plants, and raw materials such as firewood. Mangroves are also important habitats for numerous species, and play key roles in coastal protection. ˇ Annual Value of Subsistence Fishing. There is no official estimate of the value of subsistence fisheries in Fiji, though its volume is estimated at about 18,000 metric tons a year (FFD 1999). Estimates from this report (see Table 3.1, Volume I) indicate a value of F$7.7-$13.3 million a year, but this reflects only its value to food security. By lack of a better estimate, one can multiply the volume of subsistence fisheries by the average value of artisanal fisheries (F$3.0/kg), as reported by the Fisheries Division yearbook (FFD 1999), yielding an estimate of F$53.5 million in 1998. From discussions with Fisheries Division experts, mangroves are assumed to account for 35-40 percent of the coastal fisheries production. Viti Levu accounts for 58 percent of the land area in Fiji and for 77 percent of its population. It is therefore assumed that Viti Levu accounts for 58 to 77 percent of Fiji's mangrove fisheries. The higher estimate (77 percent) is justified by the fact that mangrove use is affected by population pressure and Viti Levu, as a high island, has a larger proportion of Fiji's mangroves than other islands. Thus: F$53.5 million x 35-40 % x 58-77 % (adjustment for Viti Levu) = F$10.9 to F$16.5 million The total area of mangroves in Viti Levu is estimated at 23,500 hectares (Watling 1985). Hence, the annual value per hectare of subsistence fisheries associated with mangroves is estimated to be: (F$10.9 to F$16.5 million) / 23,500 hectares = F$464 to F$702 per hectare per year ˇ Annual Value of Commercial Fishing. The Fisheries Division yearbook lists the value of artisanal fisheries catch in 1998 at F$20.7 million (FFD 1999). Industrial fisheries are not counted here, since they include primarily offshore (tuna) fishing. Using the same assumptions as for subsistence fisheries, the value of mangroves associated with commercial fishing is: F$20.7 million x 35-40% x 58-77 % (adjustment for Viti Levu) = F$4.2 to F$6.4 million F$4.2 to F$6.4 million / 23,500 hectares = F$179 to F$272 per hectare per year Annex A Page 12 ˇ Annual Value of Medicinal Plants. The value of mangroves as a source of medicinal plants is estimated at about US$200 a year per rural household in accordance with a previous worldwide study (Constanza and others 1997). Using a 1998 exchange rate of 1US$=1.96 F$, this yields a value of F$392 per rural household per year. The total number of households in Fiji in 1991 was 136,363 according to UNDP (1997). Using the population growth rate for Fiji and assuming no change in average household size, the number of households in 1998 is estimated at about 144,185 in 1998. Assuming that 77 percent of these households are in Viti Levu (in accordance with the total population distribution), and that 60 percent are rural households: 144,185 x 77 % (for Viti Levu) x 60 % ( rural households) = 66,613 rural households in Viti Levu. Assuming that 40 to 60 percent of medicinal plants originate from mangroves: 66,613 x F$392 x 40-60 % = F$10.4 to F$15.4 million Converting this value into a per hectare value yields: (F$10.4 to F$15.4 million) / 23,500 hectares = F$442 to F$655 per hectare per year. ˇ Annual Value of Raw Materials. Mangroves are an important source of fuelwood and construction materials. Based on a worldwide review by Contanza et al. (1997), the annual value of raw materials are estimated at around F$349 per hectare per year. ˇ Annual Value of Habitat Functions. Mangroves support important biodiversity. Based on a study for Thailand conducted by Christensen (1982), the annual value of habitat functions are estimated at around F$308 per hectare per year. ˇ Annual Value of Coastal Protection. Mangroves help trap sediments, protecting the coast against erosion. The value of coastal protection in Fiji was estimated by Sistro (1997) at F$2,896 per hectare per year. The total annual value of mangrove land is therefore: (F$464 to $702) + (F$179 to $272) + (F$442 to $655) + F$349 + F$308 + F$2,896 = F$4,638 to F$5,182 per hectare per year Converting to US dollars (1 US$ = F$1.96 in 1998) / US$2370-$2,644 per hectare per year Or an average of US$2,505 per hectare per year. It is assumed that on eroded land the entire value of mangroves is lost. This assumption seems reasonable from discussions with experts (note, however, that the assumption does not hold for inundated land, where a net gain in mangrove area may be possible as a result of climate change). The economic value of mangrove land lost to erosion needs to take into account the opportunity value of the land, or the stream of benefits that would occur in the future if the land was not lost to erosion. In general, people value future benefits less than present-day benefits. The value today of a stream of future benefits is called the Present Value, and the rate at which people discount future benefits is called the discount rate. In perfect markets, the discount rate is equal to the opportunity cost of capital, around 10 Annex A Page 13 percent. Assuming that the stream of annual benefits would continue for 25 years,3 the present value of mangrove land lost to erosion is as follows: Present Value (10 % discount rate, 25 years, US$2,505 per hectare per year) = US$22,738 per hectare. 1.2. Economic Value of Land in Southern Coast. The economic value of land in the Southern coast was derived from estimates made by Beagley (1998) for a 150 hectare tourism development site in southern Viti Levu, and checked against the prices of real estate as listed in the web sites of various agents (namely Fiji Real Estate 2000). The estimates are as follows: F$507,000 per hectare for resort development areas (7 percent of the land) F$41,825 per hectare for land adjacent to roads (17 percent of the land) F$5,436 per hectare for agriculture land (76 percent of the land) The value of resort development land is roughly similar to its residual value taking into account future use and therefore represents its economic value. Weighing the values of land by the area they represent (and rounding off the numbers) gives a weighted average value of: (F$507,000 x 7 %) + (F$42,000 x 17%) + (F$5,500 x 76%) = F$46,810 per hectare. These estimates are based on freehold land, which can be sold. Up to 91 percent of the land in Viti Levu, however, is held under customary tenure and can only be leased. Leased land costs generally 20 percent less than freehold land (Robert Gillett, personal communication, May 2000). Thus, an additional adjustment is needed: (F$46,810 per hectare x 9 %) + [80 % x (F$46,810 per hectare x 91%)] = F$38,300 per hectare freehold land (9%) crown or native lease (91%) Converting into US dollars (at US$1 = F$1.96 in 1998), gives an estimate of US$19,600 per hectare The higher value of mangrove land as compared to land in the southern coast reflects economic values that are not taken into account in market transactions, such as the protection, habitat, and subsistence value of mangroves. Step 2: Estimate the Economic Value of Land lost to Erosion. With the two estimates of economic value, US$22,738 per hectare for mangrove land in Tuvu and Rewa delta, and US$19,600 per hectare for the southern coast, one can easily derive the economic value of eroded land (table A.8). Example: Table A.8 shows that by 2050, under the worst case scenario, the southern coast will lose 53 hectares of land. At a value of US$19,600 per hectare, the economic losses are estimated at: 53 hectares x US$19,600 per hectare = US$1.0 million For the northern coast, under the same assumptions of sea level rise, 323 hectares of land could be lost. At a value of US$22,738 per hectare, the economic losses could total: 323 hectares x US$22,738 per hectare = US$7.3 million By adding the economic impacts on the three land types, the estimated economic losses due to erosion for Viti Levu amount to US$13.3-$26.1 million in 2025, US$26.1-$52.8 million in 2050, and US$66.0-$136.0 million in 2100 (table A.8). The lower range reflects a best guess scenario, the higher range a worst case scenario in 2050. 3After 25 years, the value today of future benefits would be so small because of the discount rate as to be negligible. Annex A Page 14 Step 3: Estimate the Annualized Value of the Losses. The last step in the economic analysis involves calculating the annualized losses. This can be estimated through the capital recovery factor: [i(1+i)n] / [(1+i)n -1] where i = rate of interest (10 percent) and n = number of years when losses can accrue. In the case of Viti Levu, n was assumed to be 25 years. Hence, [i(1+i)n] / [(1+i)n -1] = [0.1(1.1)25] / [(1.1)25 -1] = 0.11 Applying this factor to the capital losses gives annualized losses of: 0.11 x US$13.3-$26.1 million = US$1.5 to $2.9 million (in 2025) 0.11 x US$26.1-$52.8 million = US$2.9 to $5.8 million (in 2050) 0.11 x US$66.0-$136.0 million = US$7.3 to $15.0 million (in 2100) While the erosion analysis was based on profiles that represented large sections of the coast, caution is needed in interpreting the extrapolations, as the topography of Viti Levu varies greatly. Nonetheless, the extrapolated values provide an order of magnitude estimate of potential impacts due to climate change. Impact of Inundation Caused by Climate Change on Viti Levu's Coast A. Physical Impact of Inundation due to Sea Level Rise The inundation analysis was carried out in three sites: Tuvu, Korotogo and Suva Peninsula. The Rewa river delta was excluded due to the difficulty of separating natural flooding from sea level rise over very low gradient delta surfaces. The analysis of physical impacts involved six steps: Step 1: Determine Baseline Conditions. The Mean High Water Spring (MHWS) tide level in Viti Levu is 0.64 meters. A 1 in 50 year storm surge can increase MHWS by 0.98 meters, to a total of 1.62 meters. Table A.9. Water Level under Baseline Step 2: Estimate Projected Changes in Water Conditions and Sea Level Rise in Viti Levu, Level under Sea Level Rise. The water levels for Fiji the different sea level rise scenarios are as given in the first column of table A.9. Water Level Equivalent Scenario (in meters) Step 3: Estimate the Area of Inundated Land. For each of the three case study sites, the 0.64 1998 Baseline MHWS potential inundation was estimated by raising the 0.75 2025 Best Guess MHWS by sea level rise increments and estimating the land affected by inundation. The 0.87 2025 Worst Case 2050 Best Guess results are shown in table A.10. 1.10 2050 Worst Case Step 4: Estimate the Number of Structures. The 2100 Best Guess number of buildings, extent of roads, mangrove 1.67 1998 Baseline MHWS + storm surgea and sugarcane land, and railway affected by 2100 Worst Case potential inundation were counted and estimated 2025 Best Guess + storm surge for each of the case study sites (see table A.10). 1.85 2025 Worst Case + storm surge These include buildings and roads (in the Suva 2050 Best Guess + storm surge Peninsula), and mangrove, sugarcane plantations, 2.12 2050 Worst Case + storm surge and railways in Tuvu. In Korotogo, only land is 2100 Best Guess + storm surge expected to be affected. Notes: Current (baseline conditions) of sea level are shaded. aThe baseline with storm surge is 1.62 meters, but it approximates the conditions under the other scenarios. Table A.10. Potential Inundation at Case Study Sites, Viti Levu Annex A Page 15 Water Level Scenario (m) Suva Peninsula Korotogo Tuvu Land Buildings Roads Land Land Mangrove Sugarcane Railway (m2) (no.) (m) (m2) (m2) (m2) (m2) (m) 0.64 1998 Baseline MHWS 0.75 2025 Best Guess 15,000* 5,000a 5,000a 0.87 2025 Worst Case 2050 Best Guess 38,705 1,570 10,066 10,066 1.10 2050 Worst Case 2100 Best Guess 60,581 2,457 120,793 120,793 1.67 1998 Baseline MHWS + storm surge 2100 Worst Case 2025 Best Guess + storm surge 475,441 87 5,943 4,662 264,568 192,310 1.85 2025 Worst Case + storm surge 2050 Best Guess + storm surge 572,035 112 6,126 5,397 266,051 192,310 2.12 2050 Worst Case + storm surge 668,628 137 6,310 6,132 267,535 192,310 1,845 1,000 2100 Best Guess + storm surge Notes: a. Estimate only. When the difference in water levels of two scenarios were too close to allow for determination of changes in inundation, the scenarios were combined and the highest water level was depicted. Step 5: Extrapolate for Viti Levu. The linear length of coast for each of the sites surveyed is as follows: Tuvu - 1,000 meters or 1.0 km Korotogo - 440 meters or 0.44 km Suva Peninsula - 5,575 meters or 5.575 km To extrapolate the case study sites to Viti Levu, the inundation values of table A.10 were first converted into inundation per linear kilometer of coast by dividing them by the length of coast of the study sites. Example: For a 2050 Best Guess scenario, the inundation per linear kilometer of coast for the Suva Peninsula is: 38,705 m2 / 5.575 kilometers = 6,942 m2 inundated per linear kilometer of coast For Korotogo, under the same scenario: 1,570 m2 / 0.44 kilometers = 3,569 m2 inundated per linear kilometer of coast For Tuvu, under the same scenario: 10,066 m2 / 1.0 kilometers = 10,066 m2 inundated per linear kilometer of coast Second, the inundated areas were extrapolated to the length of coastline that the case study sites represent (table A.6). Example: For the same scenario above, the extrapolated land inundation in hectares for urban areas of Viti Levu is: 6,942 m2 per linear kilometer of coast x 37 kilometers = 256,854 m2 or 25.7 hectares. For Korotogo: 3,569 m2 per linear kilometer of coast x 170 kilometers = 606,730 m2 or 60.7 hectares. For Tuvu: Annex A Page 16 10,066 m2 per linear kilometer of coast x 281 kilometers = 2,828,546 m2 or 282.9 hectares. A similar extrapolation can be done for structures. For example, for a 2050 Best Guess scenario with storm surge in Suva: 112 buildings / 5.575 kilometers = 20 buildings per linear kilometer of coast and extrapolating to all urban areas: 20 buildings per linear kilometer of coast x 37 kilometers = 742 buildings (with rounding). The final results of this extrapolation can be seen on table A.11. In Tuvu, to allow for an estimate of the value of land not under mangrove or sugarcane use, a new category was created entitled "other land." This is equal to Total Land - Land under Mangrove - Land under Sugarcane. Table A.11. Potential Inundation Extrapolated to Viti Levu Water Level Scenario Southern (m) Urban Areas Coast Northern Coast Land Buildings Roads Land Mangrove Sugarcane Other Railway (ha) (no.) (km) (ha) (ha) (ha) Land (ha) (ha) 0.64 1998 Baseline MHWS 0.75 2025 Best Guess 10.0 140.5 0.87 2025 Worst Case 2050 Best Guess 25.6 60.7 282.9 1.10 2050 Worst Case 2100 Best Guess 40.1 94.9 3,394.3 1.67 1998 Baseline MHWS + storm surge 2100 Worst Case 2025 Best Guess + storm surge 314.9 576 39.4 180.1 5,403.9 2,030.4 1.85 2025 Worst Case + storm surge 2050 Best Guess + storm surge 378.9 742 40.6 208.5 5,403.9 2,072.1 2.12 2050 Worst Case + storm surge 2100 Best Guess + storm surge 442.8 907 41.8 236.9 5,403.9 51.8 2,113.8 281.0 Note: Estimates may differ slightly due to rounding. Step 6: Estimate Proportion of Coastal Land that is Inundated. As seen above, the coastal area of Viti Levu below 10 meters altitude is 60,000 hectares. By comparing the total land inundated with this figure, one can estimate the proportion of coastal land (below 10 meters) likely to be inundated by sea level rise. This is simply computed as the ratio between the total land likely to be inundated (the sum of total land inundated for urban areas, and the south and north shores) and the 60,000 hectares of coast that are found at below 10 meters altitude (table A.12). Example: The proportion of coastal land likely to be inundated under a 2050 Worst Case scenario is: (40.1 + 94.9 + 3,394.3) / 60,000 = 5.9 percent. Annex A Page 17 Table A.12 Proportion of Coastal Land Likely to be Inundated by Sea Level Rise Water Level Scenario Total Land Area Inundated (m) (in hectares) Urban South North Total Viti Proportion of Areas Coast Coast Levu Coastal Area Below 10 m 0.64 1998 Baseline MHWS 0.75 2025 Best Guess 10.0 140.7 150.7 0.3% 0.87 2025 Worst Case 2050 Best Guess 25.6 60.7 282.9 369.2 0.6% 1.10 2050 Worst Case 2100 Best Guess 40.1 94.9 3,394.3 3,529.3 5.9% 1.67 1998 Baseline MHWS + storm surge 2100 Worst Case 2025 Best Guess + storm surge 314.9 180.1 7,434.4 7,929.4 13.2% 1.85 2025 Worst Case + storm surge 2050 Best Guess + storm surge 378.9 208.5 7,476.0 8,063.4 13.4% 2.12 2050 Worst Case + storm surge 2100 Best Guess + storm surge 442.8 236.9 7,517.7 8,197.5 13.7% Note: Total area inundated in north coast equals total area mangrove + total area sugarcane + total other land on table A.11. B. Economic Impact of Inundation due to Sea Level Rise This analysis can be done in four major steps: Step 1: Estimate the economic value of land and structures lost to inundation. The erosion analysis showed how the economic value of land in the southern coast and land under mangroves could be computed. For the inundation analysis, additional estimates are needed for land, buildings and roads in urban areas, and sugarcane and railway values in the northern coast. 1.1 Economic Value of Land in Southern Coast. As estimated before, the economic value of land in the southern coast of Viti Levu is about US$19,600 per hectare. Because inundation is likely to recur, inundated land is assumed to lose all of its economic value. 1.2 Economic Value of Mangrove Land in Northern Coast. Contrary to coastal erosion, which may lead to a replacement of mangrove species and a net loss, the impact of inundation on mangrove ecosystems is likely to be neutral or even beneficial. Most mangroves should be able to adapt to slow inundation, and the deeper roots resulting from sea level rise could help trap additional nutrients (Andrew Hooten, personal communication, June 2000). Consequently, no major economic losses resulting from inundation of mangroves are expected, and the inundated mangrove land is costed at US$0 per hectare. 1.3. Economic Value of Urban Land. According to a recent survey conducted by Margaret Chung and consultations with a major real estate firm in Suva, the value of developed freehold land in the city averages about F$40,000 per acre, or F$100,000 per hectare. Assuming that only 9 percent of the land is under freehold status, and that non-freehold land sells for 20 percent less, the weighted average value of land is, at 1998 exchange rates (1 US$ = F$1.96): Annex A Page 18 (F$100,000 x 9%) + [0.8 x (F$100,000 x 91%)] = F$81,800 per hectare / 1.96 = US$41,735 per hectare It is assumed that the market price of urban land reflects its economic value. 1.4. Economic Value of Urban Buildings. The coastal survey in Fiji was conducted just prior to the coup of May 2000. Subsequently, it has been difficult to obtain accurate estimates of real estate values. Nonetheless, an estimate made by Robert Gillett (Gillett, personal communication, May 2000), indicates the following values for shorefront houses in Suva: About 1/3 of the houses are of original boxy Indian-style costing approximately F$150,000 About 1/3 of the houses have been rebuild, at values of approximately F$250,000 About 1/3 of the houses are luxurious dwellings worth about F$400,000 each For the purpose of the analysis, it is assumed that inundation would destroy about 50 percent of the boxy houses, 15 percent of the rebuilt houses, and only about 5 percent of the luxury houses (as these are generally built to withstand storms). Hence: (F$150,000 x 50%) + (F$250,000 x 15%) + (F$400,000 x 5 %) = F$132,500 or US$67,600 per building. This estimate is conservative, since it does not take into account the higher value of commercial buildings that could be affected by inundation. 