Report No. 67668-SAS Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains b Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Donald Alford, David Archer, Bodo Bookhagen, Wolfgang Grabs, Sarah Halvorson, Kenneth Hewitt, Walter Immerzeel, Ulrich Kamp, and Brandon Krumwiede i This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Acknowledgements This volume was prepared by a team led by Winston Yu (the World Bank) and Donald Alford (Consultant). Don Alford, David Archer (Newcastle University), Bodo Bookhagen (University of California Santa Barbara), and Walter Immerzeel (Utrecht University) contributed to the sections related to mountain hydrology. Wolfgang Grabs (World Meteorological Organization) developed the sections in the report on climate monitoring. Sarah Halvorsen (University of Montana) prepared the sections on indigenous glacier monitoring. Kenneth Hewitt (Wilfrid Laurier University) developed the sections on glacier mass balance monitoring. Ulrich Kamp (University of Montana) and Brandon Krumwiede (US National Weather Service) contributed to the sections on satellite imagery and digital elevation models. Editorial support of John Dawson is gratefully acknowledged. The authors benefited enormously from the many technical discussions with colleagues during the preparation of this report and strategic guidance from senior management. Generous support was provided by the World Bank and the South Asia Water Initiative. ii Contents Contents About the Authors x Abbreviations and Acronyms xiii Executive Summary xv Monitoring Objectives xv Monitoring of Glaciers, Climate, and Runoff: Main Themes xvii Climate xvii Glaciers xvii Hydrology xviii Indigenous Monitoring xviii Satellite Imagery and GIS xix Mesoscale Imagery xix Macroscale Imagery xx MODIS xx TRMM xx AVHRR xx DEMs and Geomorphometry xx Requirements for Instituting a Monitoring Program xxi 1. Introduction 1 1.1 History 2 1.2 The Problem 3 1.3 Scale and Location 3 1.4 Objectives and Procedures 5 References 6 2. Climate Monitoring 7 2.1 Monitoring Objectives 7 2.2 Previous Network Design Recommendations 8 2.3 Use of Climate Networks 9 2.3.1 Temperature 9 2.3.2 Precipitation 10 2.4 Environmental Features Affecting Variations in Climate and Glacier Mass Balance 11 2.4.1 Precipitation 11 2.4.2 Temperature 12 2.4.3 Energy Balance Variables 14 2.5 Monitoring and Analysis Needs 14 2.5.1 Using Existing Climatological Data 15 2.5.2 Assessing Stationarity and Homogeneity of Records 16 iii Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Observational and Entry Errors 17 Changes of Instrument or Measurement Practice 17 Changes in Station Location 17 Changes in Station Environment 17 2.6 Identification of and Adjustment for Bias 18 2.6.1 Statistical Methods 18 2.6.2 Investigation of Regional Consistency 19 2.7 Monitoring Network Components 19 2.7.1 Automatic Weather Stations 19 2.7.2 Communications 20 2.7.3 Measurement of Snow 21 Falling Snow 21 Snow on the Ground 21 2.8 Recommendations 22 2.8.1 Data and Metadata Acquisition and Validation Recommendations 22 2.8.2 Climate Analysis Recommendations 23 2.8.3 Monitoring and Instrumentation Recommendations 24 References 24 3. Glacier Mass Balance Monitoring 27 3.1 Monitoring Approaches 28 3.2 High Asian Context 29 3.3 National and Transboundary Issues 30 3.4 Glacier Inventories and Reference Materials 31 3.5 Past and Present Monitoring Efforts in the Region 32 3.5.1 India 32 3.5.2 China 33 3.5.3 Nepal 33 3.5.4 Pakistan 34 3.6 Current State of Direct Glacier Monitoring 34 3.6.1 Elements of Mass Balance in the HKH 36 3.6.2 Accumulation and Source Zones 38 3.6.3 High-Elevation Snowfall at Biafo Glacier, Central Karakoram 40 3.6.4 Ablation in the HKH 42 3.7 Debris-covered Glaciers 43 3.8 Water Yield from Glaciers 44 3.9 Glacier Regimes 46 3.10 Mass Balance Gradients 46 3.11 Verticality 47 3.12 Glacier Motion 48 3.13 Thermal Classes 49 3.14 Neglected Seasons 50 3.15 Discussion 51 3.15.1 Field Programs and Instrumentation 51 3.15.2 Personnel and Safety 51 References 52 iv Contents 4. Mountain Hydrology 57 4.1 Background to Mountain Hydrology 57 4.2 Monitored Streamflow of the HKH Mountains 59 4.2.1 The Indus River 59 4.2.2 Upper Indus Basin Hydrology 61 4.2.3 The Nepal Himalaya 62 4.2.4 Recession Flows 63 4.2.5 East–West Variation in Runoff 64 4.2.6 Altitudinal Gradients of Runoff 64 4.2.7 Initial Uses of the Existing Network 65 4.3 Assessing Comparative Contribution to Streamflow 65 4.4 Streamflow Monitoring 68 4.4.1 Quality of Streamflow Measurements 69 4.4.2 Site Selection 70 4.4.3 Water Level Measurement 70 4.4.4 Establishing a Relationship between Water Level and Discharge 71 4.4.5 Transforming the Record of Stage to Discharge 73 4.4.6 Evaluating Historical Discharge Records 74 4.5 New Network Requirements 75 4.6 Summary and Recommendations 75 References 76 5. Indigenous Glacier Monitoring 79 5.1 Indigenous Monitoring: Overview and Purpose 80 5.2 Vulnerability of Mountain Communities: Some Considerations 81 5.2.1 Glacial Recession 82 5.2.2 Demographics 83 5.2.3 Gaps in Knowledge and Awareness of Mountain Hazards 83 5.2.4 Male Out-migration 84 5.3 Glacier Hazard Management Issues 84 5.4 Solutions for Indigenous Glacier Monitoring in the HKH Region 84 5.5 Observations and Recommendations 85 5.5.1 Observations 85 5.5.2 Recommendations 86 General community interventions 86 Development of indigenous monitoring teams, as in “citizen scientist” programs 86 Support and enhance hazard preparedness and disaster risk reduction at local level 86 References 87 6. Satellite Imagery and Digital Elevation Models 89 6.1 Literature Review 89 6.2 Requirements for Glacier Monitoring Program 90 6.3 Mesoscale Satellite Imagery 90 6.4 Glacier Monitoring Using Satellite Imagery and DEMs 92 6.5 Monitoring Debris-free Glaciers 93 6.6 Monitoring Debris-covered Glaciers 95 v Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 6.7 Global Land Ice Measurements from Space 97 6.8 Macroscale Satellite Imagery 98 6.8.1 Moderate Resolution Imaging Spectrometer 99 6.8.2 Tropical Rainfall Monitoring Mission 100 6.8.3 Advanced Very High Resolution Radiometer 101 6.9 DEMs and Geomorphometry 101 6.9.1 Source Data 101 6.9.2 Error Calculation 104 6.9.3 Ground Control Points 104 6.9.4 Postprocessing 104 6.9.5 Software Packages 105 Satellite Imagery and DEMs 105 Satellite Imagery, DEMs, and GIS 106 6.10 DEM Analysis 106 6.10.1 Geomorphometry 106 6.10.2 Land Surface Parameters 107 6.10.3 Topographic Radiation Modeling 107 6.10.4 Altitudinal Functions 107 6.11 Summary 108 6.11.1 Satellite Imagery 108 6.11.2 DEMs 108 References 110 7. Monitoring of the HKH Cryosphere 114 7.1 Considerations and Technical Procedures for HKH Monitoring 114 7.2 Selection of Monitoring Networks and Logistical Considerations 115 7.3 Practical Procedures the HKH Cyosphere 115 7.3.1 Guiding Principles 115 7.3.2 Essential Variables 115 7.3.3 Requirements Document 116 7.3.4 Components of a Cryospere Monitoring Network 116 7.3.5 Historical Data Records 116 7.3.6 Telecommunications 117 7.4 Data Management 117 7.4.1 Access to Data and Information 117 7.4.2 Metadata 117 7.4.3 Database Management Systems 118 7.4.4 Data Integration and Management 118 7.4.5 Data Management and Reanalysis 118 7.4.6 Development of Analysis and Forecast Procedures 119 7.5 Institutional Setup and Organization 119 7.6 Cryosphere Monitoring Program Components 120 7.7 IGOS Monitoring Principles 121 7.8 General Considerations 122 7.8.1 Costs of Field Trips 122 7.8.2 Selection of Location 122 vi Contents Reference 123 Recommended Reading on Monitoring 123 General 123 Glacier Monitoring 123 Guidelines and Standards Relating to the International Glacier Monitoring Strategy 123 Guidelines and Standards Relating to Measurement of Glacier Fluctuations 123 Snow Monitoring 124 Climate Monitoring 124 Hydrological Monitoring 124 vii Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains FIGURES Figure 2.1 Seasonal Temperature and Runoff, Figure 3.8 Debris Cover on Ablation Zone of June–August, at Two Locations in Baltoro Glacier, Central Pakistan 10 Karakoram, June 44 Figure 2.2 Preceding Seasonal Precipitation Figure 3.9 Light, Scattered Debris, Upper (October–March) at Astore and Baltoro Glacier, Representative Runoff (July–September) at Two of about Two Thirds of the Locations 11 Ablation Zone, July 44 Figure 2.3 Estimates of Monthly Freezing Level Figure 4.1 Mountain Catchment Basins of and Seasonal Mean Daytime Land the Indus River 60 Surface Temperature Lapse Rates Figure 4.2 Diversity of Annual Streamflow for Upper Indus Basin 13 from Catchments in the Upper Figure 2.4 Annual Variation of the Indus Basin, One Year 61 Temperature Lapse Rate for the Figure 4.3 Recession Curves for Glacierized Sutlej River Valley 14 Basins of the Karakoram, Based Figure 2.5 Kunjerab Automatic Weather on Mean Monthly Data for Station at 4,733 m above Sea July–December 63 Level in Hunza Tributary of the Figure 4.4 Recession Curves for Besham, Upper Indus in Pakistan 20 Based on Mean Monthly Data Figure 3.1 Main Zonal, Vertical, and Mass for July–December 63 Balance Regimes of Valley Glaciers 37 Figure 4.5 East–West Variation in Specific Figure 3.2 Typical Avalanche-fed Glacier: Runoff in HKH 64 Bazhin Glacier, Nanga Parbat Figure 4.6 Regional Orographic Runoff East Face 39 Gradient for the Himalaya Figure 3.3 Avalanche-nourished Sumaiyar Bar Based on Data from Glacierized Tributary of Barpu Glacier, Central and Nonglacierized Basins 65 Karakoram 39 Figure 4.7 Estimated Glacier Melt Figure 3.4 Biafo Glacier Accumulation Zone: Contribution to Total Annual Source of Snow Pit and Drill Core Flow, HKH Mountains 67 Samples 40 Figure 4.8 Typical Arrangement for Figure 3.5 Snowfall (Water Equivalent) from Water Level Measurement