1.5. Economic Value of Urban Roads. According to a JICA proposal (1998), the cost of building a 1.1 kilometer of roadway in a residential area in Nadi was F$330,000 in mid-1997. Road construction and improvements elsewhere were estimated at F$530,000 for 1.5 kilometers of road. These estimates suggest an average cost of road construction of about F$300-$353,000 per kilometer. Adjusting to 1998 values by the housing consumer price index (from IMF reports), yields an estimate of F$314-$361,800 per kilometer or US$160-$184,600 per kilometer. The impact of inundation on roads is difficult to assess. However, discussions with infrastructure experts suggest that it is not unreasonable to assume a loss of 25 percent of the value of the road, due to the need for higher maintenance and reconstruction. Hence, the economic value of roads lost to inundation is estimated at: Average of US$160-$184,600 is US$172,300 x 25 % = US$43,000 per kilometer of road. 1.6. Economic Value of Sugarcane Land in Northern Coast. The economic value of sugarcane land depends on the opportunity cost of the land use, in this case sugarcane plantations. It is assumed that inundated land would lose all of its sugarcane production value. The average production of sugarcane accounting for possible reductions in yield due to climate change is 52.7 metric tons per hectare per year (FAO 1996 and study estimates). The price of sugarcane adjusted to 1998 values, was F$75.7 per metric ton. Hence, the value of sugarcane land is: F$75.7 per metric ton x 52.7 metric tons/ha/year = F$3,987 or US$2,034 per hectare per year. Similarly to the value of mangrove land in the erosion analysis, it is necessary to compute the future benefit stream of sugarcane production lost as a result of inundation. Using a 10 percent discount rate and a benefit stream of 25 years, the present value of sugarcane land is as follows: Present Value (US$2,034 per year, 25 years, 10% discount rate) = US$18,463 per hectare. Annex A Page 19 1.7. Economic Value of Other Land. A significant amount of land in the northern coast does not fit the description of mangrove or sugarcane areas. Unfortunately, no estimates were available of the value of this land. Until better estimates are available, the assumption was made that this land was worth 80 percent of the agricultural land value in the southern coast (F$5,500 per hectare). Hence, the value of the "other land" lost to erosion is: F$5,500 per hectare x 80 % = F$4,400 or US$2,200 per hectare 1.8. Economic Value of Railways. No data were available to compute the potential impact of inundation on railways. The railways found in the northern shore are most likely used for sugarcane transport, and are only affected under a 2050 worst case scenario or 2100 best guess scenario with storm surge. Step 2: Estimate the Economic Value of Land and Structures Lost to Inundation. From the calculation above, the value of land and structures lost to inundation is as follows: Unit values of Land and Structures Lost to Inundation: Urban land US$41,735 per hectare Urban buildings US$67,610 per structure Urban roads US$43,000 per kilometer Land in southern coast US$19,600 per hectare Mangrove land in northern coast US$0 per hectare Sugarcane land in northern coast US$18,463 per hectare Other land US$2,200 per hectare Railway not computed These estimates were used to compute the total economic losses due to inundation, by multiplying the unit values by the estimates of land and structures lost on table A.11. Example: The urban land lost to inundation under the Best Guess scenario for 2050 is 25.6 hectares. At a value of US$41,735 per hectare, the total loss is 25.6 hectares x US$41,735 per hectare = US$1,068,416 or US$1.1 million. Step 3: Estimate the Total Economic Loss Caused by Climate Change. The next step in the analysis is to estimate the losses that can be attributed to climate change. For this, it is necessary to compare future conditions with the present baseline. Future conditions with storm surge (SS) need to be compared to the appropriate baseline with storm surge (the second shaded area on table A.13). Hence, for a 2050 Best Guess (BG) scenario, the economic losses due to climate change are as follows: Without storm surge: 2050 BG - 1998 Baseline MHWS = US$2.3 million With storm surge: 2050 BG with SS - 1998 Baseline with SS = US$76.4-61.8 million = US$14.6 million The storm surge conditions reflect a 1 in 50 year storm. Assuming conservatively that there is no increase in storm frequency, conditions without storm surge are likely to occur in 49 out of 50 years. Storm surge conditions would continue to occur in 1 out of 50 years. Hence, the weighted average is: (US$2.3 million x 49/50) + (US$14.6 million x 1/50) = US$2.5 million Annex A Page 20 Table A.13. Economic Value of Land and Structures Lost to Inundation in Viti Levu (millions of 1998 US$) Water Level Scenario Southern Northern Coast Total (m) Urban Areas Coast Losses Land Buildings Roads Land Mangrove Sugarcane Other Land Land Land 0.64 1998 Baseline MHWS 0.75 2025 Best Guess 0.4 0.0 0.4 0.87 2025 Worst Case 2050 Best Guess 1.1 1.2 0.0 2.3 1.10 2050 Worst Case 2100 Best Guess 1.7 1.9 0.0 3.6 1.67 1998 Baseline MHWS + storm surge 2100 Worst Case 2025 Best Guess + storm surge 13.1 39.0 1.7 3.5 0.0 4.5 61.8 1.85 2025 Worst Case + storm surge 2050 Best Guess + storm surge 15.8 50.2 1.7 4.1 0.0 4.6 76.4 2.12 2050 Worst Case + storm surge 2100 Best Guess + storm surge 18.5 61.3 1.8 4.6 0.0 1.0 4.7 91.9 Note: Shaded areas indicates baseline values. For a 2050 Worst Case scenario, the economic losses due to climate change are as follows: Without storm surge: 2050 Worst Case (WC) - 1998 Baseline MHWS = US$3.6 million With storm surge: 2050 WC with SS - 1998 Baseline with SS= US$91.9-61.8 million=US$30.1 million Weighing the above estimates by the frequency of the storm surge: (US$3.6 million x 49/50) + (US$30.1 million x 1/50) = US$4.1 million Step 4: Estimate Annualized Losses Caused by Climate Change. Similarly than for the erosion analysis, the above costs reflect capital losses. In order for the losses to be expressed in annualized terms, it is necessary to multiply the capital losses by the capital recovery factor of 0.11: 0.11 x US$2.5 million = US$0.28 million 0.11 x US$4.1 million = US$0.45 million Though the annualized losses (US$0.3-$0.5 million) appear small, it should be remembered that in years of strong storm surge, Viti Levu could experience incremental losses in the order of US$14.6-30.1 million. If no account is made of current storm impacts, the absolute value of these losses could amount to US$76 to US$92 million by 2050. Annex A Page 21 Impacts of Climate Change on Coral Reefs of Viti Levu Climate change is expected to affect coral reefs primarily through increases in sea surface temperature (resulting in coral bleaching), sea level rise, and the impact of stronger cyclones and storm surges, which lead to extensive destruction of the reef system. These impacts need to be computed separately from those of land erosion and inundation. Step 1: Estimate the Total Area of Coral Reefs in Viti Levu. The area of coral reefs surrounding Viti Levu is unknown, though Fiji has about 1 million hectares of reef (WRI 1999). Several approaches are possible to estimate Viti Levu's reef area: the first is to assume that Viti Levu's share of Fiji's land area also applies to its coral reefs. Under this approach, Viti Levu would have 58 percent of Fiji's reefs, or 580,000 hectares. This estimate is clearly exaggerated, however. While Viti Levu is a large island, many other small islands in Fiji are surrounded by coral reefs, with a ratio of coral reef to land area much greater than Viti Levu. The more correct approach albeit subject to future confirmation is to assume that the area of coral reef is broadly proportioned to the length of the coast line. The total coastal area of Fiji is 5,010 kilometers (Robert Gillett, personal communication) while the coastal area of Viti Levu is 750 kilometers, or 15 percent of the total for Fiji. This suggests that the area of coral reefs in Viti Levu is 15 percent of that of Fiji, or 150,000 hectares. This estimate was used for the analysis. Step 2: Determine the Products or Functions Lost as a Result of Climate Change. Climate change is likely to affect the following uses and functions of coral reefs: productivity of subsistence and commercial fisheries, tourism, biodiversity or habitat values, and coastal protection. The value of these functions is examined in turn below. As a result to overexploitation and habitat degradation, some 19 percent of Fiji's coral reefs are presently considered to be at high risk, with an additional 48 percent are at moderate risk (WRI 1999). As a low bound estimate, it is assumed that all of the high risk reefs which are already under severe stress would die as a result of climate change. As a high end estimate, it is assumed that all of the high risk reefs plus half of the reefs at moderate risk (24 percent) might also die in the future. Thus, the analysis assumes a mortality of 19 to 43 percent of the total area of coral reef. This is a reasonable estimate given the latest scientific knowledge and the extent of the bleaching that affected 50-100 percent of the corals above 30 meters deep in southern Viti Levu in early 2000 (South and Skelton 2000). Step 3: Estimate the Economic Losses Resulting from Climate Change 3.1. Annual Losses in Subsistence Fishing. Dead or bleached reefs do not necessarily lose their fisheries value, as they tend to be quickly covered by algae and lead to a proliferation of herbivorous fish. In the long term, however, dead reefs will tend to break away, and more substantial losses of productivity can be expected (Clive Wilkinson, Tom Goreau, and Herman Cesar, personal communication, May 2000). It is assumed conservatively that the 19-43 percent of reefs that would die would lose 50 percent of their fisheries value. Though Viti Levu may have only 15 percent of the Fiji reefs, the majority of the artisanal and subsistence catch originates from the island, where 77 percent of the Fiji population live. Hence, it is reasonable to assume that the subsistence fisheries of Viti Levu account for 15 to 77 percent of Fiji's total. As seen in the estimation of mangrove values, subsistence fishing in Fiji was valued at approximately F$53.5 million in 1998, taking into account 1998 official estimates of subsistence catch and the average Annex A Page 22 market price for artisanal fisheries (FFD 1999). Coral reefs are assumed to account for between 35 and 45 percent of the coastal fisheries production (Esaroma Ledua, personal communication, March 2000). The value of subsistence fishing in Viti Levu is therefore: F$53.5 million x 15 to 77 % x 35 to 45 % = F$2.8 to F$18.5 million (total value of subsistence (Viti Levu's share of (Share of subsistence fisheries) ( total catch) (fisheries dependent on coral reefs) Assuming a 19 to 43 percent loss in reefs due to climate change, and a loss of 50 percent in reef fisheries productivity, the impacts of climate change on Viti Levu's subsistence fisheries are as follows: F$2.8 to F$18.5 million x 19 to 43% x 50% = F$0.3 to F$4.0 million per year (Value of subsistence fishing (Estimated reef (loss of fisheries in Viti Levu) mortality due to from dead reefs) climate change) 3.2. Annual Losses in Commercial Fishing. As per the estimation of mangrove values, artisanal fishing in Fiji was valued at F$20.7 million in 1998 (FFD 1999). Using the same assumptions as for subsistence fishing: F$20.7 million x 15 to 77 % x 35 to 45 % = F$1.1 to F$7.2 million (total value of artisanal (Viti Levu's share of (Share of artisanal fisheries) ( total catch) (fisheries dependent on coral reefs) F$1.1 to F$7.2 million x 19 to 43% x 50% = F$0.1 to F$1.5 million per year (Value of artisanal fishing (Estimated reef (loss of fisheries in Viti Levu) mortality due to from dead reefs) climate change) 3.3. Annual Losses in Tourism. While tourism is a F$568 million a year industry in Fiji, only a relatively minor proportion of tourists is likely to cease visitation due to coral reef mortality. A similar phenomena was reported in Maldives after the massive reef mortality in 1997-98. Most tourists were said to visit the Maldives for its beaches and leisure, rather than its reefs. The strange coloration of bleached reefs, and the growth of soft coral and herbivorous fish can in some cases provide additional attractions for snorkelers and divers. The most likely tourist category to be affected by coral reef mortality are dive tourists. Numerous attempts were made to obtain statistics on dive tourism in Fiji, but these data does not appear to be available. Pending further information, it is assumed that extensive coral reef mortality would result in a 15 percent drop in tourism revenues. This is in line with recent estimates from Palau that indicate a 9 percent drop in divers' willingness to pay following the 1997-98 bleaching event (Graham, Idechong and Sherwood 2000). Similar surveys in East Africa indicate that 19 percent of the tourists visiting Zanzibar (and 39 percent of those visiting Monbasa) would likely reroute their travel if they knew coral reefs were bleached (Westmacott, Cesar and Pet Soede 2000). It was similarly difficult to ascertain the share of tourism revenues that is spent on Viti Levu. A conservative assumption is to use Viti Levu's share of Fiji's land mass (58 percent). The impact of climate change on tourism linked to the health of coral reefs is therefore: F$568 million x 58% x 15% = F$49.4 million (annual tourism revenues) (% of tourists to Viti Levu) (proportion of tourism likely to be affected by coral mortality) Annex A Page 23 F$49.4 million x 19 to 43% of the reefs = F$9.4 to F$21.2 million per year (tourism revenues in Viti (estimated coral mortality Levu likely to be affected by from climate change) coral mortality) 3.4. Annual Losses in Habitat Value. In addition to their fisheries, recreational and protection values, coral reefs play important roles in biodiversity as habitats for marine life. No estimate is available for this function in Fiji. However, a parallel study for the Galapagos indicates a value (expressed in 1998 F$) of about F$16 per hectare (de Groot 1992). This estimate is used as here as a fist order approximation only. As discussed above, Viti Levu is assumed to have 15 percent of Fiji's reefs. Hence, the habitat value likely to be lost to climate change would be as follows: F$16 per hectare per year x 150,000 hectares x 19 to 43 % = F$0.5 to F$1.0 million per year (annual habitat value) (estimated area of Viti (estimated coral Levu's reefs) mortality from climate change) 3.5. Annual Losses in Coastal Protection Value. Coral reefs play a vital role in the protection of the coast. Following bleaching or mortality events due to elevations in ocean surface temperature, coral reefs are likely to sustain their production function for some time, but will eventually break and wash away (Clive Wilkinson, personal communication). While the loss of coastal protection caused by bleaching events is technically separate from the impact of sea level rise on coastal areas, it is difficult to isolate the two impacts. Hence, no estimate was done of the loss of coastal protection resulting from coral reef mortality. Estimates from other countries indicate that this value may be substantial, around US$60,000 to US$550,000 per square kilometer of reef (Cesar 1996). Step 4: Add up the Impacts. Since the estimates of the various impacts are already annualized, no further adjustment is required, as done for the erosion and inundation analyses. Hence, the impacts can be simply added to give an estimate of the impacts of climate change on the coral reefs of Viti Levu (table A.14). Table A.14. Total Estimated Annual Economic Impact of Climate Change on the Coral Reefs of Viti Levu, Fiji (millions of 1998 US$) Losses in: Annual Damages Millions of 1998 F$ Millions of 1998 US$ Subsistence fisheries 0.3 to 0.4 0.1 to 2.0 Commercial (coastal) fisheries 0.1 to 1.5 0.05 to 0.8 Tourism 9.4 to 21.2 4.8 to 10.8 Habitat 0.5 to 1.0 0.2 to 0.5 Coastal Protection Others (nonuse values, etc.) + + Total estimated damages 10.3 to 27.7 5.2 to 14.1 Notes: Not available; + Not available, but believed to be substantial. Annex A Page 24 Tarawa Atoll Impact of Erosion Caused by Climate Change on Tarawa's Coast A. Physical Impact of Coastal Erosion due to Sea Level Rise The Tarawa atoll has a land area of 30 square kilometers (about 3,000 hectares). All land in Tarawa can be classified as "coastal" since it rarely rises more than 5.0 meters above sea level, with much of the land at less than 3 meters altitude (McLean 1989). To model the impact of erosion and inundation, two case study sites were selected: Bikenibeu island, representing the type of conditions found in South Tarawa, and Buariki, representing the conditions found in North Tarawa. Two transects were surveyed in Bikenibeu and three transects were surveyed in Buariki (table A.15). Table A. 15. Case Study Sites for Tarawa Coastal Type Area Proportion of Case Study Area of Case (hectares) Tarawa atoll Sites Study Sites (hectares) South Tarawa; densely populated, some coastal protection, location of major infrastructure and government 1,577 50% Bikenibeu 104 North Tarawa: sparsely populated, less developed infrastructure, traditional dwellings, subsistence living 1,526 50% Buariki 293 Notes: Area of South Tarawa and North Tarawa said to be 3,896 and 3,771 acres, respectively, in the study's economic analysis report. Area of Bikenibeu and Buariki derived from table 3.10 of the original Kiribati Vulnerability and Adaptation report. The estimate of physical impact involved three major steps: Step 1. Estimate Potential Shoreline Erosion. The level of shoreline displacement was estimated using the Shoreline Translation Model (STM) described earlier in this Annex, and is shown on table A.16. Table A. 16. Protential Shoreline Retreat in Case Study Sites, Tarawa Sea Level Projected Area of Buariki Eroded Projected Area of Bikenibeu Eroded Rise (meters) North Transect Central Transect South Transect West Transect East Transect Shore Area Shore Area Shore Area Shore Area Shore Area Retreat Eroded Retreat Eroded Retreat Eroded Retreat Eroded Retreat Eroded (meters) (hectares) (meters) (hectares) (meters) (hectares) (meters) (hectares) (meters) (hectares) 0.1 0.6 0.1 0.9 0.2 0.5 0.1 0.5 0.2 0.4 0.2 0.2 1.2 0.2 2.3 0.5 1.0 0.2 1.0 0.4 0.6 0.2 0.3 2.6 0.4 3.0 0.7 2.3 0.4 2.0 0.8 0.8 0.3 0.4 3.2 0.5 4.4 1.0 2.9 0.5 2.5 1.0 1.0 0.4 0.5 4.5 0.7 5.0 1.1 3.6 0.6 3.5 1.4 1.2 0.5 0.6 5.2 0.8 6.5 1.4 4.2 0.7 4.0 1.6 1.5 0.6 0.7 6.5 1.0 7.3 1.6 4.8 0.8 5.0 2.0 1.8 0.7 0.8 7.2 1.2 9.1 2.0 5.5 0.9 5.5 2.2 2.2 0.9 0.9 8.4 1.3 14.2 3.1 6.1 1.0 6.5 2.6 2.7 1.1 1.0 9.5 1.5 30.5 6.7 6.7 1.1 7.0 2.8 3.2 1.3 Notes: Shaded areas indicate the scenarios of sea level rise considered in the study Annex A Page 25 Step 2. Estimate the Proportion of the Study Sites Land that is Lost to Erosion. To compute the proportion of the study sites' land that is likely to be lost to erosion, the projected area eroded in each of the transects was first added up to provide a total area eroded for Buariki and Bikernibeu. Example: For a 0.5 meters sea level rise, the area eroded in Buariki is as follows: 0.7 hectares + 1.1 hectares + 0.6 hectares = 2.4 hectares (north transect) (central transect) (south transect) For a 0.4 meters sea level rise, the area eroded in Bikenibeu is: 1.0 hectares + 0.4 hectares = 1.4 hectares (west transect) (east transect) These totals can then be compared with the area of the study sites (table A.15) to determine the proportion of land lost to erosion. Example: For a 0.5 meters sea level ride in Buariki: 2.4 hectares / 293 hectares = 0.0082 or 0.8 percent. For a 0.4 meters sea level rise in Bikenibeu: 1.4 hectares / 104 hectares = 0.013 or 1.3 percent. The proportion of land eroded for all major sea level rise scenarios is shown on table A.17. Table A.17. Estimated Potential Land Erosion as a Result of Climate Change in Tarawa Areas 2025 2050 2100 Estimate Best guess Worst case Best guess Worst case Best guess Worst case Sea level rise 0.1 0.2 0.2 0.4 0.5 1.0 North Potential land eroded in Buariki 0.4 0.9 0.9 2.0 2.4 9.4 Tarawa (in hectares) Percentage of total land (%) 0.1 0.3 0.3 0.7 0.8 3.2 Extrapolated land eroded in North Tarawa (in hectares) 2.0 4.6 4.6 10.2 12.7 48.8 Value of eroded land (in US$ million) 0.1 0.2 0.2 0.4 0.6 2.1 South Potential land eroded in Bikenibeu Tarawa (in hectares) 0.4 0.6 0.6 1.4 1.9 4.1 Percentage of total land (%) 0.4 0.6 0.6 1.3 1.8 3.9 Extrapolated land eroded in South Tarawa (in hectares) 5.7 9.8 9.6 21.1 28.5 61.8 Value of eroded land (in US$ million) 0.3 0.6 0.6 1.3 1.7 3.7 Total Tarawa 0.4 0.8 0.8 1.7 2.3 5.8 Shaded areas refer to the example given above. Numbers may not add up due to rounding. Annex A Page 26 Step 3. Extrapolate Erosion to the Rest of Tarawa. If it can be assumed that Buariki and Bikenibeu are broadly representative of the conditions found in North and South Tarawa, then the proportions of land lost to erosion can also be applied to the total area of North and South Tarawa to determine the total potential land eroded in the atoll. Example: For 0.5 meters sea level rise in North Tarawa: 0.83 percent x 1,526 hectares = 12.7 hectares (proportion of land lost (area of North Tarawa) to erosion in Buariki) For 0.4 meters sea level rise in South Tarawa: 1.34 percent x 1,577 hectares = 21.1 hectares These estimates are shown on table A.17 above. B. Economic Impact of Coastal Erosion due to Sea Level Rise The analysis of economic impacts involved two key steps: Step 1: Estimate the Economic Value of Land Lost to Erosion. To compute the economic impact of coastal erosion, it is necessary to know the value of land per hectare. The Land Management Division in Kiribati quotes the following values for land in Tarawa: For North Tarawa: Unimproved residential land A$18,000 per acre Commercial land: A$100,000 per acre For South Tarawa: Unimproved residential land: A$20,000 per acre Commercial land: A$100,000 per acre For the purposes of the estimation, and by lack of other data, it was assumed that these values reflect the economic value of land in Tarawa. In North Tarawa, it was assumed that 10 percent of the land was used for commercial purposes, while 90 percent was assumed to be unimproved residential land. In South Tarawa, 20 percent of the land was assumed to be for commercial purposes, while 80 percent was unimproved residential land. Hence, the weighted average value of land is: For North Tarawa: (A$18,000 x 90%) + (A$100,000 x 10%) = A$26,200 per acre For South Tarawa: (A$20,000 x 80%) + (A$100,000 x 20%) = A$36,000 per acre Converting into hectares (1 hectare = 2.5 acres) and US$ (1US$ = A$1.5 in 1998): For North Tarawa: A$26,200 x 2.5 / 1.5 = US$43,670 per hectare For South Tarawa: A$36,000 x 2.5 / 1.5 = US$60,000 per hectare These unit costs can then be applied to the extrapolated land eroded in North and South Tarawa (table A.17) to determine the potential losses due to erosion. Annex A Page 27 Example: For the worst case scenario of 2100, the estimated land eroded in North Tarawa is 49.3 hectares. The eroded land in South Tarawa is 60.6 hectares. Applying the unit costs above: For North Tarawa: 48,8 hectares x US$43,670 per hectare = US$2,131,096 or US$2.1 million For South Tarawa: 61.8 hectares x US$60,000 per hectare = US$3,708,000 or US$3.7 million The total value of land eroded in Tarawa for this scenario is US$5.8 million. The economic costs of erosion are therefore US$0.4-$0.8 million in 2025, US$0.8-$1.7 million in 2050, and US$2.3-$5.8 in 2100. Step 2: Estimate the Annualized Value of the Losses. The values above represent the loss of capital due to inundation. To calculate the annualized losses, it was assumed, in discussions with an expert, that the duration of most structures in Tarawa averaged about 10 years. Using a discount rate of 10 percent, and applying it to the capital recovery factor formula gives the following capital recovery factor: [i(1+i)n] / [(1+i)n -1] where i = rate of interest (10 percent) and n = number of years when losses can accrue (10 years). Hence, [i(1+i)n] / [(1+i)n -1] = [0.1(1.1)10] / [(1.1)10 -1] = 0.16 Applying this factor to the capital losses due to erosion, yields the following annualized losses: 0.16 x US$0.4 to $0.8 million = US$0.1-$0.13 million in 2025 0.16 x US$0.8 to $1.7 million = US$0.1-$0.3 million in 2050 0.16 x US$2.3 to $5.8 million = US$0.4-$0.9 million in 2100 Impact of Inundation Caused by Climate Change on Tarawa's Coast A. Physical Impact of Inundation due to Sea Level Rise The analysis of physical impacts involved four key steps: Step 1: Determine Baseline Conditions. The Mean High Water Spring (MHWS) tide level for Tarawa is 2.53 meters. A 1 in 14 year storm the baseline with storm surge raises the sea level to approximately 3.0 meters under baseline conditions. Step 2: Estimate Projected Changes in Water Level under Sea Level Rise. The water levels for the different sea level rise scenarios are displayed on the first column of table A.18. These are derived by simply adding sea level rise scenarios onto the MHWS baseline. Step 3: Estimate the Area of Inundated Land and Infrastructure. Based on the simulations of the area inundated at various water levels, the land area, number of structures, and length of road inundated were calculated from maps provided by the Kiribati Land Management Division. The results are presented on table A.18. Annex A Page 28 Table A.18. Potential Inundation at Case Study Sites, Tarawa Buariki Bikenibeu Water Level Scenario: Land area Land area (m) affected Structures Roads affected Structures Roads (hectares) (number) (kilometer) (hectares) (number) (kilometer) 2.53 1998 Baseline MHWS -- -- -- -- -- -- 2.75 2025 Worst Case 53 (18%) 196 (59%) 6.55 (77%) 0 (0%) 0 0 2025 Best Guess 3.00 1998 Baseline with storm surge 89 (30%) 213 (64%) 6.55 (77%) 1.9 (2%) 34 ( 2%) 0 2050 Worst Case 2100 Best Guess 3.2 2050 Best Guess + storm surge 161(55%) 229 (69%) 7.5 (89%) (25%) 423 (27%) 1.26 (29%) 3.56 2050 Worst Case +storm surge 233 (80%) 245 (74%) 8.5 (100%) 56 (54%) 986 (63%) 2.83 ( 66%) 2100 Best Guess+storm surge 2100 Worst Case 4.0 2100 Worst Case+ storm surge 248 (85%) 316 (95%) 8.5 (100%) 83 (80%) 1302 (84%) 4.36 (100%) Note: Projected losses in structures and roads derived from topographic maps (Land Management Division of Kiribati). Storm surge is based on a 1 in 14 year event (Solomon 1997). Figures in parenthesis indicate percentage of the land, infrastructure and/or roads inundated (relative to the total for the study sites). Step 4: Extrapolate to all Tarawa. Assuming that Buariki and Bikenibeu are broadly representative of the conditions of North and South Tarawa, respectively, a simple extrapolation can be used to derive the physical impact on the entire atoll. For land, the percentage of land that is projected to be inundated in Buariki and Bikenibeu was applied directoy to the area of North and South Tarawa. For example, for the 2025 Worst Case scenario, 18 percent of the land of Buariki is expected to be inundated. This percentage can be then applied to the total area of North Tarawa, 1,526 hectares (table A.15), yielding a total area of land inundated of approximately 275 hectares (table A.19). Similarly for structures, a simple extrapolation can be used. To calculate the number of structures for North Tarawa under a 2025 Worst Case scenario: No. of structures affected in Buariki (196) x Area of North Tarawa (1,526)/Area of Buariki (293) = 1,021 structures. The results are summarized on table A.19, which also contains the results of the economic analysis explained below. B. Economic Impact of Inundation due to Sea Level Rise The economic impact analysis involved three key steps: Step 1: Estimate the economic value of land and structures lost to inundation. From the economic analysis of erosion impacts, the average unit value of land in South Tarawa was estimated at US$60,000 per hectare, while the value of land in North Tarawa averaged US$43,670 per hectare. Annex A Page 29 Table A.19. Estimated Economic Value of Land and Structures Lost to Inundation Water Level North Tarawa South Tarawa Total (m) Scenario Tarawa Land Lost Structures Lost Land Lost Structures Lost hectares Costs No. Costs Hectare Costs No. Costs Costs (US$M) (US$M) s (US$M) (US$M) (US$M) 2.53 1998 Baseline MHWS - - - - - 2.75 2025 Worst Case 2050 Best Guess 274.7 12.0 1,021 20.4 0.0 0.0 0.0 0.0 32.4 3.00 1998 Baseline + storm surge 2050 Worst Case 2100 Best Guess 457.8 20.0 1,109 22.2 31.5 1.9 516 10.3 54.4 3.2 2050 Best Guess + storm surge 839.3 36.7 1,193 23.9 394.3 23.7 6,414 128.3 212.4 3.56 2050 Worst Case + storm surge 2100 Best Guess + storm surge 2100 Worst Case 1,220.8 53.3 1,276 25.5 851.6 51.1 14,951 299.0 429.0 4.0 2100 Worst Case + storm surge 1,297.1 56.6 1,646 32.9 1,261.6 75.7 19,743 394.9 560.1 Shaded area indicates baseline conditions. The cost of structures was estimated at A$30,000 each (US$20,000), from data from the Kiribati Land Management Division. Table A.19 shows the estimate of economic value of land and structures lost to inundation. The interpretation of the topographic maps did not allow for a differentiation of the type of structure affected. Data were also not available to estimate the economic costs of road losses to erosion. Example: In the 2050 Best Guess Scenario, about 1,021 structures are estimated to be affected by inundation. The value of these losses is therefore: 1,021 x US$20,000 or US$20.4 million. Step 3: Estimate the Net Economic Losses Caused by Climate Change. Similarly to the Viti Levu analysis, the future conditions need to be compared with the correct baseline: for example, if one is estimating conditions under 2050 Best Guess with storm surge, the correct comparator would be a scenario of 1998 Baseline conditions with storm surge. Hence, for a 2050 Best Guess (BG) scenario, the economic losses are as follows: Without storm surge: 2050 BG - 1998 Baseline MHWS = US$32.4 million With storm surge: 2050 BG with SS - 1998 Baseline with SS = US$212.4 - US$54.4 = US$158.0 million The storm conditions reflect a 1 in 14 year storm. Assuming no change in the frequency of storms, storm surge conditions are likely to occur once every 14 years, while no storm surge conditions are likely to occur in 13 out of 14 years. The weighted average loss is therefore: (US$32.4 million x 13/14) + (US$158.0 million x 1/14) = US$41.4 million For a 2050 Worst Case Scenario (WC), the economic losses would be: Without storm surge: 2050 WC - 1998 Baseline MHWS= US$54.4 million With storm surge: 2050 WC with SS - 1998 Baseline with SS = US$429.0 - US$54.4 = US$374.6 million (US$54.4 million x 13/14) + (US$374.6 million x 1/14) = US$77.3 million Annex A Page 30 Step 4: Estimate the Annualized Losses Caused by Climate Change. Similarly than for the erosion analysis, the costs above reflect capital losses. For the losses to be expressed in annualized terms, it is necessary to multiply them by the capital recovery factor, which is 0.16 (see erosion analysis). The annualized costs are therefore: 0.16 x US$ 41.4 million = US$6.6 million for a best guess scenario in 2050 0.16 x US$ 77.3 million = US$12.4 million for a worst-case scenario in 2050. In years of strong storm surge, Tarawa could experience incremental losses in the order of US$158-$375 million by mid-century. If no account is made of current storm impacts, the absolute value of these losses could amount to US$212-$429 million by 2050. Impacts on Coral Reefs and Mangroves Tarawa has lost some 70 percent of its mangroves since 1940, and only 57 hectares remain (Metz 1996). The impact of sea level rise is therefore considered to be negligible, and could be even positive as higher water levels may lead to denser root systems and expansion of mangroves shoreward. Climate change is likely to have a significant impact on coral reefs, however. Bleaching caused by warmer sea surface temperatures and the impact of sea level rise could impact subsistence and commercial fisheries, habitat functions, and the coastal protection role of coral reefs. Losses in tourism value are considered negligible as Tarawa does not have a developed dive or snorkel tourism industry. The physical and economic impact analysis involved three major steps: Step 1: Estimate the Area of Coral Reefs: The reef area of Tarawa is 129 km2 or 12,900 hectares (Stratus 2000). Step 2: Determine the Products or Functions Lost as a Result of Climate Change. No estimate is known of the present status of coral reefs in Tarawa. However, in-country discussions indicate that it is not unreasonable to assume a loss of 10 to 40 percent of the reefs by 2050 under climate change conditions. Similarly to Viti Levu, it is assumed conservatively that dead reefs would retain about 50 percent of their fisheries productivity. Step 3: Estimate the Economic Losses Resulting from Climate Change 3.1. Annual Value of Subsistence Fishing. Government statistics put the value of subsistence fishing in Kiribati at A$6.3 million in 1998. This figure needs to be adjusted for Tarawa. A population proportion (45 percent of the I-Kiribati population lives in Tarawa) rather than an area adjustment was used because several of the islands of Kiribati particularly on the Phoenix group are uninhabited. Hence, A$6.3 x 45 % = A$2.8 million (for 12,900 hectares of reefs) Assuming a 10 to 40 percent reef loss, and a 50 percent loss in fisheries productivity, the value of subsistence fisheries lost is: Annex A Page 31 0.10 to 0.40 x 0.5 x A$2.8 = A$0.14 to $0.57 million, or (at an exchange rate of 1 US$ = A$1.5) US$0.14 to $0.38 million per year 3.2. Annual Value of Commercial Fishing. Commercial fishing in Kiribati is targeted primarily at offshore tuna resources which are not directly related to coral reefs. The value of commercial reef fisheries, which are commonly designated as "artisanal fisheries", is unknown in Kiribati. However, exports of ornamental fish and seaweed were valued at A$0.7 million each in 1998, in accordance with IMF statistics. This almost certainly underestimates commercial fish production, but it is difficult to separate offshore tuna from reef-related fisheries in the fish export category found in official statistics. Hence, only ornamental fish and seaweed value were taken into account in the analysis: A$0.7 million (for seaweed) + A$0.7 million (for ornamental fish) = A$1.4 million Adjusting for Tarawa and for reef loss gives the following economic value of commercial reef fisheries lost to climate change: 0.10 to 0.40 x 0.5 x 45% x A$1.4 million = A$0.032 to A$0.126 or US$0.02- $0.08 million per year 3.3 Annual Habitat Value. A study in the Galapagos (de Groot 1992) estimates the annual habitat value of coral reefs at approximately US$8 per hectare of reef. However, the conditions in the Galapagos are too distinct from those in Kiribati for this estimate to be of relevance. The habitat value of coral reefs in Tarawa is probably substantial, but it was not quantified by the study. 3.4. Coastal Protection Value. Given the difficulties of separating the impacts of coral reef loss from those of inundation and erosion resulting from sea level rise, the coastal protection value of coral reefs was assumed to be largely accounted for in the erosion and inundation analysis, and was therefore not quantified here. Hence, the annual value of coral reefs lost due to climate change in the Tarawa atoll is estimated at US$0.16 to $0.46 million per year. Annex A Page 32 C. Impacts on Water Resources For the estimation of the impacts of climate change on water resources in Viti Levu and Tarawa, the study team followed slightly different methodologies. In Viti Levu, where cyclones are of paramount importance, historical records of these extreme events were used to derive an estimate of the future impact. The impact on climate change on surface hydrology was based on the water resources model developed by PACCLIM. The Tarawa water resources analysis, however, used a SUTRA groundwater model (Voss 1984) which took into account the mixing between freshwater and underlying seawater. This is considered superior and of greater reliability than the "sharp interface" model incorporated into PACCLIM, which assumes a sharp boundary between freshwater and seawater. The specific methods followed in the two case studies are outlined in further detail below. Viti Levu The impacts of climate change on the water resources of Viti Levu were estimated for cyclones, surface hydrology and water balance. No estimate of incremental costs could be made for El Niņo-related droughts, however, given the absence of data on baseline drought conditions. Nonetheless, the impacts of droughts is partially taken into account by the agriculture impact analysis (Section D). Impact of More Intense Cyclones Fiji has the highest incidence of cyclones in the South Pacific, an average of 1.1 cyclones a year (Pahalal and Gawander 1999). To be as representative as possible of present-day conditions, the intensity and frequency of actual cyclones in the 1992-99 period was used as a baseline. Reliance on historical data was necessary since GCM models and PACCLIM do not reflect well the effects of cyclones. The impact analysis involved three key steps. Step 1: Estimate the Future Impact of Cyclones. Jones and others (1999) estimate that cyclone intensity (maximum wind speed) could increase by 0-20 percent by mid-century as a result of climate change. In order to compute the potential impacts of this increase in intensity, it was necessary to estimate how the increase in wind speed will affect cyclone damage. Three estimates were used for this purpose: ˇ The historical records of the Fiji Meteorological Services for seven recent (1983-97) cyclones indicate damages of F$67 million (in 1998 value) for an average cyclone with wind speed of 58 knots. A 20 percent increase in wind speed over this historical average would be equivalent to cyclones Oscar (68 knots, 1983) and Eric (72 knots, 1985), which averaged F$106 million in damages a 59 percent increase (see figure A.3). ˇ From the existing theory, an increase in wind speed by 20 percent should result in a 44 percent increase (1+0.2)2 = 1.44 in wind force (J. Terry, personal communication, May 2000). ˇ A recent study (Clark 1997) indicates that a 15 percent increase in cyclone severity is likely to result in a doubling of the damages, or a 100 percent increase. Annex A Page 33 Figure A.3. Cyclone Damage as a Function of Wind Speed, Fiji Data courtesy from Fiji Meteorological Services. Maximum Cyclone damage Wind Speed (F$ million) Oscar (1983) 68 148 Average cyclone Eric (1985) 72 64 Wind Speed Damages Gavin (1985) 42 2 58 67 Sina (1990) 54 33 Joni (1992) 55 2 Average cyclone + 20% Kina (1993) 60 188 Wind Speed Damages Gavin (1997) 56 35 70 106 Relationship Between Maximum Wind Speed and Damages for Major Cyclones in Fiji (1983-1997) 200 Kina (1993) 160 Oscar (1983) F$million) 120 (1998 80 Eric (1985) Sina (1990) Gavin (1997) Damages 40 Storm Gavin (1985) Joni (1992) 0 0 10 20 30 40 50 60 70 80 Maximum Wind Speed (knots) Source: Fiji Meteorological Services.. Damages are converted to 1998 costs Taking these three estimates into consideration, an increase in wind speed of 20 percent could result in a 44 to 100 increase in total damages. This broad estimate recognizes the large variations that can result from cyclone origin, trajectory and characteristics. Step 2: Estimate the Baseline Conditions. Using the 1992-99 Table A.20. Damages from Cyclones in Fiji period as an analogue for baseline conditions, the average (1992-99) annual damages due to cyclones amount to F$28.3 million per year in 1998 values (table A.20). Cyclone Year Damages (millions of 1998 F$) The Fiji Meteorological Services confirmed that the estimated damages pertain to Fiji as a whole, and not just to Viti Levu. Joni 1992 2 Kina 1993 188 Given the correlation between population concentration and Gavin 1997 35 infrastructure, adjusting the damages by the proportion of the Gale June 1997 1 population that lives in Viti Levu is preferable to adjusting it by area. Hence, the average annual cyclone damages for Viti Average 56.5 Yearly average 1992-99 Levu the baseline for the analysis is: (8 years) 28.3 F$28.3 million per year x 77 % = F$21.8 million per year Source: Fiji Meteorological Services, and background studies to this report. Annex A Page 34 Step 3: Estimate the Future Economic Costs of a Rise in Cyclone Intensity. The impact of a 0 to 20 percent increase in maximum cyclone intensity is as follows: ˇ If the change is 0 percent, climate change has zero incremental impact; ˇ If the change is 20 percent, damages could increase by 44 to 100 percent Thus, for a 20 percent increase in cyclone intensity: 44 % to 100 % of F$21.8 million = F$9.6 to F$21.8 million , or US$4.9-$11.1 million at 1998 exchange rates. Hence, the average annual incremental impact of climate change on cyclone Table A.21. Major Drought-Related Damages and intensity is US$0 to US$11.1 million. Expenditures for the Drought of 1997-1998, Fiji (millions of 1998 F$) However, in absolute (non incremental) terms, the likely cost of an extreme event in Viti Levu would be: Item Cost F$56.5 million x 0.77 = F$43.5 $125 (Average cyclone (Correction Sugarcane crop rehabilitation programme $43.7 Losses in 92-99) for Viti Levu) Food and water rations $33a Weaning foods $3.6 b F$43.5 million / 1.96 x 0-2.0 = Food crop losses $15 c (Losses in (US$ (Projected Crop re-establishment needs $0.41 Viti Levu) conversion) increase in Welfare payments & lost income $75d intensity) School children funding $0.625 e School water tanks and gardens $1.5 = US$0-$44 million Draft animals grass-starved $2.4 f Milk production decline (50%) $10.9g Farm drains and creeks $0.26 h Impact of Droughts Irrigation to re-establish sugarcane $0.040 Boreholes for water supply $0.630 i Commercial forest $80 j With El Niņo conditions expected to Health and nutrition $1.8 k prevail in the future, Fiji could experience Macro impacts + l accentuated droughts. The 1997-98 drought Fire damage (up to 10% of forest areas) + which caused damages of F$275 to F$325 Tourism + million could well become the norm in the Total F$313.9+ future. The drought affected food supplies, commercial crops, livestock, and the water Notes: a. Ten months at $3.3 million per month. supply of schools and communities (table b. F$360,000 per month times 10 months. A. 21). The damages of the 1997/98 c. Estimate of revenue losses from failed new plantings, by Director of Ag Extension. drought a 1 in 100 years event are d. F$2,000 per family times 15,000 farm families, covering 6 months of considered indicative of the type of recovery, prorated to also include 9 months of zero income. There is some droughts that might be expected in the potential double counting with the sugarcane crop loss value on line 1. e. 10,000 children missed school because of hardship. F$500K was future. Since approximately F$125 fundraising target to reach 8,000 still in need. million were losses to sugarcane (mostly f. Molasses block feeding program to restore work animals to work condition. incurred in Viti Levu) and F$15 million g. F$39,600 loss per day, for 275 days (75% of a year). h. One year of 2 year grant program to clean up drains. were losses to other crops, one can assume i. 675 families served by boreholes, at F$937 per family. the following indicative (absolute) costs for j. Re-establish dead plants only, only for mahogany plantation. k. Foreign aid food assistance received. future droughts in Viti Levu: l. Budget deficit of 2.4% GDP; negative GDP growth of 4.0% rather than anticipated 3.0% positive growth. Cost to government is 30% of budget (or 3 times the revenue loss). Source: UNDAC (1998) supplemented with original analysis. Annex A Page 35 The total costs, F$275$325 million, are equivalent to US$140$165 million The agricultural losses, F$125 + F$15 million are equivalent to US$71 million Thus, the non-croped related losses in Viti Levu are: (US$140$165 million) ­ US$71 million = US$69$94 million x 0.77 = US$50-$70 million (rounded) The crop related losses in Viti Levu include all the losses of sugarcane plus 77 percent of the losses in other crops: US$64 million + (US$7.7 million x 0.77) = US$70 million From 1983 to 1998, Fiji experienced four El Niņo related droughts, or 1 every 4 years. With the exception of the 1997-98 event, these more normal droughts could be considered the baseline variability conditions against which to measure a more severe event (such as the 1997-98 drought). Unfortunately, no information exists on the impact of these earlier droughts apart from agricultural yields, and one cannot rule out the possibility that the 1983-98 period already reflects some incremental impacts of climate change. Given these constraints, the incremental impact of climate change on droughts could not be assessed at this stage. About 40-50 percent of the costs of the 1997-98 drought involve agricultural impacts, and these were assessed separately in Section D. Impact of Changes in Precipitation and Temperature on River Flow The impact of rainfall changes on the water resources of Viti Levu was estimated for two water courses, the Nakauvadra Creek and the Teidamu Creek. The Nakauvadra Creek drains 28 km2 of steeplands in northern Viti Levu. The Teidamu Creek drains a 56 km2 watershed in the northwestern part of Viti Levu. Both streams have been monitored by the Hydrology Section of the Fiji Public Works Department since the early 1980s. Projections of rainfall for Fiji under climate change conditions vary according to the two GCM models used by the study, which predict both an increase (with the CSIRO9M2 model) and a decrease in rainfall (DKRZ model). Using the PACCLIM water resources impact model, the results suggest that under the scenarios of climate change considered by the study, there would be a change of ą 10 % in the water flow by 2050 and of ą 20 % by 2100 (table A.22). Given the similar climate, land uses, topography and commercial areas between these creeks and the major Nadi and Ba rivers, similar changes in flow may occur there. A possible increase of 10-20% in flood volume could cause extensive damage to industrial and commercial areas. Conversely, if a lower rainfall scenario materializes, the Nadi, Ba and Sigatoka rivers could experience salt intrusion causing problems for agricultural irrigation. The economic impact of these scenarios could not be assessed due to lack of data. Annex A Page 36 Table A.22 Predicted Future 1 in 10 year Low and High Daily Flows for the Teidamu and Nakauvadra Creeks, under Climate Change Scenarios (Viti Levu) CGM Temperature Rainfall 2025 2050 2100 Model Change Changes Flow conditions Flow conditions Flow conditions oC (%) (% change from baseline) (% change from baseline) (% change from baseline) Low flow High flow Low flow High flow Low flow High flow Teidamu Creek CSIRO9M2 B2 mid 0.5 3.3 +3.9 +3.3 +5.9 +5.7 +9.8 +9.6 A2 high 0.6 3.7 +3.9 +3.7 +8.8 +8.2 +20.6 +20.3 DKRZ B2 mid 0.5 -3.3 -2.9 -3.9 -3.9 -5.7 -9.8 -9.7 A2 high 0.6 -3.7 -2.9 -7.8 -7.8 -8.2 -19.6 -20.3 Nakauvadra Creek CSIRO9M2 B2 mid 0.5 3.3 +3.4 +3.3 +5.7 +5.7 +9.2 +9.7 A2 high 0.6 3.7 +3.4 +3.7 +8.0 +8.2 +19.5 +20.3 DKRZ B2 mid 0.5 -3.3 -3.5 -3.7 -5.7 -5.7 -10.3 -9.7 A2 high 0.6 -3.7 -3.5 -3.7 -8.0 -8.2 -20.7 -20.3 Impact of Changes in Precipitation and Temperature on Water Supply A simulation of climate change scenarios impact was made for the Nadi-Lautoka Regional Water Supply Scheme. The scheme is the second largest in Fiji, serving an estimated 123,000 people in western Viti Levu (PNG Consultants 1996). The safe sustainable yield of the water sources (the Vaturu dam and the Lautoka sources) is 98 million liters a day (ML/day). This is based on a water availability of a 1 in 15 year drought event. The Vaturu dam currently supplies 45 ML/day and the other sources 14.5 ML/day. About 29 percent of the water supply is unaccounted for water, due to leakage and illegal connections. In 1996, the per capita demand for water was 330 l/day (JICA 1998). This is assumed to represent the average per capita consumption of both commercial and domestic users. Future water demand was calculated based on a conservative 300 l/day average per capita, a 25 percent loss rate, and the population growth rates estimated in this study (table A.23). Under the baseline conditions, demand would exceed the sustainable supply in 2051 and 2041 for the mid and high population growth scenarios. Using PACCLIM to estimate the future water supply under climate change conditions, and the results of the Teidamu creek as a proxy, it can be seen that the effects of climate change only become significant by 2050. With a low population scenario, climate change does not have significant impacts apart from the high scenario for 2100. Table A.23. Projected Approximations of Potential Water Surplus (+) or Potential Deficit (-) of the Nadi-Lautoka Water Supply Scheme (presented as deviation from sustainable yield with and without climate change) 2026 2051 2096 Climate No DKRZ B2 DKRZ A2 No DKRZ DKRZ A2 No DKRZ B2 DKRZ A2 scenario/ climate Mid High climate B2 Mid High climate Mid High Population change change change projection Low +24 +21 +20 +12 +7 +4 +10 0 -10 Mid-range +20 +17 +16 0 -5 -8 -18 -28 -38 High +17 +14 +13 -11 -16 -19 -56 -66 -76 Annex A Page 37 Figure A.4. Projections of Future Water Demand and Supply under Climate Change Conditions 160 ) High 120 Sustainable (ML Supply Mid meul B2 80 Vo Population Low A2 Projections tera W 40 0 1990 2040 2090 Year As for the water flow analysis, no data were available to permit an estimation of the economic impacts of climate change. Tarawa The analysis of climate change impacts on water resources of Tarawa used a water balance model to estimate water recharge and then a groundwater model to analyze the impact of decreased recharge and changes in sea level on the thickness of the freshwater lens (Alam and Falkland 1997). The study did not account for changes in evapotranspiration. Impact on Water Recharge In low coral islands with a shallow water table, the water balance equation can be expressed as: R = P - (EI + ES + TL) + dV where P is rainfall, E is actual evapotranspiration, composed of: interception (EI), evaporation and transpiration from the soil zone (ES); and transpiration of deep rooted vegetation directly from groundwater (TL). R is recharge to groundwater, and dV is the change in soil moisture. A study by Alam and Falkland (1997) estimated monthly recharge estimates using the water balance model for the Bonriki freshwater lens, the main water source for Tarawa. The water balance model used by this study (originally for 1954-1991) was adjusted to cover the 1954-1996 period, using daily rainfall data and estimates of daily evaporation data collected at the Betio meteorological station in Tarawa (Station No. J61000). The soil moisture zone at Bonriki is typically 500 millimeters thick based on observations from shallow pits and wells. The field capacity was assumed to be 0.15 based on typical values for sand-type soils (UNESCO 1991). The operating range of soil moisture was assumed to be Annex A Page 38 from 25 to 75 millimeters, and the amount of evaporation from the soil moisture zone was assumed to be linearly related to the available soil moisture content. The predominant vegetation types on the island are coconut trees and a variety of grasses and other shallow rooted vegetation. Crop factors of 1.0 and 0.8 were assumed for, respectively, grasses and shallow rooted vegetation, and coconut trees. The potential evaporation rate for coconut trees was taken to be 80 percent of potential evaporation while that for grasses or other shallow rooted vegetation was assumed to be equal to potential evaporation. Using the water balance model described above, the impact of changes in rainfall on the water recharge at Bonriki was derived by linearly adjusting all daily rainfall values by the selected percentage change. The results are summarized on table A.24. Table A.24. Impacts of Changes in Rainfall on Water Recharge in Bonriki, Tarawa Mean Annual Rainfall % of Current Mean Mean Annual Water % of Current Mean (mm) Annual Rainfall Recharge (mm) Annual Water Recharge 1327 65 415 41 1668* 75 583* 57 1632 80 667 65 1838 90 843 82 2040 (Current) 100 1023 100 2245 110 1207 118 2449 120 1393 136 * - Extrapolated values from table 4.1 in Taueua et al. (2000). The climate change scenarios used by this study project up to -11 to + 7 percent change in rainfall by 2050 and up to -27 to +18 percent change by 2100. As table A.24 indicates, a 10 percent change in rainfall could lead to an almost 20 percent change in the water recharge. Under the worst case scenario in 2100, the water recharge could change by up to - 43 percent (for a 25 percent decline in rainfall) to +36 percent (for a 20 percent increase in rainfall). Analysis of Impacts on Groundwater A. Physical Impacts of Climate Change on the Groundwater Resources of Tarawa The water balance of a small island groundwater system can be expressed as follows: R = GF + D + Q + S where R is net recharge, GF is groundwater outflow (to the sea), D is dispersion at the base of the groundwater, Q is groundwater extraction (normally by pumping), and S is change in freshwater zone storage. The SUTRA groundwater model (Voss 1984) was applied to a selected cross section of the Bonriki freshwater lens to analyze the impact of climate change and sea level rise scenarios on the freshwater thickness. The SUTRA model was calibrated and verified using measured groundwater conditions in the mid to late 1980's. Groundwater data was available from a network of salinity monitoring boreholes at Bonriki installed in the early to mid 1980s. Annex A Page 39 A baseline scenario was created for current conditions (Alam and Falkland 1997). These included recharge to groundwater based on a 20 percent coverage of deep rooted trees (predominantly coconut trees) and a pumping rate of 1,000 cubic m3/day, which is equal to the current pumping rate. The critical time selected for comparison was that occurring in November 1985 at the end of a long drought, which produced the thinnest freshwater lens condition. The impact of eight scenarios of rainfall change, sea level rise, and potential loss of island width on the Bonriki freshwater lens are summarized on table A.25. The scenarios reflect the 2050 conditions forecasted by the models used by the study: they include a best guess (0.2 meters) and worst case (0.4 meters) scenario of sea level rise, and the projected rainfall changes (+ 7 to - 10 percent) under the worst case scenarios4. The analysis also took into account a reduced island width due to inundation of 230 meters, or 19 percent of the total width of the island. This corresponds to the likely area inundated at the station sampled. Table A. 25. Impacts on the Bonriki Freshwater Lens Thickness due to Possible Climate Change Scenarios Average change in freshwater thickness Climate Change and Sea Level Rise Scenario compared with baseline scenario (meters) (%) 1. Baseline scenario (current mean sea level and rainfall; average freshwater thickness = 12.1 m) - - 2. Current MSL, 7% increase in rainfall +0.66 +5.5 3. Current MSL, 10% reduced rainfall -1.7 -14 4. 0.2 meters MSL rise, current rainfall -0.1 -0.9 5. 0.4 meters MSL rise, current rainfall +0.25 +2.0 6. 0.4 meters MSL rise, 10% reduced rainfall -1.4 -12 7. 0.4 meters MSL rise, current rainfall, reduced island width -3.5 -29 8. 0.4 meters MSL rise, 7% increased rainfall, reduced island width -2.3 -19 9. 