by Selected Sites on Biafo Glacier Pressure Transducer in the HKH 71 and Adjacent Basins, 1983–88 41 Figure 4.9 Typical Discharge Measurement Figure 3.6 Accumulation Profile Exposed in Devices in the HKH 72 a Crevasse, Biafo Glacier 42 Figure 4.10 Typical Examples of ADCP in Figure 3.7 Ablation Season Weather Use for Discharge Measurement 73 Observations for On-ice and Figure 6.1 Landsat ETM+ Index Map for Off-ice Stations at Same Elevation the HKH Region 91 and 1.5 km Apart at Baintha Profile, Biafo Glacier, 4,050 m, Figure 6.2 ASTER Image of Glaciers in the 1986 43 Himalaya of Bhutan and China 92 viii Contents TABLES Figure 6.3 ALOS AVNIR-2 Scene Covering Table 2.1 Recommended Minimum Density Sagarmatha National Park, of Precipitation Stations 9 Nepal 92 Table 4.1 Descriptive Statistics of the Basins Figure 6.4 Delineation Results for Glaciers Considered in the Study 60 in the Northern Tien Shan 94 Table 4.2 Descriptive Statistics of the Figure 6.5 Simple Threshold Ratio Mapping Glacierized Catchment Basins Approach Using Landsat 7 Bands of the Nepal Himalaya 62 4 and 7 for Parts of the Himalaya in India (33°N 77°E) 94 Figure 6.6 Glacier Mapping Results Using Different Band Ratios Applied to a Landsat Image of Ikh Turgen Range 95 Figure 6.7 Characteristics of Supraglacial Debris of Glaciers in Northern Pakistan Derived from SPOT Imagery Multispectral Analysis 95 Figure 6.8 Results from Different Glacier Mapping Steps for the Nun Kun Mountains in Zanskar 96 Figure 6.9 Results from Morphometric Glacier Mapping (MGM) of Glaciers in Himalaya Range of Zanskar, India, Using ASTER Satellite Imagery and ASTER DEMs 96 Figure 6.10 Viewing GLIMS ASTER Browse Data within Google Earth 98 Figure 6.11 SRTM Index Map for the HKH Region 103 ix Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains About the Authors Donald Alford is a consultant in mountain Bodo Bookhagen is Associate Professor at the hydrology in Billings, Montana. He studied at Department of Geography, University of California Montana State University and the Institute of at Santa Barbara. He received a PhD (summa Hydrology and Glaciology, University of Zurich, cum laude) from Potsdam University (Geology). and received a PhD for a dissertation on cirque His professional interests include understanding glaciers from the Institute of Arctic and Alpine Quaternary climate change, geomorphic processes, Research, University of Colorado. He developed landscape evolution, and tectonic processes through the High Altitude Research Program at the Cold integrated studies involving cosmogenic radionuclide Regions Research and Engineering Laboratory, dating, recent and past climatic records, remote where he studied snow and glaciers in the Rockies sensing, numerical modeling, and field observations. and St. Elias Range, as well as participating in snow stratigraphy and seismic traverses of Wolfgang Grabs is Chief of the Hydrological northern Greenland. Alford specializes in the Forecasting Division of the World Meteorological study of mountain hydrological systems. He has Organization (WMO) in Geneva. Before joining participated in studies of mountain glaciology, WMO in 1999, he worked at the Global Runoff hydrology, applied water resources development, Data Center in Koblenz, Germany, and also and mountain hazard and risk assessment has extensive experience working in Africa and and management for 35 years in geophysical Asia. He is responsible for the development and environments ranging from northern Greenland to implementation of the Mekong Hydrological Cycle the subtropics of Southeast Asia. Observation System (HYCOS) and the Arctic HYCOS, and the development of the Hindu Kush- David Archer is at the Water Resource Systems Himalaya HYCOS. He established the Glacier and Research Laboratory, School of Civil Engineering and Climate Research Group in the Nepal Department Geosciences, Newcastle University, United Kingdom, of Hydrology and Meteorology, which currently and is a consultant with JBA, Consulting Engineers maintains a network of high-altitude climate and Scientists, North Yorkshire. He has worked in monitoring stations in the Nepal Himalaya. academia at the United Kingdom’s University of Newcastle and as a hydrologist at the Northumbrian Sarah Halvorson is Professor at the Department Water and National Rivers Authority. He has been a of Geography, University of Montana in Missoula. consultant in development environments in Asia and Her teaching and research interests span broad Africa for 12 years and has published more than and diverse areas including gender and social 50 academic papers and two books. His research aspects of water resources and environmental includes studies of climate change impacts on river hazards; medical and health geography; gender flow in the upper Indus basin, at the western end of geography; international development in the Himalaya-Karakoram-Hindu Kush, as well as Central and South Asia and Africa; and water in Africa. He is currently conducting research at the and landscape transformation in the Rocky School of Civil Engineering and Geomatics at the Mountain West. In the 1990s, she carried out University of Newcastle. His studies of the climate ethnographic fieldwork in mountain communities of the upper Indus basin and western Himalaya are in the Karakoram of northern Pakistan. This work generally considered to be the definitive standard. culminated in a doctoral dissertation entitled x About the Authors Geographies of Children’s Vulnerabilities: (NWO) and working on seasonal forecasting Households and Water-Related Disease Hazard in of Asian river discharges from the Himalayan Northern Pakistan, from the University of Colorado. cryosphere and monsoon feedbacks in close Since 2000, she has carried out field studies in the collaboration with Utrecht University. He currently Bitterroot valley of Montana, Royal Kingdom of works as a postdoctoral researcher at Utrecht Bhutan, Republic of Georgia, Kyrgyzstan, Turkey, University and ETH Zurich and is responsible for a Tajikistan, and the Xinjiang Uyghur Autonomous number of projects at the cutting edge of climate Region of China. change and hydrology. In 2011, he was awarded a prestigious NWO-VENI grant to support his research Kenneth Hewitt is Professor Emeritus in Geography on the impacts of climate change on the hydrology and Environmental Studies and is a Research of the Himalaya and Karakoram mountain ranges. Associate at the Cold Regions Research Centre at Wilfrid Laurier University in Ontario, Canada. He Ulrich Kamp is Associate Professor at the received his PhD in Geomorphology from London Department of Geography, University of Montana University. His main research interests are in in Missoula. He began his career at the Institute for glaciers, catastrophic landslides, and environmental Space Sciences at Freie Universität, Berlin, where disasters. His regional specializations are mainly in he focused on airborne remote sensing and water high-mountain environments worldwide, especially quality monitoring of lakes and rivers in urban the Karakoram Himalaya, inner Asia, with 16 field areas. In 1999, he received his PhD in Geography seasons there. He has published extensively on from Technische Universität, Berlin, with a thesis these topics and is one of the leading authorities on about Quaternary geomorphology and glaciations the glaciers of the western Hindu Kush-Himalaya in the Pakistani Hindu Kush. He then carried out Mountains. postdoctoral studies at the Department of Geography and Geology at the University of Nebraska, Omaha, Walter Immerzeel has 12 years’ experience in in remote sensing of glaciers in the Himalaya. He geo-informatics, water resources management, and then spent three years as an assistant Professor of climate change and is skilled in hydrometeorological Geography and Environmental Science at DePaul monitoring, the use of remote sensing, simulation University in Chicago before joining the University of models, and spatial analysis. He has been doing Montana in summer 2005. As a research fellow of research on Himalayan hydrology since 2002. the Alexander von Humboldt Foundation of Bonn, He holds a PhD in Physical Geography from Germany, Kamp spent the academic year 2010–11 Utrecht University and his research focused on the at the Institute for Space Sciences at Freie Universität, interface of mountain hydrology, climate change, where he worked on monitoring of glaciers in the and agriculture. From December 2002 until June Altai Mountains of Mongolia. His research includes 2004, he was attached to the International Centre mountain geography, geomorphology, Quaternary for Integrated Mountain Development (ICIMOD) glaciations, glacier monitoring, natural hazards, in Nepal as associate expert in GIS and natural remote sensing, and environmental studies. He resource management. From 2008 to 2011, he has carried out fieldwork in Algeria, India, Jordan, worked as a CASIMIR fellow supported by the Lesotho, Mongolia, Pakistan, Peru, South Africa, and Netherlands Organization for Scientific Research Venezuela. xi Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Brandon Krumwiede is a GIS specialist and and streamflow of the upper Indus basin, and project assistant with the National Weather was responsible for the development of all Service’s National Operational Hydrologic SRTM models of the catchment basins as well Remote Sensing Center in Chanhassen, as the shapefile for the Karakoram and western Minnesota. He has a master’s degree in Himalayan glaciers. His MS thesis topic was: geomatics from the University of Montana, where Mapping Glacier Variations from 1990 to 2006 he was a student of Ulrich Kamp. He was a in the Central Mongolian Altai. Prior to returning technical adviser on GIS and satellite imagery to school, Krumwiede was a GIS specialist with for the recent World Bank study on the glaciers Eastview Cartographics, of Minneapolis, MN. xii Abbreviations and Acronyms Abbreviations and Acronyms ADCP acoustic doppler current profiler ALOS Advanced Land Observing Satellite AMSR-E Advanced Microwave Scanning Radiometer for the Earth Observing System AMSU-B Advanced Microwave Sounding Unit ASTER Advanced Spaceborne Thermal Emission and Reflection Radiometer AVHRR Advanced Very High Resolution Radiometer AVNIR-2 Advanced Visible and Near Infrared Radiometer type 2 CAREERI Cold and Arid Regions Environmental and Engineering Research Institute CDMA code division multiple access CIS Commonwealth of Independent States cm centimeter DEM digital elevation model DMSP Defense Meteorological Satellite Program ELA equilibrium line altitude ERS European Remote Sensing ETM+ Enhanced Thematic Mapper Plus FAO Food and Agriculture Organization of the United Nations GCOS Global Climate Observing System GCP ground control point GIS geographic information system GLIMS Global Land Ice Measurements from Space GLOF glacial lake outburst flood GPRS general packet radio service GPS global positioning system GSDQ gauging station data quality HF high frequency HKH Hindu Kush-Himalaya ICIMOD International Center for Integrated Mountain Development IDW inverse distance weighted IGOS Integrated Global Observing Strategy IPCC Intergovernmental Panel on Climate Change IRS Indian Remote Sensing ISO International Organization for Standardization IT information technology km kilometer km2 square kilometer km 3 cubic kilometer LiDAR light detection and ranging m meter m2 square meter xiii Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains m3 cubic meter MGM morphometric glacier mapping MIR mid-infrared mm millimeter mm (we) millimeters water equivalent MODIS Moderate Resolution Imaging Spectrometer MSS Multispectral Scanner NASA National Aeronautics and Space Administration NDSI normalized difference snow index NIR near infrared RMSE root mean square error SAR synthetic aperture radar SNHT standard normal homogeneity test SPOT Système Pour l’Observation de la Terre SRTM Shuttle Radar Topography Mission SSM/I special sensor microwave/imager TIN triangular irregular network TIR thermal infrared TM Thematic Mapper TMI TRMM microwave imager TRMM Tropical Rainfall Monitoring Mission UNESCO United Nations Educational, Scientific and Cultural Organization UNFCCC United Nations Framework Convention on Climate Change USGS United States Geological Survey VNIR near infrared VIR visible infrared VIS visible WAPDA Water and Power Development Authority WMO World Meteorological Organization xiv Executive Summary Executive Summary Effective monitoring of the hydrometeorological There is no hydrometeorological monitoring environment of the Hindu Kush-Himalaya (HKH) “cookbook,” as such, for a region as complex Mountains – the collection of information defining as the HKH Mountains. The first step in planning the climate, hydrology, and glaciers of these a monitoring effort must be a clear statement of mountains – has proven difficult because of purpose and an understanding of the general problems of accessibility, the complex nature of the characteristics of the area to be monitored. mountain environment, lack of conceptual models of the mountain hydrometeorological environment, and Monitoring Objectives inadequate analysis of the existing databases from Hydrometeorological monitoring encompasses a monitoring of the Indus and Ganges River basins of set of activities that characterize the environment India, Nepal, and Pakistan. of the hydrosphere. The development of a credible hydrometeorological monitoring network must be A realistic monitoring program will need to consider approached as a problem in technology transfer that the interactions of climate, glaciers, and stream involves: (a) instrument selection and placement; flow in the Himalaya headwater catchment basins (b) instrument maintenance; (c) data acquisition; as a factor in monitoring network design. To date, (d) data synthesis and digitization; (e) data analysis; a majority of the descriptions of elements of the (f) data storage; (g) user training; and (h) data sharing. HKH hydrometeorological monitoring regime have Further, it will involve development of standard involved traditional “black box” statistical analyses, operating procedures, including: (a) integrated data based on the gross aggregate mean of temperature collection and analysis procedures; (b) funding and and precipitation measured at only a few sites in maintenance responsibilities; (c) personnel training; each basin to forecast lowland water supply. While (d) procedures related to scale and modeling; and this approach has provided realistic data for many (e) ensuring accessibility of monitoring sites. types of lowland water use problems, it is apparent that viewing the headwater catchment as a black box Developing a regional monitoring program in located above the altitudes at which the processes of hydrometeorology faces a number of obstacles: (a) there is no history of any serious, sustained energy and water exchange appear to be maximized collaboration on water resources problems among will provide little guidance in monitoring instrument countries of the HKH region; (b) there is no history placement or in interpreting the data they produce. of serious, continuing, independent field research The challenge is to design a monitoring station in the mountains by scientists of the region; (c) network that is properly located, relatively accessible, although hydrometeorological databases for the and at a scale appropriate for the mountain HKH mountain catchment basins do exist, they are topography, and in which all processes are defined often unanalyzed and unshared; (d) the extreme in credible terms. This report will begin a discussion topography of the mountain catchment basins of of the composition and design of a climate, the major rivers of the region limit accessibility glacier, and runoff monitoring network for the HKH and scale;1 and (e) a generally accepted set of Mountains of India, Nepal, and Pakistan. monitoring procedures does not exist. 1 Accessibility determines the effort needed to reach a particular study or monitoring site, maintain a presence there, and undertake meaningful research. This is a problem in studying the region’s glaciers, which are commonly located in roadless areas at altitudes of 3,000–7,000 meters above sea level. Scale determines the appropriateness of data density, and of both data collection procedures and analyses. xv Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains This study was undertaken at the request of the This report consists of assessments of: World Bank. The purpose of the study was to (a) the nature of the major elements of the assess the current status of major factors related hydrometeorological regime of the mountain to the development of a regional approach to headwaters of the Indus and Ganges Rivers, the management of the water resources of the namely climate, glaciers, and hydrology; mountain headwaters of the Indus and Ganges (b) the primary tools available to monitor these Rivers of South Asia. These factors were identified elements and the current and potential status and as: (a) data availability and sharing; (b) status applicability of these tools, primarily automatic of the current hydrometeorological and glacier weather stations, satellite imagery, geographic observation networks in South Asia; (c) adequacy information systems (GIS), and distributed process of the existing systems to support the assessment hydrological models; and (c) recommendations of climate change implications; (d) modern for improvement of monitoring procedures and hydrometeorological observation systems (ground data management. The contributors to this report and satellite based) and related information have undertaken field studies of aspects of the technology (IT) improvements; and (e) harmonization hydrometeorology, culture, and data management and exchange of hydrometeorological data among in the mountains of South Asia, and are generally riparian states. The geographic scope of the study recognized as knowledgeable authorities in was the headwaters of the Ganges and Indus Rivers their respective fields. This report is intended for in the HKH Mountains in South Asia, encompassing anyone with an interest in monitoring in a high- the mountain arc from eastern Nepal to eastern mountain environment, but is primarily aimed at Afghanistan. two principal audiences: (a) those responsible for planning the future course of hydrometeorological From lists provided by the governments of Nepal monitoring in the headwater basins of the and Pakistan, and limited information from HKH Mountains; and (b) those charged with the literature regarding Indian monitoring, a implementing those plans. total of 493 hydrometeorological monitoring stations were identified in the defined study Establishing a regional hydrometeorological region, of which approximately 90 percent are research facility in the HKH Mountains will involve located below 1,000 meters (m) above sea level. developing solutions in the areas of integrated Recent studies have indicated that the primary data collection and analysis procedures, instrument altitudinal zones of specific runoff (millimeters selection placement, compatibility of monitoring per meter), total runoff volume (cubic kilometers instruments, procedures, and analyses, training (km3)), ice cover area (square meters (m2)), and of personnel, procedures related to scale and glacier ablation zones are generally