0.4 meters MSL rise, 10% reduced rainfall, reduced island width -4.7 -38 Note: MSL = mean sea level As expected, the freshwater thickness increases or decreases depending on whether a scenario of increased or decreased rainfall materializes. A rise in mean sea level per se could increase the freshwater lens volume, as the groundwater table (the top of the freshwater lens) rises while its base remains relatively unaffected. However, if sea level rise was accompanied by a reduction of island width due to inundation (which is likely), the freshwater thickness would be reduced by 19 to 38 percent, depending on whether scenarios of increased or decreased rainfall materialized (see scenarios 8 and 9 on table A.25). These estimates reflect the combined impacts of a reduced island width, and `worst case' scenarios of sea level rise and rainfall changes in 2050. A `best guess' scenario of rainfall change could result in relatively lesser impacts, but this might be offset by the lower rise in groundwater table caused by a smaller (0.2 meters) sea level rise. 4Best guess scenarios (+5 to -7.5 percent) fall within these two extremes. Annex A Page 40 B. Economic Impact of Climate Change on the Groundwater Resources of Tarawa For the purposes of the economic analysis, it is assumed that the Bonriki freshwater lens is representative of other groundwater systems in Tarawa, and that changes in the thickness of the lens reflect changes in volume. The current yield of the Bonriki and Buota freshwater lenses is 1,300 m3/day, serving a population of about 26,000. The public water system currently fails to meet demand due to an estimated 50 percent leakage (SOPAC 1998) and the presence of illegal connections. If in the future all islands of North Tarawa were developed for groundwater use, the estimated additional yield would amount to nearly 3,900 m3/day, or a total of 5,200 m3/day with the Bonriki and Buota water supplies. Hence, 5,200 m3/day is taken as the total groundwater capacity for Tarawa, and the basis for the estimated impact. The economic cost of climate change can be estimated based on what it would cost to replace the lost groundwater capacity. Two replacement sources were considered: (i) the costs of rehabilitation and expansion into additional groundwater sources; and (ii) desalination. A third possibility rainwater collectors is likely to be less expensive than desalination, but could not be computed at this stage. The analysis involved three major steps: Step 1: Estimate the Unit Costs of Substitutes. The costs of rehabilitation and expansion into new groundwater sources in North Tarawa are estimated at A$2.6/m3 in 1995 (ADB 1996). The 1995 costs of desalination are estimated at between A$4.0 to A$6.2/m3 (ADB 1996). The higher figure appears to be more accurate. Nonetheless, as desalination costs are likely to decrease in the future, an average of the two estimates was used, or A$5.1/m3. Step 2: Convert into 1998 US dollars. The unit costs above are 1995 estimates. To convert them into 1998 U.S. dollars, an index of inflation such as the Tarawa Retail Price Index needs to be applied: Year Retail Price Index (% change)* 1996 1.5 1997 2.2 1998 4.7 * IMF estimates. For the costs of groundwater expansion: A$2.6 x 1.015 = A$2.64 A$2.64 x 1.022 = A$2.70 A$2.70 x 1.047 = A$2.82 Converting into US dollars by the 1998 exchange rate (1US$ = A$1.5): A$2.82 / 1.5 = US$1.9/ m3. For the costs of desalinated water: A$5.1 x 1.015 = A$5.18 A$5.18 x 1.022 = A$5.29 A$5.29 x 1.047 = A$5.54 Annex A Page 41 Converting into US dollars: A$5.54 / 1.5 = US$3.7/ m3. Step 3: Estimate the Costs of Substitution. Assuming that the percentage reductions in groundwater thickness apply to the total groundwater supply of Tarawa (5,200 m3/day), the economic costs of having to substitute existing supplies by alternative sources as a result of climate change are as shown on table A.26. Table A.26. Estimated Annual Economic Costs of Climate Change on Water Resources in Tarawa, 2050 (millions of 1998 US$) % change in Economic costs of substitution by Climate Change and Sea Level Rise Scenario groundwater alternative sources thickness Expansion into new Desalination groundwater sources - Baseline scenario (current mean sea level and rainfall; - average freshwater thickness = 12.1 m) Current MSL, 7% increase in rainfall +5.5 +0.2 +0.4 Current MSL, 10% reduced rainfall -14 -0.5 -1.0 0.2 meters MSL rise, current rainfall -0.9 -0.0 -0.1 0.4 meters MSL rise, current rainfall +2.0 +0.1 +0.1 0.4 meters MSL rise, 10% reduced rainfall -12 -0.4 -0.8 0.4 meters MSL rise, current rainfall, reduced island width -29 -1.0 -2.0 0.4 meters MSL rise, 7% increased rainfall, reduced island width -19 -0.7 -1.3 0.4 meters MSL rise, 10% reduced rainfall, reduced island width -38 -1.4 -2.7 Total costs (under a reduced rainfall scenario) -1.4 to -2.7 Total costs (under an increased rainfall scenario) -0.7 to -1.3 -0.7 to -2.7 Total costs Example: For a reduction of 38 percent in total supply, the costs of expanding into new groundwater sources are: 5,200 m3/day x 0.38 = 1976 m3/day x US$1.9/m3 = US$3,754 x 365 (to account for yearly costs) = US$1.4 million. The costs of expanding into desalinated water are: 5,200 m3/day x 0.38 = 1976 m3/day x US$3.7/m3 = US$7,311 x 365 (to account for yearly costs) = US$2.7 million. Several caveats need to be attached to the analysis. First, the conditions in Bonriki may not apply to other groundwater sources in Tarawa. Second, the costs of alternative rainwater collectors are likely to be lower than those of desalination. Third, if population pressures continue, there may be no additional groundwater sources onto which to expand supply. Background reports to this study, however, estimate that an additional 2,600 m3/day of groundwater supply might be obtained by reclaiming land in Temaiku Bight, on the southeastern part of the Tarawa atoll. Hence, the assumptions made appear defensible. Annex A Page 42 D. Impacts on Agriculture Viti Levu The impacts of climate change on the agriculture production of Viti Levu were estimated using the PLANTGRO model (Hackett 1988, 1991), which was incorporated into PACCLIM and was linked to the scenario generator and Fiji data. PLANTGRO allows the derivation of notional relationships between plant response (suitability) and different levels of 23 climate and soil factors. The outputs can be in the form of: ˇ Yield: relative yield in relation to potential maximum yield; ˇ Growing season length (for annual crops); ˇ Greatest limitation: the most critical limiting factor at a given point in time; ˇ Overall limitation rating: a composite index taking into account soil and climate conditions in each site. The PLANTGRO model was used to model the impacts of climate change on the production of yam, taro, and cassava. The study team attempted to use PLANTGRO to assess the suitability and yield of sugarcane under present conditions, but the model under-predicted yields in the western part of Viti Levu, where the industry is centered. Estimates of climate change impact for sugarcane were therefore derived based on historical estimates and future projections of climate variability. Impact of Climate Change and Variability on Sugarcane Production Sugarcane production is affected by both rainfall and by climate variability, especially El Niņo-induced droughts. The cumulative effects of a sequence of extreme events can be particularly damaging to sugarcane. The 1998 drought, which followed a sequence of natural disasters in Fiji, led to a nearly 50 percent reduction in sugarcane production estimated to cost US$64 million (Figure A.5 and table A.27): Figure A.5: Impacts of El Niņo- induced Droughts on Table A.27. Production of Sugarcane Sugarcane Production in Fiji (Cane Crushed) 1983-98 in Fiji, 1983-88 (`000 MT) 500 0 Year Production E l N ino-Re la te d D rou ghts (`000 MT) 450 0 1983 2202 400 0 1984 4290 1985 3042 350 0 1986 4109 )T 1987 2960 M 300 0 0 1988 3185 00'( 1989 4099 noitcudorP 250 0 1990 4016 200 0 1991 3380 1992 3533 150 0 1993 3703 1994 4063 100 0 1995 4110 1996 4379 50 0 1997 3280 1998 2100 0 83 48 58 86 87 88 9 90 19 2 93 49 59 96 97 89 19 19 19 19 19 19 198 19 19 199 19 19 19 19 19 19 Average 3528 Ye ar Annex A Page 43 The present expectation in Fiji is for an annual sugarcane production in the order of four million tonnes. As can be seen in Figure A.5, this level of production has been attained in only seven of the last 15 years. Of the other eight years, all but one (1985) were associated with El Niņo-droughts, which also tend to be associated with a greater frequency of tropical cyclones of hurricane force winds (Pahalad and Gawander 1999). Intermediary production years (such as 1993) reflect a slow recovery from the effects of droughts and cyclones. The analysis of physical and economic impact involved three major steps: Step 1: Estimate the Baseline Conditions. The baseline conditions reflect the impact of climate variability (droughts and cyclones) at present times. Hence, future production levels should not be compared to a normal year (such as 1990, with 4 million metric tons), but to an average year over a representative period of time. Taking 1983-19985 as a basis for the analysis, the average baseline production during this 16 year period was 3.53 million metric tons (table A.27). Step 2: Estimate Future Production under Climate Change and Climate Variability. Under future climate change conditions, the adverse effects of warmer temperatures (due to increased evapotranspiration and heat stress) may be offset by the possibility of higher rainfall. However, it is likely that the effects of bad years might be worsened by warmer and possibly drier conditions. The 1997/98 El Niņo drought, regarded as a 1 in 100 year event, may become more of the norm during El Niņo years. If this were the case, and based on the same frequency of droughts as observed in the 1983- 98 period (Figure A.5), the following production might be expected within the next 25 to 50 years: ˇ 25 percent of the years (4 out of 16 years) would have half the normal production: 2 million MT ˇ 31 percent of the years (5 out of 16 years) would have three-fourths of the normal production: 3 million MT ˇ 44 percent of the years (7 out of 16 years) would have normal production: 4 million MT. The frequency of droughts and `recovery' years is based on the historical records of Figure A.5. The weighted average production in a given year in the future is therefore: (0.25 x 2 million MT) + (0.31 x 3 million MT) x (0.44 x 4 million MT) = 3.2 million MT This represents a 9 percent drop relative to the average for the 1983-98 period (3.5 million MT) Step 3: Estimate the Economic Impact of Future Climate Change and Climate Variability. The 1998 price paid to sugarcane growers was F$81.79 per metric ton, or US$41.73 at the 1998 exchange rate of 1US$ = F$1.96. The average annual economic damages to surgarcane caused by future climate change and variability can then be computed as follows: (3.528 million MT x US$41.73/MT) - (3.2 million MT x US$41.73/MT) = US$147.2 - US$133.5 million = US$13.7 million It is assumed that all of the sugar production is harvested in Viti Levu. 5Readers may note that this period differs from that used as the analogue for cyclone impacts (1992-99). The reason is that droughts, occurring less frequently than cyclones, need a longer historical period on which to base the analysis. Also, 1983-98 was the period for which data on sugarcane was available. Annex A Page 44 Impact of Climate Change and Variability on Other Crops (Taro, Yam and Cassava) The PLANTGRO model in PACCLIM was used to model the impact of climate change and climate variability on taro, yams and cassava. The model was run for the following conditions: Baseline Conditions: Average year (1990) Current El Niņo: +0.5°C, -50% rainfall Current La Niņa -0.5°C, +50% rainfall 2050 Conditions: CSIRO9M2 Global Circulation Model and Best Guess Scenario +0.9°C, +5.7% rainfall CSIRO9M2 Global Circulation Model and Worst Case Scenario +1.3°C, +8.2% rainfall DKRZ Global Circulation Model and Best Guess Scenario +0.9°C, -5.7% rainfall DKRZ Global Circulation Model and Worst Case Scenario +1.3°C, -8.2% rainfall 2050 El Niņo +1.5°C, -60% rainfall 2050 La Niņa +0.5°C, +60% rainfall The 2050 El Niņo and La Niņa models include both average changes in climatic conditions, as well as in climate variability. Similar models were run for 2025 and 2100, but were not used in the final analysis. For each of the above models, the images produced were imported into an IDRISI Global Information System (Eastman 1985), and the land area in each yield class (0-5 metric tons/ha; 5-10 metric tons/ha; 10- 15 metric tons/ha) was calculated. This represents areas of suitability for the different yield classes, rather than actual area in production. However, the changes observed in the suitable land areas as climatic conditions are simulated allow for an estimate of the relative impacts of climate change on production yields (table A.28). Table A.28. Relative Impact of Climate Change and Climate Variability on Taro Yield Baseline (1990) 2050 Conditions 2050 Normal Year 1990 1990 1990 2050 2050 Normal El Niņo La Niņa CSIRO CSIRO DKRZ DKRZ El Niņo La Year Year Year Best Guess Worst Case Best Guess Worst Case Year Niņa Year 1.No. of hectares suitable for cultivation with yield of: 0-5 t/ha 2232.7 6456.3 877.0 2065.5 1992.0 2619.0 2766.5 8789.3 759.5 5-10 t/ha 4935.3 3477.0 5291.0 5127.3 5221.3 4723.3 4688.8 1720.3 5395.8 10-15 t/ha 3397.5 634.3 4397.5 3372.8 3352.3 3223.3 3110.3 56.0 4410.3 2. Percentage of hectares per land class: 0-5 t/ha 21% 61% 8% 20% 19% 25% 26% 83% 7% 5-10 t/ha 47% 33% 50% 49% 49% 45% 44% 16% 51% 10-15 t/ha 32% 6% 42% 32% 32% 31% 29% 1% 42% 3. Weighted average yield (in tons/hectare) 8.1 4.7 9.2 8.1 8.1 7.8 7.7 3.4 9.2 4.Change from baseline (relative to a 1990 normal year, in percentage) 0% -41.4% +13.8% +0.8% +1.1% -3.3% -4.8% -58.2% +14.6% Note: Shaded figures relate to the examples on page 45. Numbers may not add up due to rounding. Annex A Page 45 On the table above, the first set of estimates represents the number of hectares suitable for cultivation in each of the three land classes (0-5 tons/ha; 5-10 tons/ha; 10-15 tons/ha). This figure changes according with each of the 9 simulated scenarios, and is provided through PACCLIM and PLANTGRO. The second set of estimates represents the percentage of land suitable for cultivation in each of the three land classes. Example: The 1990 average baseline conditions project 3397.5 hectares of land suitable for cultivation with a yield of 10 to 15 metric tons per hectare. First, one computes the total number of hectares suitable for cultivation: 2,232.7 + 4,935.3 + 3,397.5 = 10,565.5 Then one computes the proportion of area suitable for cultivation having a yield of 10 to 15 metric tons per hectare: 3,397.5 hectares : 10,565.5 hectares = 32 percent. Hence, under this scenario, 32 percent of the area suitable for cultivation would have yields of 10-15 tons/ha. The third set of estimates represents the weighted average yield of each of the scenarios: Example: The 1990 average baseline conditions are distributed as follows: 0-5 tons/ha land: 2,232.7 hectares or 21% of the total 5-10 tons/ha land: 4,935.3 hectares or 47% of the total 10-15 tons/ha land: 3,397.5 hectares or 32 % of the total The weighted average yield uses the mid-value of each land type. Thus, for 0-5 tons/ha land, the mid-value would be 2.5 tons/ha; for 5-10 tons/ha land, the mid-value would be 7.5 tons/ha; etc. The weighted average uses these mid-values, multiplied by the percentage of each land class, is as follows: (2.5 tons/ha x 0.21) + (7.5 tons/ha x 0.47) + (12.5 tons/ha x 0.32) = 8.1 tons/ha. Finally, the fourth and last set of estimates computes how the average yield for each scenario differs from the 1990 baseline average: Example: The 1990 weighted average yield is 8.1 tons/ha. The yield during a 1990 El Niņo is only 4.7 tons/ha. Thus: 4.7 / 8.1 = 0.586. The yield during a 1990 El Niņo is only 58.6 percent that of a normal year (on average). This means that the average yield during an El Niņo event decreases by 1-0.586 = 0.414 or -41.4 percent. Tip: To estimate the percent decline or increase relative to the baseline, use [{scenario} / {baseline} - 1] x 100 to determine the percentage reduction or increase. For the 2050 El Niņo and La Niņa conditions, the relevant comparison is with the baseline El Niņo and La Niņa conditions: Example: The 1990 weighted average yield during an El Niņo year is 4.74 tons/ha. The estimated yield during a 2050 El Niņo year is 3.36 tons/ha (the numbers are rounded on table A.28). The incremental difference is: [(3.36 / 4.74) -1)] x 100 = 29.1 percent. Similarly than for taro, one can also estimate the change in yield of yam and cassava caused by the combined effects of climate change and climate variability (tables A.29 and A.30) Annex A Page 46 Table A.29. Relative Impact of Climate Change and Climate Variability on Yam Yield Baseline (1990) 2050 Conditions 1990 1990 1990 2050 Average Conditions 2050 2050 Average El Niņo La Niņa CSIRO CSIRO DKRZ DKRZ El Niņo La Niņa Conditions Year Year Best Guess Worst Case Best Guess Worst Case Year Year No. of hectares suitable for cultivation with yield of : 0-5 t/ha 1749.3 779.0 6677.5 2171.8 2389.3 1348.0 1228.0 873.3 7598.8 5-10 t/ha 3694.3 5420.5 1506.5 3458.8 3335.3 3887.8 3969.3 6229.5 1148.3 10-15 t/ha 5122.0 4366.0 2381.5 4935.0 4841.0 5329.8 5369.3 3462.8 1818.5 % hectares per land class: 0-5 t/ha 17% 7% 63% 21% 23% 13% 12% 8% 72% 5-10 t/ha 35% 51% 14% 33% 32% 37% 38% 59% 11% 10-15 t/ha 48% 41% 23% 47% 46% 50% 51% 33% 17% Weighted average yield (in tons/hectare) 9.1 9.2 5.5 8.8 8.7 9.4 9.5 8.7 4.8 4.Change from baseline (relative to a 1990 normal year, in percentage) 0% +1.1% -39.9% -3.2% -4.8% +3.2% +4.0% -4.1% -47.6% Note: Numbers may not add up due to rounding. Table A.30. Relative Impact of Climate Change and Climate Variability on Cassava Yield Baseline (1990) 2050 Conditions 1990 1990 1990 2050 Average Conditions 2050 2050 Average El Niņo La Niņa CSIRO CSIRO DKRZ DKRZ El Niņo La Niņa Conditions Year Year Best Guess Worst Case Best Guess Worst Case Year Year No. of hectares suitable for cultivation with yield of : 0-5 t/ha 728.0 736.5 2178.3 728.0 735.0 1110.0 1131.5 935.8 3847.5 5-10 t/ha 7400.8 7880.0 5955.3 8208.8 8479.3 7763.0 8051.8 8450.5 5108.3 10-15 t/ha 1929.8 1442.0 1925.0 1121.8 844.3 1185.5 875.3 672.3 1102.8 % hectares per land class: 0-5 t/ha 7% 7% 22% 7% 7% 11% 11% 9% 38% 5-10 t/ha 74% 78% 59% 82% 84% 77% 80% 84% 51% 10-15 t/ha 19% 14% 19% 11% 8% 12% 9% 7% 11% Weighted average yield (in tons/hectare) 8.1 7.9 7.4 7.7 7.6 7.5 7.4 7.4 6.1 4.Change from baseline (relative to a 1990 normal year, in percentage) 0% -3.0% -8.9% -5.0% -6.7% -6.9% -9.0% -9.0% -24.2% Note: Numbers may not add up due to rounding. Annex A Page 47 In order to estimate the economic impact of climate change and variability on food crops, the following five steps were followed: Step 1: Compute Baseline Production for Viti Levu. Tables A.28-30 are useful to estimate the percentage change in average yields caused by the different climatic conditions. However, the `suitability' areas do not reflect actual production. To estimate the baseline production in Viti Levu, one needs to use actual production statistics. FAO (1996) provides the following estimates of production in 1990, an `average' year in Viti Levu (table A.31). Table A.31. Production of Major Fijian Food Crops (1990) Land Area Yield Production Price Value Crop (hectares) (tons/ha) (thousand tons) (F$/ton) (millions F$) Taro 4,632 1.9 8.8 230 2.0 Yam 1,099 7.1 7.8 500 3.9 Cassava 4,352 5.4 23.5 220 5.2 Source: Statistics on yield, production, and price come from the FAOSTAT database (FAO 1996) for the base year 1990. Note: Acreage and value are derived from these basic statistics. Tons represent metric tons. The figures above are for Fiji as a whole. To estimate the production for Viti Levu, the production figures need to be adjusted by the proportion of total land that Viti Levu represents (58 percent), assuming relative uniform yields across Fiji: Crop Production Adjustment for Viti Levu Total Production Viti Levu (thousands tons) (thousand tons) Taro 8,800 x 0.58 5,104 Yam 7,800 x 0.58 4,524 Cassava 23,500 x 0.58 13,630 Step 2: Adjust the Value of Production to 1998. To convert the 1990 prices to 1998 values, one can use the consumer price index (CPI) for Fiji (MFNP 1999): Year CPI (% change) Taro Yam Cassava (F$/ton) (F$/ton) (F$/ton) 1990 baseline 0.0 230.0 500.0 220.0 1991 6.5 245.0 532.5 234.3 1992 4.9 257.0 558.6 245.8 1993 5.2 270.3 587.6 258.6 1994 0.6 271.9 591.2 260.1 1995 2.2 277.9 604.2 265.8 1996 2.4 284.6 618.7 272.2 1997 2.9 292.8 636.6 280.1 1998 8.1 316.6 688.2 302.8 (see page 4 for an explanation of inflation adjustment) Annex A Page 48 The 1998 baseline production value of the three crops in Viti Levu is therefore: Crop Production Price Production Value (metric tons) (1998 F$/ton) (1998 F$ thousands) Taro 5,104 316.6 1,616 Yam 4,524 688.2 3,113 Cassava 13,630 302.8 4,127 Or, converting into 1998 US$ (at an exchange rate of 1 US$ = F$1.96): Crop Production Value (1998 US$ thousands) Taro 824 Yam 1,588 Cassava 2,106 Step 3: Compute Baseline Production taking into Account Climate Variability. As for sugar, it would be incorrect to compare future climatic conditions with an average production year, as conditions such as El Niņo or La Niņa play a strong role in the present-day production of food crops. Hence, it is necessary to compute a baseline that takes present-day climate variability into account. This can be done by observing the sequence of El Niņo/La Niņa events in Fiji (see Figure A.5). While La Niņa does not play a vital role in sugarcane, it is quite important for food crops, in particular for yam and cassava. Hence, the occurrence of La Niņa years needs also to be taken into account. During the 16 year period of 1983-98, there were 7 years that could be considered `normal', 4 El Niņo years, 2 La Niņa years, and 3 `intermediary' years. Since El Niņo years are the most important for taro, while La Niņa plays a stronger role in the production of yam and cassava, the intermediary years are taken to involve levels of production between a `normal' and a `El Niņo' year for taro, and between a `normal' and a `La Niņa' year for yam and cassava. An occurrence of once every 4 years for El Niņo is equivalent to a probability of occurrence of 25 percent in any given year (4 : 16). Similarly, an occurrence of 2 in 16 for La Niņa is equivalent to a probability of nearly 15 percent.6 Normal years occur with a probability of 40 percent, and intermediary years with a probability of 20 percent. Hence, for all three food crops, the probability of occurrence is as follows: Baseline Conditions Taro Yam Cassava Probability of occurrence El Niņo1990 25% 25% 25% La Niņa 1990 15% 15% 15% Normal Year 1990 40% 40% 40% Intermediary Year (El Niņo to normal) 20% Intermediary Year (La Niņa to normal) 20% 20% 615 percent was used, rather than the more technically correct 12.5 percent, because it is debatable whether Fiji experienced 2 or 3 La Niņas during the 1983-98 period. Annex A Page 49 If it is recalled, the above baseline conditions result in the following changes in average yield from a normal year (see tables A.28 to A.30): Baseline Conditions Taro Yam Cassava Changes in yield from the baseline (in percentage) El Niņo 1990 -41.4% +1.1% -3.0% La Niņa 1990 +13.8% -39.9% -8.9% Normal Year 1990 0% 0% 0% Intermediary year (El Niņo to normal)1 -20.7% Intermediary year (La Niņa to normal)1 -20.0% -4.5% 1- average of a normal year and an El Niņo year (or a La Niņa year) It is then relatively simple to compute a weighted average of the two tables above to derive the variability coefficient: Baseline Conditions Taro Yam Cassava Probability of Occurrence Change in yield from the baseline (in percentage) El Niņo 1990 25% -41.4% +1.1% -3.0% La Niņa 1990 15% +13.8% -39.9% -8.9% Normal Year 1990 40% 0% 0% 0% Intermediary year (El Niņo to normal) 20% -20.7% Intermediary year (La Niņa to normal) 20% -20.0% -4.5% Variability Coefficient -12.4% -9.7% -3.0% Example: The variability coefficient for taro is: (0.25 x -41.4%) + (0.15 x +13.8%) + (0.4 x 0%) + (0.2 x -20.7%) = -10.35 + 2.07 + 0 - 4.14 = -12.4% One can then derive a baseline production which takes into account climate variability: Table A. 32. Baseline Production (1990) Taking Into Account Climate Variability Crop Normal Production Variability Coefficient Baseline Production w/ Production Value w/ (metric tons) (% change in yield due climate variability climate variability to climate variability) (metric tons) (thousands of 1998 US$) Taro 5,104 -12.4 % 4,524 722 Yam 4,524 -9.7% 4,085 1,434 Cassava 13,630 -3.0% 13,630 2,043 Annex A Page 50 Example: For yam, the baseline production value (taking into account climate variability) is: 4,524 tons x (1-0.097) = 4,085 tons (average production (variability (average production) in a normal year in coefficient) taking into account Viti Levu) climate variability) 4,085 x F$688.2 = F$2,811,300 (average production (1998 price (total value w/ variability) per ton) of production w/ variability) F$2,811,300 / 1.96 = US$1,44,335 (total value of production (exchange rate (total value of production in in Fijian dollars to US$ in 1998) 1998 US dollars, taking w/ variability) variability into account) Step 4: Compute Impact of Climate Change on Average Conditions. The impact of climate change on average conditions is simply the difference in yield and production value of an average year in 2050 relative to an average year in 1990 (table A.23). In order to account for uncertainty, a range of predicted yield changes was constructed by taking into account all four scenarios of tables A.28 to A.30 (CSIRO Best Guess and Worst Case Scenarios, and DKRZ Best Guess and Worst Case Scenarios). Table A. 33. Impact of Climate Change (Rainfall and Temperature) on Average Climatic Conditions, 2050 Crop Average Baseline Conditions (in 1990) Impact of Climate Change on Average Climatic Conditions (2050) Average Production Production Value Change in Average Yields Costs of Climate Change (metric tons) (in 1998 US$ thousands) (in 1998 US$ thousands) Talo 5,104 824 -4.8% to +1.1% -40 to + 9 Yam 4,524 1,580 -4.8% to +4.0% -76 to + 63 Cassava 13,630 2,101 -9.0% to -5.0% -189 to -105 Notes: Minus signs denote costs of climate change. Plus signs denote a benefit if scenarios of higher rainfall (as predicted by the DKRZ model) materialize. Example: For cassava, the impact of climate change on average climatic conditions is as follows (table A.30): The predicted change in average yields under the CSIRO Best Guess scenario is -5.0% The predicted change in average yields under the CSIRO Worst Case scenario is -6.7% The predicted change in average yields under the DKRZ Best Guess scenario is -6.9% The predicted change in average yields under the DKRZ Worst Case scenario is -9.0% Hence, the range of predicted changes is -9.0% to - 5.0%. Applying these changes to the baseline production value: -0.09 x US$2,101,000 = - US$ 189,090 -0.05 x US$2,101,000 = - US$ 105,050 The cost of changes in average climatic conditions in 2050 ranges from US$189,000 to $105,000. Annex A Page 51 Step 5: Compute the Impact of Climate Change and Climate Variability. To estimate the impact of climate change and climate variability in 2050, one can use a method similar to that of Step 3, by computing variability coefficients for the conditions likely to prevail in 2050 relative to a normal year in 1990. Tables A.28 to A. 30 provide the expected changes in yields of 2050 conditions relative to the 1990 baseline: Taro Yam Cassava Changes in Yield from the 1990 Baseline (in percentage) El Niņo 2050 -58.2% -4.1% -9.0% La Niņa 2050 +14.6% -47.6% -24.2% Normal Year 2050 -4.8% to +1.1% -4.8% to +4.0% -9.0% to -5.0% To compute the variability coefficients, one needs to apply the weights considered in Step 3. This assumes no change in the frequency of El Niņo and La Niņa events in the future: Taro Yam Cassava Probability of Change in yield from the baseline (in percentage) Occurrence El Niņo 2050 25% -58.2% -4.1% -9.0% La Niņa 2050 15% +14.6% -47.6% -24.2% Normal Year 2050 40% -4.8% to +1.1% -4.8% to +4.0% -9.0% to -5.0% Intermediary year (El Niņo to normal) 20% -58.2% to +1.1% Intermediary year (La Niņa to normal) 20% -47.6% to +4.0% -24.2% to -5.0% "Intermediary years" are assigned the range of values between a normal and abnormal year. Hence, for taro, the change in yield during an intermediary year could range from -58.2% (the change during an El Niņo year) to +1.1% (the change in average climatic conditions). The variability coefficients for 2050 conditions are therefore: Taro Yam Cassava Variability coefficients -25.9% to -11.7% -19.6% to -5.8% -14.3% to -8.9% Example: The variability coefficient for yam is: (0.25 x -4.1%) + (0.15 x -47.6%) + (0.4 x -4.8%) + (0.2 x -47.67%) = -19.6% for the lower bound and (0.25 x -4.1%) + (0.15 x -47.6%) + (0.4 x +4.0) + (0.2 x +4.0%) = -5.8% for the upper bound One can then apply these variability coefficients to the normal production in 1990 to derive the expected production in 2050 under both climate change and climate variability (table A.34). Annex A Page 52 Table A. 34. Expected Production in 2050 with Climate Change and Climate Variability Crop Average Baseline Conditions Variability Expected Conditions in 2050 due to Climate Change (in 1990) Coefficient in 2050 and Climate Variability Production Value Production Value (metric tons) (in 1998 US$ (metric tons) (in 1998 US$ thousands) thousands) Talo 5,104 824 -25.9% to -11.7% 3,782 to 4,507 611 to 728 Yam 4,524 1,580 -19.6% to -5.8% 3,637 to 4,262 1,270 to 1,488 Cassava 13,630 2,101 -14.3% to -8.9% 11,681 to 12,417 1,800 to 1,914 Example: The expected production of cassava in 2050 is as follows: The variability coefficient of -14.3% to -8.9% indicates the proportion by which the 1990 normal yield is reduced under the future conditions. 13,630 tons x (1-0.143) = 11,681 (for the lower bound) 13,630 tons x (1-0.089) = 12,417 (for the upper bound) Table A.34 shows the expected production and value under future conditions of climate change (rainfall and temperature) and variability (El Niņo/La Niņa conditions). It does not indicate, however, the costs of these changes. To do this, one needs to compute the difference between the 2050 conditions with climate variability and the 1990 baseline conditions (also with climate variability). This is shown on table A.35 below. Table A. 35. Impact of Climate Change (Rainfall and Temperature) and Climate Variability (ENSO) on Food Crops, 2050 Crop 1990 Baseline Conditions 2050 Conditions with Climate 2050 Impacts of Climate Change and with Climate Variability Change and Variability Variability (2050) Production Value Production Value Changes in Yield (in Economic Costs (metric tons) (in 1998 US$ (metric tons) (in 1998 US$ percentage) (in 1998 US$ thousands) thousands thousands) Taro 4,471 722 3,782 to 4,507 611 to 728 -15.4% to +0.8% -111 to +6 Yam 4,085 1,434 3,637 to 4,262 1,270 to 1,488 -11.0% to +4.3% -164 to +54 Cassava 13,221 2,042 11,681 to 12,417 1,800 to 1,914 -11.7% to -6.1% -242 to -128 Example: For Taro, the production in 2050 is 3,782 to 4,507 metric tons. The production under the 1990 baseline is 4,471 metric tons. Hence, the change in yield is as follows: [(3,782 / 4,471) -1 ] x 100 = -15.4 percent (for the lower bound) and [4,507 / 4,471) - 1 ] x 100 = +0.8 percent (for the upper bound) The economic costs of climate change and variability are therefore: 611 - 722 = -US$111,000 (for the lower bound) and 728 - 722 = + US$6,000 (for the upper bound) The upper bound represents a net economic gain under a scenario of rainfall increase. The lower bound represents a net economic loss under a scenario of rainfall decrease. The overall impact of climate change on the agriculture sector of Viti Levu is summarized in table A.36. Annex A Page 53 Table A.36. Estimated Economic Impact of Climate on Change on Agriculture in Viti Levu, Fiji, 2050 (thousands of 1998 US$) Impact of change in average rainfall Impact of change in rainfall, temperature, and temperature and climate variability (ENSO) Current Change in Change in production Economic Impact average yield Economic Impact average yield Crop (US$ thousands) (US$ thousands) (percent) (US$ thousands) (percent) Sugarcane 147.2 -- -- -13,700 -9 Dalo (Taro) 800 -40 ­ +9 -5 ­ +1 -111 ­ +6 -15 ­ +1 Yam 1,600 -76 ­ +63 -5 ­ +4 -164 ­ +54 -11 ­ +4 Cassava 2,100 -189 ­ -105 -9 ­ -5 -242 ­ -128 -12 ­ -6 Total -13,800 ­ 14,200 -- Not available. Minus signs indicate an economic cost. Plus signs indicate an economic benefit (from rainfall increases). Note: Ranges reflect best-guess and worst­case scenarios under two different climate change models. Source: Background studies to this report. Tarawa The physical and economic impact of climate change on the agriculture of the Tarawa atoll could not be computed quantitatively due to lack of statistical data. However, some qualitative judgments can be made. Climate change is most likely to impact agricultural crops through changes in rainfall. Coconuts, for example, require an annual rainfall of 1,000-1,500 millimeters or more. Copra production is closely related to rainfall, albeit with a two year lag (Figure A.6). Te babai (the giant taro) is also extremely sensitive to reductions in groundwater, and prone to saltwater intrusion as a result of storm surge and overwash. If wetter conditions prevail in the future, production of water sensitive crops -- coconut, breadfruit and te babai -- is Figure A.6: Variation in Copra Production with Rainfall in likely to increase. If rainfall decreases as Kiribati a result of climate change, copra and te babai 14000 400 production would be adversely affected. Of Copra production equal interest is the likely effect of climate Rainfall 12000 350 variability on agricultural crops. Droughts are most likely to occur during La Niņa 300 10000 years. An intensification and/or increase in )t 250 ) frequency of La Niņa events, coupled with M( 8000 m higher average temperatures, could have onit 200 m(lla significant negative effects on the major oducrP 6000 infaR crops. 150 4000 Sea level rise could affect agriculture crops 100 in two ways. The first is through salt-water 2000 50 intrusion, which would affect te babai production in particular. The second is 0 0 through loss of coastal land due to 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Year inundation, which could reduce the production of copra, breadfruit or pandanus. Note: Rainfall is offset by two years relative to copra production Annex A Page 54 E. Impacts on Health Viti Levu Climate is one of the important determinants of human health. Average climate conditions, climate variability and climate extremes influence human health in Fiji either through direct mechanisms (such as flooding), indirect mechanisms (such as the distribution of vector-borne diseases) or cumulative mechanisms (as exemplified by the effects of a cyclone or drought on the national economy and living standards). The potential impacts of climate change on human health in Viti Levu were analyzed quantitatively for dengue fever and diarrheal disease. These illnesses have a clear climate linkage and are amenable to quantitative analysis. Other effects such as nutrition deficiencies caused by more intense cyclones were taken into account in the water resources analysis (though the impact of droughts on nutrition could not be quantified). The potential impact of more intense cyclones on human fatalities can be estimated, but it involves controversial estimates of the value of a statistical life. This method of valuation is based on the monetary value individuals place on lowering their mortality risks. With estimates of this type only available in developing countries, extrapolations to the Fiji conditions would require comparing the value of a statistical life in Fiji with that of the United States (for example) a procedure which remains highly controversial. While this method is illustrated here in the economic analysis of dengue fever, it was not applied to cyclone fatalities as this would have required estimating average fatality rates, an estimate which is highly uncertain (as it depends on individual cyclone paths and intensity). The impact of climate change on malaria, while important, does not appear to be a substantial threat in Fiji due to the absence of the vector. Filiarisis, another vector-borne disease, is expected to be eradicated within 5-10 years. While changes in sea surface temperatures are believed to influence, in some sites, the incidence of ciguatera (an illness caused by ingestion of fish contaminated by ciguatoxins), such relation was not found in Fiji (Hales et al. 1999). Impacts of Climate Change on Dengue Fever Figure A.7. Dengue Fever Transmission Cycle Dengue fever, a mosquito-borne viral disease, is a significant and potentially Immunity increasing public health problem in Fiji. The most recent epidemic -- in 1998 -- Susceptibles Infected involved an estimated 24,000 cases and Human population resulted in 13 deaths (Basu et al, 1999; WHO, 1998). Changes in rainfall, ambient Mosquito bite Mosquito bite temperature and humidity influence the life- cycle of the mosquito vectors and are powerful determinants of the distribution and size of vector populations. Rising Mosquito population temperatures also increase the epidemic risk Extrinsic by increasing the biting rate of the mosquito incubation period vector and the replication rate of the dengue virus (Figure A.7). Infectious Susceptible mosquitoes mosquitoes Annex A Page 55 The model used in PACCLIM to assess the impact of climate change on dengue fever was based on the CIMSIN and DENSIM models relating epidemic potential--an index reflecting the efficiency of transmission--to ambient temperature (Patz et al. 1998). This relationship was the basis of the simple biophysical index model used in PACCLIM, which describes relative changes in dengue fever risk resulting from changes in ambient temperature as projected by the scenarios used in the study. Rainfall was not considered in this model because the main dengue fever vector, Aedes aegypti, is adapted to the urban and domestic environment where breeding sites may be created by artificial water containers. These breeding sites may be equally available under low rainfall conditions. Since both of the general circulation models used by this study project similar changes in temperature (see page A.2) and the health impacts are modeled by temperature alone, only the CSIRO9M2 global circulation model was used for scenario construction. The scenarios took into consideration both a `best guess' emissions scenario as well as a `worst case' scenario (as outlined in table A.1). Present monthly patterns of epidemic potential in Nadi and Suva and spatial characteristics of epidemic potential in Viti Levu were entered into the model, and analyzed under future climatic conditions in 2025, 2050, and 2100. The major factor affecting epidemic potential in this model is the extrinsic incubation periodthe time taken for viral replication in the host mosquito. The model could not account for the range of climatic, environmental and ecological factors affecting mosquito abundance, or distribution and effects specific to vector species other than Aedes aegypti. Changes in the frequency and severity of extreme events and social aspects of transmission dynamics such as human population density, travel, immunity and housing conditions which may influence disease transmission (Hales et al. 1999) were also not captured under the model. Nonetheless, the dengue fever model in PACCLIM is the best semi-quantitative tool available to assess the possible impacts of climate change on human health in Fiji. In order to analyze the impact of climate change on epidemic potential, different areas of Viti Levu were classified according to their epidemic risk. Five categories of epidemic risk were constructed (table A.37). The risk was considered to be low where the model-predicted epidemic potential was less than 0.1. An area with an epidemic potential of 0.3 or higher was considered to be at extreme risk of an epidemic. Increases in epidemic risk were then evaluated for different areas of Viti Levu. PACCLIM produces both Geographical Information System simulations as well as quantitative projections of the epidemic distribution risk. Monthly variations in epidemic potential were also included in the model in order to estimate changes in the seasonality of dengue fever. Annual averages employed to estimate geographical and time-based impacts. Table A. 37. Risk Categories of a Dengue Fever Epidemic Based on Model- predicted Epidemic Potential in Viti Levu, Fiji Category Epidemic Potential Category description 0 less than 0.01 (or not a land area) No risk 1 0.01 to just less than 0.1 Low risk 2 0.1 to just less than 0.2 Moderate risk 3 0.2 to just less than 0.3 High risk 4 0.3 to 1.0 Extreme risk Note: Epidemic potential is an index that reflects the efficiency of transmission in a particular area. The analysis of climate change impacts on dengue fever involved five major steps: Annex A Page 56 Step 1. Estimate the Baseline Conditions. Under the present conditions (1990 baseline), 53 percent of the area of Viti Levu is considered to be at a low risk of a dengue fever epidemic. About 47 percent is at a moderate risk of an epidemic. Dengue fever epidemic potential is highest in the coastal areas and decreases towards the highlands. Epidemic potentials in the west, especially in coastal and hill land areas, are higher than in the central area. Also, the coastal areas in