at altitudes modeling, ensuring accessibility of monitoring sites, of 3,000–6,000 m from eastern Nepal to the and management, analysis, and archiving of the Karakoram. This altitudinal zone should be the acquired data. The major themes of this report are focus of any program of enhanced monitoring the monitoring of climate, glaciers, and streamflow; in the HKH Mountains. While findings based on the appropriate use of the mountain peoples in a statistical correlations between the measured monitoring program; the types of satellite imagery low-altitude climate and the glaciers and that are available to supplement ground-based hydrometeorology of the higher altitudes have activities; and the administrative needs of a credible produced useful results, they may also be a factor monitoring effort. The following are general in some of the more extreme concerns regarding summaries of those themes as reflected in the main climate change and glacier retreat. text of this report. xvi Executive Summary Monitoring of Glaciers, Climate, of existing data should take priority over further development of the climate monitoring network. and Runoff: Main Themes Climate Glaciers There are four ways to address glacier monitoring: Climate, defined here as the long-term trend of (a) direct field measurements and instrumentation meteorological processes determining the water in glacier basins; (b) indirect approaches using and energy balance at a site, varies widely in hydrometeorological data from outside glacier the HKH Mountains. Mean seasonal and annual temperatures may differ by as much as 20–30°C basins; (c) remote sensing; and (d) modeling. A between the low-altitude climate stations now in strategic choice and integration of all four seems use, mean altitudes of a majority of the mountain the best approach. Attempts to derive mass balance catchment basins, and glaciers they contain. estimates and changes in the HKH have been The dominant precipitation source, rainfall in based largely on temperature and precipitation the eastern Himalaya resulting from the summer data extrapolated from weather stations outside monsoon, becomes a mixture of rain and snow the glacier zones, or climate models, sometimes in the western Himalaya of Himachal Pradesh including assumptions about snowlines and and Jammu and Kashmir, and is primarily winter equilibrium line altitudes (ELAs). Conditions known snowfall in the Karakoram Range as the summer to influence mass balance in the HKH but largely monsoon weakens from east to west along the lacking in direct measurements include high- mountain front, and is replaced by winter westerly elevation snowfall, avalanche and wind redistribution lows in the west. of that snow, avalanche-fed glaciers, all-year conditions and cycles in glacier basins, and glacier Temperature monitoring could be improved thermal regimes and movement. Going forward, this by an expanded network of climate stations at will entail regionally appropriate innovation, and not intermediate altitudes (2,000–5,000 m) in the simply relying on greater knowledge and instruments mountain basins. Measurement of precipitation is from elsewhere. Even without the special conditions much more problematic. For snow, the important outlined, any agency or country for which glacier variable is the water content of the winter snow hydrology is required cannot avoid having a setup layer. Under the best circumstances, measuring for continuous engagement with and experience on snow water equivalent depths at remote sites with glaciers. existing instrumentation has proven challenging. The fundamental lesson is that the most reliable Two strategies that are the norm in regions with measurements require an observer at a site to well-established monitoring may not work in the actually measure the water content of each storm HKH: a set of “benchmark” glaciers or a glacier accumulation. Precipitation gauges or pressure network. Both imply mass balance monitoring for pillows can only provide approximations. A few whole glaciers. The former has succeeded mainly accurate measurements of snow water equivalent by choosing small, relatively simple glaciers that by trained villagers at intermediate altitudes in seem, nonetheless, representative for the region. It the mountains would provide much more reliable is doubtful this can work in the HKH. It is here that input to the rainfall–runoff forecast models than an strategic engagement between field and indirect expansion of the precipitation gauge network. approaches is needed. Glaciers would need to be chosen for their suitability for training, ground Only a very limited analysis has been carried out on control, historical reconstruction of glacier change, climate data in the HKH. Acquisition and analysis and experimental efforts. xvii Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Caution is urged with respect to more expensive, Many of the stream flow data for HKH mountain state-of-the-art instrumentation and techniques. basins are not readily available, either as a result of Some are attractive for the high mountains and can a formal policy, as is the case in India, or due to a overcome difficulties found in the HKH. However, general lack of procedures for data management, the region is littered with “advanced” setups and as in Nepal and Pakistan. Until this data access devices that are broken or were quickly deemed problem is resolved, many of the analyses of the inappropriate, or could not be maintained with hydrometeorology of the mountain basins will be local resources. Working with simpler and well-tried based on extrapolation of lowland stream flow and methods is often more reliable, and a better basis for climate records. The challenge is development of training and building glacier experience. realistic extrapolation procedures for the extreme In each country, direct observations will require one three-dimensional topography of those basins. or more teams trained and permanently ready to work in glacierized areas. None of this is likely to In the eastern Himalaya, runoff is produced primarily happen or be successful without a core of personnel by rainfall associated with the summer southeast experienced in mountain environments, usually with monsoon. In the Karakoram, in the extreme western mountaineering and winter skills, and enthusiastic portion of the mountain chain, stream flow results about the work in which they are engaged. However, primarily from the summer melt of the previous this will not happen without addressing important winter snowpack, with the addition of a glacier melt and special problems of safety, equipment, and component. A comparison of mean basin-specific training. runoff (in millimeters) with mean basin altitude (in m) shows a curvilinear trend of runoff depth with Hydrology altitude, with specific values reaching a maximum at intermediate altitudes, and minima at the altitude For the purposes of this discussion, mountain extremes. This suggests that any expansion of the hydrology is defined as the methodologies hydrometric monitoring network should be focused associated with the monitoring and measurement on the zone between 3,000 and 5,000 m above sea of the water balance of the catchment basins of level, with a particular emphasis on the hydrology the HKH Mountains. Traditionally, hydrological of the glacier and periglacial environments. monitoring undertaken for purposes of water While the annual hydrographs of runoff from the resources planning or management has been eastern and western portions of the HKH are very based on “rainfall–runoff” or “black box” similar, with maximum values occurring during the correlation modeling, in which input, as measured summer months, it is apparent that the underlying precipitation, is correlated with output, as hydrometeorological processes are quite different measured streamflow, to provide an estimate of the between the two. timing and volume of streamflow from a basin. This type of modeling produces very useful information Indigenous Monitoring for engineers and water managers concerned with the lowland rivers originating in the mountain In terms of existing glacier-related research activities, basins. This modeling approach, however, provides there is an absence of explicit involvement and relatively little insight into questions concerning the participation of mountain communities in monitoring relative contribution of rain, snow, or glacier melt and assessing change. Proposals to scientifically to streamflow volumes, or the role of glacier retreat monitor glaciers, weather, and environmental and climate change in the streamflow regimes of changes in ways that directly involve people living in the major rivers of South Asia, a major topic of an upper basin catchments have not been advanced in ongoing debate. the region. xviii Executive Summary Indigenous glacier monitoring goals in terms of or snow cover over much larger areas such as an selected impact indicators should be related to the entire mountain range. needs of the scientific community, glacier hazard risk reduction, vulnerabilities and capacities of Mesoscale Imagery mountain communities, awareness and knowledge of the public, evidence-based adaptation planning, Until the early 1970s, aerial photography was and prioritization of community objectives. the primary remote sensing technology in glacier Intermediate outcome indicators could be targeted mapping and monitoring. Although this technology for achievement in the next three to five years has many advantages, it also has many restrictions, and could include the following: (a) training for example, in the extent of ground coverage, and participation of local villagers as research its availability for many study areas such as the assistants and technicians; (b) increased scientific HKH region, and high costs of aircraft and flight knowledge among the mountain-based population; campaigns. This led to the introduction of satellite (c) implementation of new curricula in glaciology, imagery analysis in studies of the cryosphere. mapping, data analysis, hydrology, and hazards Since the early 1970s, medium-resolution planning at regional institutes of higher education; (10–90 m) optical satellite data have become and (d) creative and innovative solutions to reduce available, particularly with the launch of sensors risks and hazards. such as Landsat Multispectral Scanner (MSS), Landsat Thematic Mapper (TM), Système pour Satellite Imagery and GIS l’Observation de la Terre (SPOT), Indian Remote Sensing (IRS) including Cartosat and Resourcesat, Satellite imagery is becoming increasingly available Landsat 7 ETM+, ASTER, and Advanced Land for use in studies of elements of the water and Observing Satellite (ALOS). Today, large-scale (less energy characteristics of high-mountain basins, such than 10 m) imagery suitable for detailed glacier as glacier and snow cover extent. A basic problem studies at basin scale is available from, for example, involves the need to ensure that the scale of the IKONOS, Quickbird, and GeoEye-1. However, the imagery is compatible with the scale of the variable narrow swath, long revisit cycles, and high costs limit being studied. Imagery scale is defined by the size its use for systematic glacier monitoring of larger of pixels (a physical point in an image, the smallest regions such as the HKH. CORONA data from controllable element of a picture represented in 1960 to 1972 were declassified in 1995 but are the image). Landsat 7 Enhanced Thematic Mapper only available for some glacierized areas within the Plus (ETM+) and Terra Advanced Spaceborne HKH region. Thermal Emission and Reflection Radiometer (ASTER) images with a pixel size of 15 m are considered Some of the potential datasets that can be obtained here as mesoscale, or intermediate scale, and through satellite imagery, digital elevation models are appropriate for measurements at the scale of (DEMs), and GIS include elevation values, glacier individual glaciers or mountain basins. Moderate hypsometry, basin hypsometry, glacier longitudinal Resolution Imaging Spectrometer (MODIS), profiles, glacier ELAs, slope, surface curvature, Tropical Rainfall Monitoring Mission (TRMM), and aspect, and surface roughness. Through the use of Advanced Very High Resolution Radiometer (AVHRR) satellite imagery, DEMs, and GIS-derived datasets images have a pixel size of about 250 m, and are in combination with empirical measurements, it will defined here as macroscale imagery, which is more be possible to develop a better understanding of the appropriate for measuring the extent of the glacier HKH environment. xix Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Macroscale Imagery discharge data in the upper Indus suggest that TRMM estimates provide a quantitative index of MODIS monthly precipitation rather than a measure of absolute magnitude. In this, they are similar to the MODIS instruments of the National Aeronautics and local long-record meteorological observations, Space Administration (NASA) capture data in which also do not directly represent catchmentwide 36 spectral bands ranging in wavelength from 0.4 to precipitation but do correlate well as indicators 14.4 micrometers and at varying spatial resolutions. of mass inputs for seasonal snowmelt-driven They are designed to provide measurements of catchments. large-scale global dynamics, including changes in Earth’s cloud cover and radiation budget, and AVHRR processes occurring in the oceans, on land, and in the lower atmosphere. NASA software extracts AVHRR is a radiation detection imager that can time series datasets with given resolution and be used for determining cloud cover, surface time averaging from specific sensed wavelengths. temperature, and snow cover extent. Although Records are available from early 2000 to the with more limited spectral resolution, this long present. A major problem encountered in the use record offers the potential to: (a) greatly increase of MODIS imagery in mountain terrain is the lack the overlap of the spatial data products with of correspondence between the spatial scale of the local observations, thus refining quantification of MODIS image and the scale at which processes of relationships between them; and (b) extend the water and energy exchange vary over the surface of range of observations by MODIS to better capture the mountain basin. present spatiotemporal climate variability. TRMM DEMs and Geomorphometry TRMM high-resolution observations provide indirect DEMs are digital representations of the Earth’s data on rainfall through a correlation between surface. In glacier monitoring, they are required for rainfall depth and lightning frequency. The TRMM image orthorectification and radiometric calibration, specific observations are merged with additional debris-covered glacier mapping, surface energy passive microwave observations from several other balance studies, glacier ice volume loss and mass satellite-borne instruments (such as special sensor balance estimates, glacier hypsometry, and ELA microwave/imager (SSM/I), Defense Meteorological estimation. DEMs are generated from digitized Satellite Program (DMSP), Advanced Microwave topographic maps, satellite stereo-imagery (for Scanning Radiometer for the Earth Observing System example, ASTER, IRS, SPOT), and data derived from (AMSR-E), and Advanced Microwave Sounding radar interferometry (for example, Shuttle Radar Unit (AMSU-B)) as well as the near-continuous Topography Mission (SRTM), TerraSAR-X) and laser low-resolution infrared and thermal imagery from altimetry (for example, light detection and ranging geostationary weather satellites. (LiDAR)). Digital terrain modeling is a complex process involving acquisition of source data, TRMM may provide reliable quantitative estimates interpolation techniques, and surface modeling; of summer monsoon convective rainfall, but the in addition, there is the need for quality control, application to orographically enhanced winter including accuracy assessment (overall planimetric snowfall from westerly systems may prove more and vertical accuracy), data management, problematic. Comparisons with available local interpretation, and application. Accurate glacier long-record observations of precipitation and river assessment using topographic information is xx Executive Summary frequently an issue of DEM quality. Great care is Depending on local conditions, logistic required to ensure that the data selected for an arrangements are made by the executing entity or application are appropriate, processing is carried with the assistance of a well-established trekking out with a high level of expertise, and errors in agency as partner. Finding human resources (such any derived data are accurately reported, so that as porters) for logistic support has recently become real geophysical patterns and features can be difficult and more expensive because, at least for differentiated from image and processing artifacts. the conditions in Nepal, large numbers of younger people are migrating out of the country for better Geomorphometry is defined as the science of job opportunities. In general, field visits and station quantitative land surface analysis and draws from maintenance need to be undertaken using local mathematics, computer science, and geosciences. facilities and possibilities. For cost-effectiveness and The field of geomorphometry has two modes of sustainability of the installed infrastructure, it is not study: the study of individual or specific landforms advisable to leave all observations, maintenance, and features (for example, glaciers), and study of and station surveillance to office staff back in the the general land surface or region (for example, the city but, to the largest extent possible, delegate such Himalaya). Glacier mapping usually includes the functions to locally available staff, who may take geomorphometric analysis of the glacier surface, great pride in doing these works if their services and most software packages include relevant tools. are adequately recognized and acknowledged. For However, as much as geomorphometric parameters example, after the end of a project, helicopters for help in identifying, describing, and classifying station supply are not a realistic and sustainable glaciers, their quality depends on the accuracy of the option for a government organization or any input DEM. other locally operating entity. It is essential for the monitoring programs to have local, well-trained Requirements for Instituting technical personnel to reduce travel and mission a Monitoring Program costs and time lags in reaching a station after a problem has occurred. This is technically feasible Carrying out monitoring programs in high-mountain through adequate capacity-building programs that areas, and especially in the HKH region, have been enable local personnel to perform essential technical challenged by the rough environment, insufficient functions based on well-defined, station-specific funding of continued, long-term observation standard operating procedures. programs, weak institutions, and difficulty of dispatching government officials on a regular basis As a lesson of past projects, an agreed data policy to high-mountain regions on missions that often last needs to be developed covering the different several weeks. Also, under current civil service rules data streams that the project will establish. Such in all three countries, there