the north and west of the island have a higher epidemic potential than the southern and eastern coasts, especially near Nadi, Lautoka and western part of the Coral Coast. Step 2. Estimate the Impact of Climate Change. Figure A.8 and table A.38 summarize the impacts of climate change predicted by the model: ˇ Changes in Epidemic Potential and Seasonality in Nadi and Suva. By 2050, the epidemic potential is projected to increase by 20-30 percent in Nadi and Suva. By 2100, the epidemic potential could increase by 40-100 percent. Dengue fever epidemics, which now occur seasonally from November to April and with a frequency of once every 10 years could become endemic, occurring all year round and with increasing frequency by the end of the century. Figure A.8. Projected Changes in Epidemic Risk of Dengue Fever in Viti Levu (2025, 2050, (Best Guess Scenario) (Worst Case Scenario) 100% 100% 80% 80% 60% 60% High risk Extreme risk 40% Medium risk 40% High risk Medium risk Low risk Low risk 20% 20% 0% 0% 1990 2025 2050 2100 1990 2025 2050 2100 Table A. 38. Potential Impact of Climate Change on Dengue Fever in Viti Levu, Fiji Likely Changes Baseline 2025 2050 2100 (1990) Estimated Change Epidemic Potential 0% 10% 20-30% 40-100% Epidemic Potential in Viti Levu: Low 53% 38-39% 25-31% 7-21% Medium 47% 61-62% 69-72% 48-72% High -- -- 0-3% 7-41% Extreme -- -- -- 0-4% Seasonality Nadi Seasonal Seasonal Seasonal to All Year All year Suva Seasonal Seasonal Seasonal to Prologued season Extended Season to all year Frequency of epidemics 1 in 10 years Likely Increase Severity of strains Likely Increase Note: Ranges represent the best guess and worst case scenarios. Only the CSIRO GCM model was used here. Annex A Page 57 ˇ Spatial Changes. Spatial changes in epidemic potential can be analyzed using mapping functions in PACCLIM. The results are summarized in Figure A.8. By 2050, some 69-72 percent of Viti Levu is expected to be at moderate risk of an epidemic (a 50 percent increase over 1990). By 2100, under a high scenario, nearly half of Viti Levu (45 percent) is projected to be at high or extreme risk. At greatest risk will be the coastal areas of the Western Division. ˇ Size of epidemics. With an increase in epidemic potential, exacerbated by rising population densities and urbanization, future epidemics are expected to escalate more easily and involve a greater number of people than currently observed. Climate change may also have direct effects on vector abundance. ˇ Likelihood of Severe Strains. The increase in the frequency and size of the epidemics and shift to endemicity could inrease the risks of more severe forms of dengue fever, such as dengue shock syndrome (DSS) and dengue haemorrhagic fever (DHF). This could increase the number of fatalities relative to what is observed today. Step 3: Estimate the Change in Number of Dengue Fever Cases. In order to calculate the economic costs of climate change on dengue fever, it is necessary to estimate the possible increase in the number of cases attributable to climate change. Several factors make this estimate particularly challenging: (a) dengue fever epidemics rely on complex interactions between the mosquito biology (such as breeding sites, predators, competition), and socio-economic factors (such as the living environment of villages and towns); (b) changes in epidemic potential may not necessarily correlate with changes in the number of cases; and (c) it is difficult to predict the improvements in medical services, treatments and possible vaccine development that may occur in the future. For the purposes of the economic analysis, however, it is assumed that the percentage change in dengue fever epidemic potential is indicative of the possible changes in the number of cases averaged out over a 10-20 year period. The estimated change in the number of cases in Viti Levu is therefore: Estimated Changes in the Number of Dengue Fever Caused by Climate Change Baseline (1990) 2025 2050 2100 -- 10% 20-30% 40-100% where the ranges indicate the best guess (lower bound) and worst case scenarios (upper bound). Step 4: Estimate the Unit Cost of an Epidemic in the Baseline Scenario. The most recent dengue fever epidemic in Fiji, in 1997/98, was the worst in Fiji's history, causing an estimated 24,000 cases of illness, 1,700 hospital admissions and 13 deaths. The epidemic occurred during a severe drought related to the 1997/98 El Niņo event, and cost an estimated F$6.5 million (Koroivueta, personal communication, 1999). This estimate accounts only for hospitalization and personnel costs, treatment and medication, intra- venous fluid costs and laboratory services. Apart from the possible loss of human life, dengue fever epidemics involve many direct and indirect economic costs (Basu et al, 1999): ˇ Costs of hospitalization, medical treatment and laboratory services; ˇ Loss of productivity due to illness; ˇ Emergency vector control costs; ˇ Loss in tourism revenues; ˇ Costs incurred at the household level in coping with ill family members. Annex A Page 58 A typical epidemic in the future may involve the following cost breakdown (modified from NZMoH 1996): Table A.39. Potential Impact of a Dengue Fever Epidemic in the Future Cost Items Potential Impact Nature of epidemic 1000 cases of classical dengue 10 cases of Dengue Haemorrhagic Fever 1 death Medical costs 100 hospitalizations for 1 week 10 intensive care cases for one week (with a further 2 weeks of hospitalization) Time off worka 500 for 1 week 400 for 2 weeks 100 for 3 weeks 9 for 4 weeks (1 death) Note: aAssumes all cases are working adults. The economic analysis used the cost breakdown of table A.39. However, the total number of cases considered in the baseline estimate was based on the historical evidence of the 1989/90 epidemic (3,700 cases) and the 1997/98 epidemic (24,000 cases). This range (3,700-24,000 cases per epidemic) was considered to best represent the baseline conditions of the 1990s. Based on the available information, four components of epidemic costs were quantified: ˇ Medical costs ˇ Loss of productivity due to illness ˇ Willingness to pay to avoid illness ˇ Willingness to pay to avoid premature fatality 4.1. Medical Costs: The medical costs of dengue fever epidemics can be computed from historical records. Using the costs of the 1997/98 epidemic as a basis (F$6.5 million), the medical costs per case are as follows: F$6.5 million : 24,000 = F$271 or US$138 per case (no. of cases of the 1997/98 epidemic) The baseline conditions are assumed to involve 3,700 to 24,000 cases per epidemic. Hence, US$138 per case x 3,700 cases = US $511,580 US$138 per case x 24,000 cases = US$3,316,300 The medical costs per epidemic (for the baseline conditions) are therefore US$0.5-$3.3 million. Annex A Page 59 4.2. Loss of Productivity due to Illness. The loss of productivity due to illness was computed based on the `time off work' estimates of table A.39. Assuming each week to have 5 working days, the number of working days lost to illness is as follows: 500 cases x 5 working days (1 week) = 2,500 working days lost 400 cases x 10 working days (2 weeks) = 4,000 working days lost 100 cases x 15 working days (3 weeks) = 1,500 working days lost 9 cases x 20 working days (4 weeks) = 160 working days lost Total working days lost = 8,180 To determine the average number of working days lost per case of dengue fever, the above number is divided by the number of cases in table A.39 (1,011): 8,180 : 1,011 = 8.09 working days per case Multiplying this by the number of cases under the baseline conditions (3,700-24,000): 8.09 x 3,700 = 29,937 8.09 x 24,000 = 194,184 or approximately 29,940-194,180 working days lost per epidemic The economic value of a working day can be approximated by the Gross Domestic Product (GDP) per capita: Fiji's GDP in 1998 (US$1,383 million) / Population of Fiji (772,655 in 1998) = US$1,790 This value can then be divided by the number of working days a year (5 working days x 52 weeks = 260 working days) to yield the economic value of a working day: US$1,790 / 260 = US$6.88 per day of lost work This analysis involves several broad assumptions. Not all Fijians work, for example. If this had been considered in the analysis, the value of a working dayfor the proportion of the population that workswould be higher than US$6.88. At the same time, not all Fijians affected by dengue fever are workers. These two factors are assumed to cancel themselves out: in other words, the value of a day of lost work is assumed to be the average of a working day for workers, and a day of inactivity for nonworking Fijians. The lost productivity due to illness is: 29,940 working days lost x US$6.88 per day of lost work = US$206,000 (for the lower bound) 194,184 working days lost x US$6.88 per day of lost work = US$1,336,000 (for the upper bound) The lost productivity due to illness for the baseline conditions is therefore valued at US$0.2 to US$1.3 million per epidemic. Annex A Page 60 4.3. Willingness to Pay to Avoid Illness. The willingness to pay to avoid illness is the welfare that dengue fever sufferers lose due to pain, suffering, and disruption to their normal life routine. Technically, it is the value that individuals would be willing to pay to avoid a day of restricted activity due to dengue fever. This type of estimates of willingness to pay require surveys of `contingent valuation' that are not available in Fiji. For lack of a better measure, the value of avoiding a restricted activity day was derived from average estimates in the United States (US$50 per person). The applicability of this estimate to Fiji needs to be considered in the context of differentials in per capita income, as well as of differences in how disease is perceived and valued in the local culture. To make a rough extrapolation to the Fijian context, the value of a restricted activity day in the United States (US$50 per person) was first converted to a percentage of the per capita GDP of the United States (US$22,000). US$50 / US$22,000 = 0.227 percent of per capita GDP This was then applied to the per capita GDP for Fiji, US$1,790 to obtain the adjusted value of a restricted activity day in Fiji: 0.227 percent x US$1,790 = US$4.07 per restricted activity day. The number of restricted activity days can be determined from table A.39. However, since individuals can value a restricted activity day independently to whether it is a working or a weekend day, one week of illness is assumed to correspond to 7 days of restricted activity: 500 x 7 restricted activity days (1 week) = 3,500 400 x 14 restricted activity days (2 weeks) = 5,600 100 x 21 restricted activity days (3 weeks) = 2,100 9 x 28 restricted activity days (4 weeks) = 252 Total restricted activity days = 11,452 This number can then be divided by the number of cases in table A.39, and adjusted by the number of cases of the 1989/90 and 1987/98 epidemics: 11,452 / 1,011 = 11.33 11.33 x 3,700 = 41,911 11.33 x 24,000 = 271,858 or 41,910 to 271,860 days of restricted activity Multiplying this by the estimated value of a restricted activity day yields the following results: US$4.07 x 41,910 = US$170,600 US$4.07 x 271,860 = US$1,106,500 The willingness to pay to avoid illness for the baseline conditions is therefore valued at US$0.17 to US$1.1 million per epidemic. Annex A Page 61 4.4. Willingness to Pay to Avoid Premature Fatality. The willingness to pay to avoid premature fatality can be assessed through what is commonly called the value of a statistical life. As stated before, the extrapolation of this value across different countries and cultures remains highly controversial and should be interpreted with extreme caution. Based predominantly on premature fatality valuation studies in the United States, the generally accepted range for a value of a statistical life ranges from US$2 million to US$16 million, with US$5.8 million the average value currently used by the United States Environmental Protection Agency (EPA 1999). Similarly to the procedure illustrated above, the value of a statistical life can be computed as a percentage of the per capita GDP: US$5.8 million / US$22,000 = 26,364 percent of per capita GDP Applying this percentage to the per capita GDP for Fiji (US$1,790): 26,364 percent x US$1,790 = US$417,900 According to table A.39, an epidemic affecting 1,011 people is likely to result in one fatality. Extrapolating this fatality rate to the baseline case: 1 / 1,011 = 0.00099 0.00099 x 3,700 = 3.7 0.00099 x 24,000 = 23.7 Thus, the fatality rate in the baseline case is assumed to range from 3.7 to 23.7 deaths. Using the estimated willingness to pay to avoid premature fatality: US$417,900 x 3.7 = US$1,546,200 US$417,900 x 23.7 = US$9,904,230 The estimated willingness to pay to avoid premature fatality in the baseline case is therefore estimated at US$1.5 to US$9.9 million. 4.5. Total Baseline Costs of a Dengue Fever Epidemic in Fiji. Adding the estimates of sections 4.1 to 4.2 together gives the following baseline costs for a dengue fever epidemic in Fiji: US$ million a. Medical costs 0.5-3.3 b. Lost Productivity due to Illness 0.2-1.3 c. Willingness to Pay to Avoid Illness 0.2-1.1 d. Willingness to Pay to Avoid Premature Fatality 1.5-9.9 Total US$2.4-15.6 million per epidemic Since the costs were based on 1998 estimates of GDP and medical costs, the baseline costs are expressed in 1998 US dollars. This estimate illustrates the potential value of loss of lives compared to other medical costs of a dengue fever epidemic. Insofar as fatalities caused by cyclones were not computed, it can be seen that the real costs of climate change on public health are likely to be substantially higher than the estimates of this analysis indicate. Annex A Page 62 Step 5: Estimate the Potential Impacts of Climate Change. To assess the potential impact of climate change on future dengue fever epidemics, two figures are needed: (i) the estimated increase in number of cases predicted by the climate change models; and (ii) the forecasted increase in population. These can be derived from tables A.4 and A38: Baseline (mid-1990s)a 2025 2050 2100 Population 772,655 +44-57% +63-110% +66-198% Number of cases/epidemic 3,700-24,000 +10% +20-30% +40-100% Note: a. Population baseline is based on 1996 census. The baseline number of cases is based on the 1989/90 and 1997/98 dengue fever epidemics. Example: From table A.4, The projected population at the end of the century is 1,280,000 (in a low population growth scenario) to 2,300,000 (in a high population growth scenario). This is equivalent to an increase of 66 and 198 percent relative to the baseline population: (1,280,000 / 772,655) -1 = 1.66 - 1 = 66 percent (2,300,000 / 772,655) - 1 = 2.98 -1 = 198 percent These growth rates can then be applied to the baseline costs (US$2.4-$15.6 million) to derive the costs of dengue fever epidemics in the future. In order to derive the average annual cost of climate change in Viti Levu, the following calculations need to be made (see table A.40): ˇ Determine the average costs of future epidemics ˇ Assess the incremental costs that can be attributed to climate change ˇ Transform these incremental costs into annual averages ˇ Compute the equivalent costs for Viti Levu 5.1. Determine the Cost per Epidemic in the Future. To determine the costs of future epidemics (expressed in 1998 dollar value), one simply multiplies the baseline cost per epidemic with the population growth rate and projected increase in the number of cases (see third row in table A.40). Example: The cost per epidemic in 2100 is: US$2.4-$15.6 million x 1.66 to 2.98 = US$4.0 to US$46.2 million (1998 baseline) (population growth) (future costs adjusted by population growth) US$4.0-$46.2 million x 1.4 to 2.0 = US$ 5.6 to US$93.0 million (future costs adjusted (projected increase (future costs adjusted by population growth and by population growth) in number of cases) projected increase in number of cases) 5.2. Assess the Incremental Costs due to Climate Change. The above costs reflect the total average economic costs of future epidemics under climate change conditions. To compute the incremental impact of climate change, one needs to subtract the baseline costs from this estimate. Annex A Page 63 Example: The baseline costs are US$2.4-$15.6 million. The cost of an epidemic in 2100 are US$5.6-$93.0 million. Hence, the incremental costs due to climate change are: (US$5.6-$93.0 million) - (US$2.4-$15.6 million) = US$3.2 - $77.4 million per epidemic. 5.3. Transform Incremental Epidemic Costs into Annual Averages. For the last 30 years, there have been 8 dengue fever epidemics in Fiji (Basu and others 1999), a frequency of once every 3.75 years. It is expected that under future climate change conditions, dengue fever may become endemic (occurring every year). However, this change is to a certain extent already captured in the expected increase in the number of cases, which is based on the estimated increase in epidemic potential. To add to that an increase in the frequency of epidemics could introduce double counting. For the purposes of the analysis, the baseline frequency of epidemics was assumed to remain unchanged from the baseline conditions. To compute the annual average costs due to climate change, the incremental costs due to climate change are simply divided by 3.75. Example: The incremental costs due to climate change in 2100 are US$3.2-$77.4 million per epidemic. The annual average costs are therefore: US$3.2-$77.4 million / 3.75 = US$0.9-$20.6 million 5.4. Compute the Equivalent Costs for Viti Levu. The above estimates apply to Fiji as a whole. To compute the average incremental costs of climate change on dengue fever epidemics in Viti Levu, the numbers need to be adjusted by 77 percent, the proportion of the total Fijian population that resides in Viti Levu. This is a conservative assumption, as the incidence of dengue fever is likely to be higher under the crowded conditions of Viti Levu's towns, and Viti Levu is expected to grow faster than other places in Fiji. However, to remain consistent with the assumptions used in impacts on other sectors, a simple population adjustment was used. Example: Adjusting the incremental annual costs due to climate change in 2100 to Viti Levu: US$0.9-$20.6 million x 0.77 = US$0.7-$15.9 million (see table A.40) The results are summarized in table A.40. Table A.40 Estimated Costs of Climate Change on Dengue Fever in Viti Levu (millions of 1998 US$) Baseline 2025 2050 2100 (mid-1990s) Population 772,655 +44-57% +63-110% +66-198% Number of Cases per Epidemic 3,700-24,000 +10% +20-30% +40-100% Costs per Epidemic (US$ million) 2.4-15.6 3.8-26.9 4.7-42.6 5.6-93.0 Incremental Costs per Epidemic due to Climate Change (US$ million) -- 1.4-11.3 2.3-27.0 3.2-77.4 Annual Incremental Costs of Climate Change (US$ million) -- 0.4-3.0 0.6-7.2 0.9-20.6 Annual Incremental Costs of Climate Change in Viti Levu -- 0.3-2.3 0.5-5.5 0.7-15.9 (US$ million) Annex A Page 64 Impacts of Climate Change on Diarrheal Disease Diarrheal diseases are caused by a range of pathogens and influenced by many different factors. The marked seasonality of diarrhea occurrence and association of diarrhea with rainfall extremes in Fiji suggests a relationship with climate: the first three months of the yearwhich are typically the warmest and wettestare also those when diarrheal diseases are typically the most common. The CSIRO Global Circulation Model used in this study projects an increase in precipitation along with gradual warming (table A.1). If the assumption is made that the increased incidence of diarrheal disease in the first three months of the year is related to the warmer and wetter conditions typical of these months, then the longer, warmer and wetter conditions induced by climate change would be expected to result in an increased incidence of diarrheal disease. Scenarios of increased climate variability and extremes are also relevant, as an increase in the frequency and severity of droughts and floods in Viti Levu would likely result in a higher incidence of diarroeal disease. Analysis of monthly reports of diarrhea in infants between 1978 and 1989 in Fiji suggests that the effect of climate cannot be distinguished statistically from seasonal patterns. However, if one assumes that the seasonal pattern is mainly attributable to changes in monthly average temperature and rainfall, then it is possible to estimate the independent effects of these variables on diarrhea incidence. The analysis of climate change impacts on diarrheal disease involves two major steps: Step 1: Assess the Expected Increase in the Number of Cases due to Climate Change. To estimate the impact of temperature changes on diarrhea incidence, one can use a simple regression: Diarrhea disease = (average temperature) where: "Diarrhea disease" are the monthly reports of infant diarrhea in Fiji during 1978-89; and "Average temperature" are average monthly temperatures recorded for Fiji The results of this regression indicate that an increase in temperature of 1°C is associated with 100 additional reports of infant diarrhea per month, with a probability of error of less than 5 percent. Data for 1990-1991 are missing, but the model was a good predictor of 1992-1998 diarrhea reports: the correlation coefficient between model predictions and true values was 0.47, with a probability of this correlation being due to chance alone of less than 0.1 percent. Since the true incidence of diarrhea is likely to be at least 10 times the number of reported cases, an increase of 1°C is estimated to be associated with at least 1000 additional cases per month based on Fiji's current population. This result can also be used, albeit with lower confidence, to estimate the potential impacts on children and adult diarrhea. Thus, an increase of 1°C is estimated to be associated with 12,000 additional cases of diarrhea per year (1,000 cases x 12 months per year), based on the current population. Annex A Page 65 Step 2: Estimate the Economic Costs of Climate Change. The costs of diarrheal disease can be estimated based on the value of loss productivity due to disease. It is assumed that each diarrhea episode lasts 2-5 days, or 3.5 days on average, which is consistent with world health estimates. The incremental annual costs attributable to climate change can then be estimated as follows: ˇ Compute the increase in the number of annual diarrhea cases due to climate change ˇ Estimate the total number of days of lost productivity per year ˇ Compute the costs of lost productivity ˇ Adjust the estimates to the population of Viti Levu. 2.1. Compute the Increase in the Number of Annual Diarrhea Cases due to Climate Change. As seen above, a 1°C increase in temperature is associated with an estimated 12,000 incremental cases of diarrhea a year. To estimate the impact of climate change, it is necessary to adjust the number of cases first by the expected population growth, and second by the expected change in temperature. Example: In 2050, the Fijian population is projected to grow by 63 to 110 percent (1.63-2.10 fold), and the temperature is projected to increase by 0.9-1.3 °C. Hence, the incremental number of annual diarrhea cases attributable to climate change are as follows: 12,000 x (1.63 to 2.10) = 19,560 to 25,200 (increase in number of diarrhea cases in the 2050 Fijian population due to a 1 °C increase in average temperature) 19,560 to 25,200 x (0.9 to 1.3 ) = 17,604 to 32,760 (increase in number of diarrhea cases in the 2050 Fijian population due to the temperature changes predicted by the climate change models) 2.2. Estimate the Total Number of Days of Lost Productivity per Year. The number of days of lost productivity is simply the incremental number of annual diarrhea cases due to climate change times 3.5 days (the estimated average duration of the episodes): Example: In 2050, the total number of days of lost productivity due to climate change is: 17,604 to 32,760 x 3.5 days = 61,674 to 114,660 days of lost productivity 2.3. Compute the Costs of Lost Productivity. As seen in the dengue fever estimates, the average cost of a day of lost productivity is US$6.88. It can be argued that many of the diarrhea cases involve children. However, the estimate of US$6.88 takes into account the productivity value of an average day for the entire Fijian population (including those that do not work) and is therefore considered to be a valid estimate. Example: In 2050, the annual costs of lost productivity would be: 61,674 to 114,660 days x US$6.88 per day = US$0.4 to US$0.8 million Since the projected increases in the number of diarrhea cases are yearly values, no adjustment is necessary to estimate the annual costs of climate change. Annex A Page 66 2.4. Adjust the Estimates to the Population of Viti Levu. As it was done for dengue fever, the costs of lost productivity caused by climate change need to be adjusted by a factor of 0.77, to account for the proportion of the population that resides in Viti Levu. Example: In 2050, the costs of climate change in terms of lost productivity due to diarrheal disease in Viti Levu are: US$0.4 to US$0.8 million x 0.77 = US$0.3 to US$0.6 million The results are summarized in table A.41 below: Table A.41. Estimated Costs of Climate Change on Diarrheal Disease in Viti Levu (millions of 1998 US$) Baseline (mid-1990s) 2025 2050 2100 Population 772,655 +44-57% +63-110% +66-198% Temperature Change +0.5-0.6o C +0.9-1.3o C +1.6-3.3o C Increase in Annual Cases due to Climate Change 8,640-11,304 17,604-32,760 31,872-118,008 Total Number of Lost Productivity Days 30,240-39,564 61,674-114,660 111,552-413,028 Annual Incremental Costs of Climate Change in Fiji 0.2-0.3 0.4-0.8 0.9-2.8 (US$ million) Annual Incremental Costs of Climate Change in -- Viti Levu (US$ million) 0.2 0.3-0.6 0.6-2.2 The overall impact of climate change on the health sector of Viti Levu is summarized in table A.42. Table A.42. Estimated Annual Economic Impact of Climate Change on Public Health in Viti Levu, Fiji, 2025-2100 (millions of 1998 US$) Estimated Incremental Economic Costs due to Climate Change 2025 2050 2100 Cyclones and Droughtsa Likely to be substantial Dengue Fever 0.3-2.3 0.5-5.5 0.7-15.9 Diarrheal Diseases 0.2 0.3-0.6 0.6-2.2 Nutriton-Related Illnesses + + + Total Estimated Costs 0.5-2.5 0.8-6.1 1.3-18.1 Notes: Not available. + No quantifiable data available, but damages likely to be substantial. a. The effect of cyclones and droughts on health could not be calculated, but the overall impact of cyclones was captured on the water resources section. Annex A Page 67 Tarawa The potential impacts of climate change on public health in Tarawa were analyzed quantitatively for ciguatera poisoning and dengue fever. Apart from preliminary work on ciguatera, however, there were no available studies examining the relationship between health and climate change in Kiribati. Hence, no economic analysis could be performed, and the findings of this report need to be interpreted in light of these limitations. Impacts of Climate Change on Dengue Fever Kiribati has two vectors for dengue fever, Aedes aegypti and Aedes albopictus. There have been four recorded outbreaks of dengue fever , in 1971/72, 1974, 1980/81, and 1988. South Tarawa has several factors which contribute to dengue fever risk. It is the main international port of entry into Kiribati. The crowded urban areas of Betio, Bairiki and Bikenibeu increase the risk of transmission. Discarded container items such as tins, empty bottles and used tyres as well as unscreened rainwater tanks, home flower vases, tree holes, and shells provide habitat for the vectors and increase the epidemic risk. The current climate with an average temperature of 31o C and average monthly rainfall of 500 mm is highly suitable for vector survival and multiplication. In the analysis of dengue fever, it was postulatedas for Viti Levuthat temperature is an important determinant of epidemic risk where a capable vector population is present. According to Patz et al. (1998), epidemic potential is negligible below 23o C, escalates rapidly from about 30o C and drops precipitously at about 40o C due to a rapid increase in mosquito mortality. This relationship between temperature and epidemic potential was used to model changes in epidemic potential in the PACCLIM model (table A. 43): Table A.43. Potential Impact of Climate Change on Dengue Fever in Tarawa, Kiribati Likely Changes Baseline 2025 2050 2100 (1990) Projected Changes in Epidemic 0.18 0.20 0.22-0.24 0.25-0.36 Potential Percentage Change from the Baseline 11% 22-33% 39-100% Note: Ranges represent the best guess and worst case scenarios. Only the CSIRO GCM model was used here. Given Tarawa's limited size, high population densities and availability of vector breeding sites, it is likely that the vast majority of the population in South Tarawa would be exposed during an epidemiceven in the absence of climate change. Consequently, while climate change could increase the transmission efficiency, and therefore the rate at which the epidemic grows, it may not greatly affect the total number of cases produced by an epidemic. Climate change, however, may influence the numbers, density and distribution of the vector. It is also likely that the impact of climate change worldwide could increase the prevalence of all dengue virus serotypes, leading to a higher incidence of severe forms of dengue fever. Moreover, given that smaller vector populations will be able to sustain an epidemic, future efforts at vector control will need to reach increasingly higher levels of the vector population in order to achieve the same result in terms of epidemic risk reduction. Annex A Page 68 Impacts of Climate Change on Ciguatera Poisoning Kiribati has one of the highest incidences of ciguatera in the Pacific (Lewis and Ruff 1993). Ciguatera is contracted by ingesting reef fish contaminated with ciguatoxins, produced by dinoflagellate organisms that accumulate as they move up the food chain. Symptoms can last from days to even months. Lewis (1992) suggests that a small increase in ciguatera poisoning in several Pacific Island countries may be related to El Niņo event. It is likely that climate is only one of several factors affecting ciguatera, others being reef disturbance and pollution. A study of ciguatera in eight Pacific Island countries found positive correlations between the annual incidence of ciguatera and local warming of the sea surface in one group of islands (including Kiribati) which experienced local warming during El Niņo conditions (Hales et al. 1999). The study found a statistically significant relationship between sea surface temperature and the reported incidence of ciguatera fish poisoning in Kiribati (figure A.9). Figure A.9. Relationship Between Sea Surface Temperature and Incidence of Ciguatera in Kiribati Kiribati 1.0 C) 0.8 0.6 rees 0.4 eg (d 0.2 y 0.0 al mo -0.2 -0.4 an -0.6 SST -0.8 -1.0 0 5 10 15 20 25 30 cases per year Source: Hales et al. (1999). This relationship was used in PACCLIM to model future changes in ciguatera incidence based on projected future trends in temperature. As Sea Surface Temperature is not a parameter available in PACCLIM, projected changes in atmospheric temperature at sea level were used as a proxy. The results are shown on table A. 44: Table A.44. Potential Impact of Climate Change on Ciguatera Incidence, Kiribati Likely Changes Baseline 2025 2050 2100 (1990) Predicted Number of Reported Casesa 7 21-24 32-43 49-101 Predicted Ciguatera Incidenceb 35-70 105-240 160-430 245-980 Note: Ranges represent the best guess and worst case scenarios for both the CSIRO and the DKRZ models (which gave similar results). a. Incidence rate per 1000 population b. Assuming a reporting rate of 10-20 percent. Annex A Page 69 These results should be interpreted with considerable caution, as the model has not been validated and the data set from which it was derived was limited. It is possible that many other unknown factors may influence the incidence of ciguatera, and that low present reported incidence rates may result in a high degree of error in future projections. The overall impact of climate change on ciguatera should be measured perhaps not in terms of the number of cases but rather in terms of how people respond to increase risk, including changes in diet, decreased protein intake, loss of revenue from fisheries, etc. Unfortunately, no data were available to permit an estimate of the economic value. Annex A Page 70 G. Impacts on Regional Tuna Fisheries The analysis of the possible impacts of climate change on tuna fisheries was done based on recent studies of impact of ENSO on tuna resources in the western central Pacific ocean and on simulation results from a spatial population dynamics model, SEPODYM. In this model, the movement of tuna and the effects of environmental variability are considered. Movement is governed by a diffusion-advection equation in which the advective term is proportional to a habitat gradient index that is a function of forage density (linked to primary productivity) and sea surface temperature. Sea surface temperature, ocean currents, and primary production are used in the model to delineate tuna spawning areas, transport larvae and juveniles, and stimulate tuna forage distribution. The economic impacts of climate change and variability on regional tuna fisheries could only be assessed qualitatively. Nonetheless, it is worthwhile to compare the present condition of the fishery with the likely future conditions, particularly with regards to (a) changes in total tuna abundance; (b) changes in spatial distribution (and consequently benefit sharing among coastal states); and (c) changes in the economic incentives of the fleet. The Present The abundance of tuna stocks in the Central and Western Pacific is influenced by the primary productivity of the ocean, which in turn varies with ENSO events. In general, El Niņo years tend to result in a higher than average recruitment of skipjack tuna in subsequent months, while La Niņa events tend to result in higher recruitment of albacore tuna in subsequent years. The relationship is less clear for yellowfin and bigeye tuna, which are more widely distributed. The spatial distribution of tuna is also affected by ENSO events. As the Western Pacific Warm Pooland the highly productive `cold tongue' that separates it from the eastern equatorial Pacificextends eastwards during El Niņo events, tuna (in particular skipjack) migrate eastwards, and countries like Kiribati and Samoa experience higher purse seine catches (figures A.10-11). Conversely, countries in the western Pacific like the Solomon Islands and the Marshall Islands enjoy higher catches during La Niņa years (figure A.12). Figure A.10. Movement of Tagged Skipjack Tuna in the Central and Western Pacific with ENSO Figure A.11. Estimated Change in Tuna Catch in Kiribati with Intensification of El Estimated change in catch of tuna in Kiribati by one unit decrease in the El Nino index (ISO) (A declining ISO indicates an El Nino event) 1000 800 600 Yellowfin Skipjack mt 400 Bigeye Albacore 200 0 -200 1st 2nd 3rd 4th Quarter of the year Sources: Lehodey et al. (1997), SPC, and background reports to this study. Annex A Page 71 Figure A.12. Share of Total Tuna Catch Captured by Coastal States During ENSO Marshall Islands -- Average Share of Total Tuna Kiribati -- Average Share of Total Tuna Catch Catch (1980-98) (1980-98) 12.00% La Nina Years 9.00% 10.00% La Nina Years El Nino Years 8.00% El Nino Years h 8.00% tc 7.00% Cal 6.00% hctaClat 6.00% taoTfo 5.00% To 4.00% 4.00% erahS 3.00% 2.00% oferahS 2.00% 1.00% 0.00% 0.00% Longline Pole and Line Purse Seine Total Catch Longline Pole and Line Purse Seine Total Catch Major Fisheries Major Fisheries Solomon Islands -- Average Share of Total Tuna Samoa -- Average Share of Total Tuna Catch Catch (1980-98) (1980-98) La Nina Years La Nina Years 45.00% 0.30% El Nino Years 40.00% El Nino Years hctaC 0.25% 35.00% 30.00% 0.20% altoTfo 25.00% 0.15% 20.00% hctaClatoTfo e 15.00% e 0.10% arhS10.00% arhS0.05% 5.00% 0.00% 0.00% Longline Pole and Line Purse Seine Total Catch Longline Pole and Line Purse Seine Major Fisheries Major Fisheries In Kiribati, a decline of one unit in the Southern Oscillation Index (indicating a move towards El Niņo conditions) has led, on average, to a 200-800 metric tons increase in average tuna catch. Such relationship is much harder to discern for the total tuna catch of the Central and Western Pacific. An important factor in determining the future impact of climate change is the degree of dependency of coastal states on fisheries. Micronesian countries like Kiribati, the Federated States of Micronesia, and the Marshall Islands are much more vulnerable to changes in relative abundance of tuna stocks than countries such as Fiji, Samoa, and the Solomon Islands, where the economies are more diversified (table A.44). Table A.45. Coastal States' Relative Dependence on Fisheries Coastal States Fisheries as a Percentage of GDP Fisheries as a Percentage of Exports Kiribati 13 27 Federated States of Micronesia 2 89 Marshall Islands 9 89 Fiji 5 6 Samoa 6 47 Solomon Islands 6 23 Sources: IMF and country economic statistics. See Volume III of this report. Annex A Page 72 The Likely Future Climate change is likely to affect the tuna fisheries of the Central and Western Pacific in two major ways: first, the average sea surface temperatures may evolve toward a `mean El Niņo' state. Second, inter- annual variability (alternating between El Niņo and La Niņa events) may increase. A mean state El Niņo would impose a permanent change in the system which has not been observed until now. Thus, while the historical evidence suggests that countries in the central Pacific (such as Kiribati) might benefit from a redistribution of the stocks, modeling results indicate the opposite effect. The primary productivity of the upwelling system in the central and eastern equatorial Pacific is likely to decline, affecting the abundance of bigeye and adult yellowfin population. Tuna stocks may redistribute to higher latitudes and toward the western equatorial Pacific. Countries in the central Pacific, such as Kiribati, are likely to suffer disproportionally as a result. While purse seine catches are not expected to be significantly affected, the decline in bigeye and adult yellowfin stocks could lead to over-exploitation. Worldwide demand for sashimi has risen steadily. A reduction in catch in the central Pacific could lead to higher prices, further placing pressure on the resources. The difficulties in adapting sufficiently rapidly to new conditions may be particularly challenging if it turns out that the increased variability scenario comes into effect. The need to compensate for losses due to a drop in catch during one season may lead to over-fishing in the next period. Distant water fishing nations' fleets would be able to move among EEZs to capitalize on these variations; for domestic fleets, however, these increased fluctuations could be particularly difficult. Stronger regional collaboration among coastal states in tuna management will be needed to counteract these trends.