are no special provisions data will have a multitude of origins, mostly for extra allowances that would make it attractive to however from national sources (such as national staff to work under harsh conditions. hydrometeorological networks). xxi xxii Introduction 1. Introduction Hydrometeorological monitoring, as discussed here, this zonation (together with the questions concerning describes the activities required to characterize these controls), installation and maintenance of the properties and processes of the hydrosphere instruments, balance between “ground-truth” and as it exists in the three-dimensional mesoscale remote sensing, storage and digitization, and environment of the high-mountain catchment basins analysis and sharing must be defined before the of the Hindu Kush-Himalaya (HKH) Mountains. nature of the monitoring network can be specified Credible monitoring involves: (a) functional with any confidence. institutions; (b) operational instruments; (c) trained, motivated individuals; (d) scientific procedures; Much of the literature describing the hydrology and (e) dedicated funding. Establishing a regional and glaciology of the Himalaya is in the form of hydrometeorological research facility in the HKH “snapshots” from single or discontinuous visits to Mountains will involve developing solutions in the particular locations or is based on the extrapolation areas of integrated data collection and analysis of lowland, gross aggregate databases. There procedures, instrument selection and placement, are few continuous records other than low- compatibility of monitoring instruments and altitude streamflow and climate data for the procedures, training of personnel, procedures hydrometeorological or glaciological environments related to scale and modeling, ensuring accessibility of these mountains, and still fewer models that would of monitoring sites, and management, analysis, and permit the synthesis and analysis of these data, many archiving of the acquired data, all in the context of which are not readily available. Of necessity, of processes within the mountain basins, not in the much of the literature is speculative and based on adjacent lowlands. relationships developed from other mountain regions in Asia, Europe, and North America. An excellent Mountain hydrometeorology is defined by a exception is Bruijneel and Bremer 1989, which set of complex, three-dimensional, biophysical provides a general overview of the hydrology of environments, produced by interactions among the mountain basins of the Ganges River, based on terrain, geology, and meteorology. The homogeneity analyses of studies of those basins. seen from the distant lowlands becomes a complex mosaic of environments within the headwater basins. A realistic monitoring program involving climate, Altitude determines the properties of an atmospheric glaciers, and streamflow in the Himalaya headwater column extending upwards from a point within the catchment basins will be needed to tie the three mountains. These atmospheric properties determine elements together as a factor in network design. To the potential water and energy budgets at a point, date, a majority of the descriptions of elements of or within a basin, in the mountains. Relief – slope the HKH hydrometeorological regime have involved aspect and angle – defines local topography. traditional “black box” statistical analyses based These terrain properties, in turn, create the three- on the use of gross aggregate means of mountain dimensional spatial mosaic of water and energy processes of water and energy exchange. While budgets that characterize mountain catchment this approach has provided realistic input into basins – the mountain topoclimatology. Controls on many types of lowland water use problems, it is 1 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains apparent that a view of the headwater catchment This situation has changed dramatically in the past as a black box, located above the altitudes at decade as mountain glaciers have become an which the processes of energy and water exchange icon of climate change and anthropogenic global appear to be maximized, will provide little guidance warming. The result was a sudden interest in glaciers in monitoring instrument placement, or in the at all levels, from sensational stories in the media interpretation of the data produced by those to the now-discredited statement that the glaciers of instruments. The challenge is to design a monitoring the Himalaya would be gone by 2035 (IPCC 2007, station network that is properly located, relatively chapter 10.6). It is now recognized that these initial accessible, and in which all processes are defined statements and resulting alarms were made possible in credible terms. This report is intended to begin a by the lack of hard data on which to base credible discussion of the realistic composition and design of assessments of the role of climate change in the a climate, glacier, and runoff monitoring network for growth and shrinkage of glaciers, in addition to a the HKH Mountains of India, Nepal, and Pakistan. general lack of familiarity, among both technical and nontechnical publics, with mountain environments 1.1 History in general and specifically the very high mountains of Asia, and analyses of existing databases by only Traditionally, the study of mountain climates, a handful of concerned, qualified scientists. The glaciers, and hydrology has been the province global climate is changing and the global ice cover of small groups of scientists working with limited is shrinking, but the societal impacts of this change funding, in relative obscurity, within the context of on the hydrometeorological environments of the conceptual and theoretical frameworks provided Himalaya, particularly on the timing and volume of largely by studies from the adjacent lowlands. the flow of the major rivers with headwaters in those This has led to the development of two separate mountains, remain uncertain. approaches to the study of what might be termed “mountain science”: (a) development of concepts All the major rivers of the region have headwaters and models of the mountain environment based in the HKH mountain ranges. The fundamental on gross aggregate means of a range of elements, challenge is to assess the quantitative importance as measured in the adjacent lowlands, or at a few of glaciers in determining the annual volume and low-altitude sites within the mountain basins, or timing of rivers with headwaters in mountains at a few, limited sites within selected basins; and containing glaciers. Recent studies have suggested (b) a fragmented collection of studies, scattered that this importance may vary widely in the HKH throughout the scientific literature, describing Mountains. The primary problems stem from the primarily results of site-specific climate, glacier facts that (a) climate change is a complex process, mass balance, and water budget studies within the most probably involving both water and energy mountains. A majority of these latter studies are input and output, and does not respond solely to from the European Alps and the mountain ranges temperature fluctuations; and (b) the mountain of North America. This has resulted in a specialized climate is a result of a complex interaction between literature that traditionally has been of immediate the mountain topography and the surrounding interest to a relative handful of individuals, and atmosphere (or lack of it), involving a three- that may or may not have relevance to dimensional mosaic of topoclimates, defined the hydrometeorological environments of primarily by local variations of altitude, aspect, the Himalaya. and slope. 2 Introduction Perhaps more importantly, a reliable analysis of • The extreme topography of the mountain HKH mountain hydrometeorology is becoming catchment basins of the major rivers of the increasingly relevant in the context of the economic region complicates all attempts to study the (and population) growth being experienced by problem. The extremity of the topography is the countries of South Asia. At the same time, characterized by the factors of accessibility and procedures for undertaking monitoring or research scale. (Accessibility determines the effort needed within the high-altitude mountain basins of the HKH to reach a particular study or monitoring site, have evolved slowly, primarily on an ad hoc basis for maintain a presence there, and undertake each project, with limited formal documentation in meaningful research. This is a problem in the technical literature. studying the region’s glaciers, which are commonly located in roadless areas at altitudes 1.2 The Problem of 3,000–7,000 meter (m) above sea level. Scale determines the appropriateness of data density, Recent concerns about climate change and and of both data collection procedures and retreating glaciers and their effects on river flows, analyses.); and specifically in the Himalaya, have illustrated • There is currently no generally accepted set of how little the scientific and water management “best practices” for conducting monitoring or communities know about the role the mountain research of the mesoscale hydrometeorology of headwaters play in the annual flow of the major river large mountain ranges such as the HKH. systems of Asia. The water available from these rivers determines the supply use problems the region’s Given current concerns related to glaciers, climate, countries now face and has caused concerns over and rivers, a monitoring program might establish the the future availability of water resources for all uses following objectives: (a) to either measure directly in the countries of South and Central Asia. The or develop credible estimation techniques for the solution must include an understanding of how annual cycle of mass gain and loss of the mountain important the Himalayan glaciers are as a source glaciers; (b) to develop analytical methodologies of the major rivers of the region by defining the to link mass balance fluctuations of these glaciers role of glaciers as a component of the hydrological to the regional and global climate cycles; and (c) cycle of the mountain basins. The countries of South to link these fluctuations to variations in the timing Asia face a number of challenges in developing a and volume of the rivers flowing from the mountain credible hydrometeorological monitoring program. catchments. They are as follows: 1.3 Scale and Location • Countries of the HKH region have no history of any earnest and sustained collaboration on water While satellite-derived data are becoming resources problems; increasingly important in the study of the • A history of independent field research in the hydrometeorology of the mountains of Asia, it is mountains among scientists of the region is essential to use imagery with a resolution compatible lacking; with that of the glaciers, snowfields, and catchment • Databases related to these problems exist in basins of these mountains to produce realistic each of the HKH countries, but are generally results defining the interrelations and interactions unorganized, unanalyzed, and, most importantly, of properties and processes in the mountain basins. unshared; Development of procedures for testing and ensuring 3 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains the accuracy, reliability, and reproducibility of the influences will produce local variations in factors various types of satellite imagery now in use, or such as the equilibrium line altitude (ELA) or the proposed, should be a priority. ablation gradient among the glaciers in the basin. These differences in the scale of the glaciers, and In recent years, macroscale satellite imagery and of the macroscale tools currently used for most global-scale climate models have been the primary glacier monitoring in the HKH Mountains, make it tools of climatologists studying the effect of climate virtually impossible to reach a consensus view of the change on mountain climates, glaciers, and relationship between climate change, glaciers, and streamflow. This has necessitated the introduction of streamflow, as it will vary from basin to basin. procedures for generating regional components of climates – at the macroscale of the global circulation The primary locations of an empirical study of the model, with grid spacing of up to 50 kilometer (km) glacier hydrology of a mountain basin in the HKH or a pixel resolution of 500 m – that can then be Mountains are those of the hydrometric stations and tested statistically against monthly or annual values of the glaciers that are separated horizontally by tens of runoff volume. This is basically a continuation of kilometers and vertically by thousands of meters. of the traditional approach to studies of mountain At the terminus of the glacier, most streamflow is hydrology, involving a lumped parameter, or “black a result of glacier melt. As distance downstream box,” approach. In considering scale as a factor in increases, the glacier contribution will be diluted by hydrological studies, it has been argued that: other sources of input, such as snowmelt or rainfall. Seen in this context, it is easier to understand why [L]evels of scale at which a meaningful recent statements that some fixed percentage of the conceptualization of physical processes is volume of flow of a given river is the result of glacier possible are not arbitrary and their range is melt may contain substantial error. In order to test not continuous. Formulations appropriate at the hypothesis that the current glacier retreat that a given level usually are not applicable at is occurring in the eastern Himalaya is a result of the immediately adjoining levels. This is seen climate change, it will be necessary to demonstrate as one of the important reasons for the slow a correlation between the two that will in addition progress of hydrological science on basin explain the current advance of the glaciers of the scale (Klemes 1983). Karakoram Range, in the western HKH region. In general, the processes controlling water and energy One of the implicit arguments for the use of exchange in mountain basins are operating at the macroscale satellite imagery and global climate mesoscale or intermediate-scale level, controlled by models has been a lack of suitable topographic topographic elements of slope, aspect, and altitude. maps for much of the HKH region. Development Glaciers originate in favored mountain basins, or of digital elevation models (DEMs) based on “cirques,” where accumulation, as snow, is maximized geographic information system (GIS) principles for by wind drifting or avalanching of snow, and ablation, any portion of the HKH Mountains is now a relatively or melt, is minimized by shielding from radiation straightforward matter, and can be a part of by terrain or the low temperatures associated with hydrological or glaciological modeling or monitoring high altitudes. These mesoscale aspects controlling in mountain basins. It has been recognized for glacier growth and shrinkage are not captured by some time that knowledge of local conditions either macroscale satellite imagery or global-scale of topography and meteorology is necessary to climate models. These local mesoscale topographic understand the basin-scale hydrology of mountains: 4 Introduction [M]ountain hydrology modeling makes surface area in Nepal and the upper Indus basin; painfully obvious … the importance of and areal mapping of hydrological and other • Regional variations in glacier melt as a geophysical variables and the inadequacy component of streamflow. of the traditional point measurements which are the legacy of the century old technology The objectives of a climatological and hydrological (Klemes 1990). monitoring network are as follows: 1.4 Objectives and Procedures • Provide sufficient information to determine areal averages of moisture and energy inputs The primary objective of a monitoring program and outputs and storages (water and energy is to develop general procedures for undertaking balances) at a range of space scales and time assessment and monitoring of the components intervals; of the water and energy budgets of the mountain • Provide a sufficient duration of record to assess catchments of the Himalaya. The two basic trends and periodicities in climatic variables assumptions of the approach discussed in this relating to moisture and energy. Data on trends report are: may be required over different time intervals, including annual, seasonal, monthly, and • While any improvement in the existing monitoring extremes. The shorter the time interval, notably for network of climatological and hydrometric extremes, the more the demand on the network stations would be a positive step, the most useful and monitoring requirement; and informative results will be obtained from • Provide data on which day-to-day resource stations established in some defined relationship management decisions may be based; with the existing station locations, and justified • Provide sufficient climatic information to assess in the context of existing analyses and models of trends and year-by-year changes in glacier the three-dimensional Himalayan environmental mass balance, depending on moisture and matrix; and precipitation inputs and energy and melt outputs; • All data collected by this program must be and available to international scientific and technical • Provide sufficient information to assess year- communities. by-year and decadal variations and trends in the comparative contribution to streamflow The value of any particular number and location from glaciers, seasonal snowmelt, and liquid of instruments in a monitoring network, the mix precipitation. of ground-based instruments, and the regional data collected from satellite imagery can only The density of the network required to satisfy these be determined from the extent to which the data needs depends on the accuracy with which individual obtained provide answers to major questions measurements can be made (a particular problem related to the system being monitored. Examples of with snow measurement), the spatial variability of relationships in the hydrometeorology of the HKH the variable (how spatially correlated it is within the Mountains that could be better defined are: region), whether the variations are systematic (for example, through lapse rates of temperature), and • East–west variation in stream flow components; the time interval over which the balance is required • Recent glacier retreat and temperature increases; (for example, a denser network for flood forecasting • Seasonal sources of monthly streamflow; than for glacial mass balance studies or water • Significance of area–altitude distribution of resources management). 5 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains References Panel on Climate Change (M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden, and Bruijneel, I., and C. Bremer. 1989. Highland- C.E. Hanson, eds.) Cambridge: Cambridge Lowland Interactions in the Ganges Brahmaputra University Press. River Basin: A Review of Published Literature. Klemes, V. 1983. “Conceptualization and Scale ICIMOD Occasional Paper 11. Kathmandu, in Hydrology.” Journal of Hydrology 65 (1–3): Nepal: International Centre for Integrated 1–23. Mountain Development. Klemes, V. 1990. “The Modeling of Mountain IPCC (Intergovernmental Panel on Climate Hydrology: The Ultimate Challenge.” In Change). 2007. Climate Change 2007: Hydrology of Mountainous Areas, IAHS Impacts, Adaptation and Vulnerability. Publication 190, International Association of Contribution of Working Group II to the Fourth Hydrological Sciences, Saskatchewan, Assessment Report of the Intergovernmental Canada. 6 Climate Monitoring 2. Climate Monitoring Climate and climate change have always played a In the last 20 years, the retreat of glaciers in the central role in the culture and economy of the Indian central Himalaya has generated heated debate on subcontinent. The great early Indus civilization of the the role of anthropogenically driven climate change. third millennium BCE, with centers at Harappa and Uncertainty has arisen primarily owing to the Mohenjo-Daro, collapsed towards 1800 BCE and its inadequacy of climate, glacier, and streamflow data, cities were abandoned, almost certainly as the result due to the difficulty of ground-based measurements of climate change (Wood 2007). In the Mughal at high elevations where climate has most impact era, the Emperor Akbar’s magnificent palace and on glacier mass balance, snowmelt, and streamflow. capital at Fatehpur Sikri was abandoned, probably In addition, political sensitivity over shared water as a result of an unanticipated shortage of water. In resources has restricted the dissemination of climate the 19th century, the British Raj used information on data by national agencies. rainfall and its impact on agricultural productivity as a basis for taxation. Climate measurement, In the following text, network design is discussed therefore, became a part of administrative duty, in relation to high-altitude environments, how and rainfall records exist from the middle of the such networks have been typically used in previous 19th century. The India Meteorological Department analysis, and the relationship between topography was founded in 1875, and, in 1891, the system of and climatic variables. The use of historical climate monitoring and reporting was formalized. climate data and the necessity of assessing the Measurements include temperature and precipitation homogeneity of such data for climate change studies – principally used for climate change analysis – are discussed. Measurement methodologies are and wind speed, humidity, cloud cover, sunshine, considered in relation to the extreme environment. and, more recently, radiation, all of which are now Finally, recommendations are made with respect to considered important variables for understanding monitoring, instrumentation, and analysis. climate change. 2.1 Monitoring Objectives The climate of the mountain fringe was considered important when hill stations were developed for The objectives of a climatological and hydrological administrators as mountain retreats from the summer monitoring network include the following: heat of the plains. The headquarters of the India Meteorological Department was based at such a • Information to determine areal averages of station, Shimla. Himalayan climate had been an moisture and energy inputs and outputs and early subject of scientific study in its own right (Hill storages (water and energy balances) at a range 1881) and as a basis for forecasting the monsoon of space scales and time intervals; (Blanford 1884). The Himalaya is believed to play a • Sufficient duration of record to assess trends significant role in building the land–ocean thermal and periodicities in climatic variables relating contrast that helps drive South Asian summer to moisture and energy. Trends may be required monsoon rainfall (Douville and Royer 1996). over different time intervals, including annual, 7 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains seasonal, monthly, and extremes (5th and 95th given spatial accuracy (thus a greater density for percentiles). The shorter the time interval, notably convective than for frontal precipitation). Space for extremes, the more demanding the network scale is also important in the required density; at and monitoring requirement; small scale, local factors such as orientation, slope, • Climatic information to assess trends and year-by- soils, geology, and glacier cover assume greater year changes in glacier mass balance, depending importance than in large basins where local factors on moisture and precipitation inputs and energy are averaged. In other contexts and environments, and melt outputs; and Bleasdale (1965) has suggested that the density • Information to assess year-by-year and decadal of gauges to define monthly estimates of average variations and trends in the comparative areal rainfall varies by a factor of 25 for catchments contribution to streamflow from glaciers, seasonal between 25 and 7,500 square kilometers (km2). snowmelt, and liquid precipitation. 2.2 Previous Network Design The density of the network required to satisfy these Recommendations needs depends on the accuracy with which individual measurements can be made (a particular problem Previous consideration in the literature is with snow measurement), the spatial variability of concentrated on the design of precipitation networks. the variable (how spatially correlated it is within the The discussion below focuses on precipitation but is region), whether the variations are systematic (for relevant to network design for other variables. The example, through lapse rates of temperature), and World Meteorological Organization (WMO) (2010) the time interval over which the balance is required gives a general guide to the density of precipitation (for example, a denser network for flood forecasting stations required for different physiographic units, as than for glacial mass balance studies or water abbreviated in Table 2.1. resources management). In general, the more variable the areal distribution Starting from the simplest theoretical variable of precipitation, as in convective storms or in that is uniform throughout a region, only a single mountainous areas, the more gauges are needed point measurement would be necessary to define to give an adequate sample. Thus, because of the the spatial average of the variable. As a second perceived variability of mountain precipitation, the stage, a variable that has a defined and unique recommended density is more than double that systematic relationship over a range of geographic required in other regions. However, actual densities physical features such as altitude could still require are much lower than those recommended. For only a single measurement. The use of a single Nepal, with 281 rainfall stations on an area of station to define certain aspects of the hydrology of 147,000 km2, the density is 523 km2 per gauge, large Himalayan catchments is not as outrageous while for the upper Indus and Jhelum in Pakistan, as it may seem at first. In some meteorological with an area approaching 200,000 km2, the density settings, the measurements of a single station may is about 5,000 km2 per gauge. Of course, in be sufficient to characterize the system. However, a both cases, the gauges are concentrated at lower large network of stations is still required to establish elevations, with limited representation at those levels such relationships. where most precipitation occurs and where most flow is generated from melting of snow and glaciers. With increasing spatial variability, the number In the Himalayan region, existing methods for of point measurements required increases for a network design and the creation of gridded climatic 8 Climate Monitoring Table 2.1 Recommended Minimum Density of Precipitation Stations Physiographic unit Minimum density (km2/gauge) Nonrecording (daily) Recording Coastal 900 9,000 Mountainous 250 2,500 Interior plains 575 5,750 Hilly/undulating 575 5,750 Small islands 25 250 Urban areas 10–20 Polar/arid 10,000 100,000 Source: WMO 2010. databases are seriously limited. The ground-based to the point of measurement but also over a wide climate network currently in place to validate area and in elevation zones in which measurements models and interpolation methods is sparse and are limited or entirely absent. This conclusion is unrepresentative. The study in this report was based on an analysis of the spatial correlation of the designed to specify where and how such limitations climate variables of precipitation and temperature can be overcome. While some global datasets in the upper Indus basin in Pakistan (Archer 2004; are important inputs to regional climate models, Archer and Fowler 2004; Fowler and Archer 2006) none of the globally gridded data are sufficient to and on links to runoff (Archer 2003; Archer and capture key elements of meteorological variables Fowler 2008). in the Himalaya. For example, data of the Climate Research Unit, the ERA model of the European 2.3.1 Temperature Centre for Medium-Range Weather Forecasts, or the Modern-Era Retrospective Analysis for Research and Thus, for example, with respect to temperature Applications of the National Aeronautics and Space from April to June, station seasonal temperatures Administration (NASA) do not capture the distinct are significantly correlated across the entire region, orographic rainfall bands that have been identified stretching 340 km east to west and 200 km north with higher-resolution data, for example, from the to south. At opposite ends of the upper Indus TRMM or from Asian Precipitation–Highly Resolved basin, Chitral (1,499 m) and Srinagar (1,587 m), Observational Data Integration towards Evaluation which have major mountain barriers in between, of Water Resources. have a correlation coefficient of 0.82. It seems reasonable to suggest that the correlation applies 2.3 Use of Climate Networks also to higher elevations; limited data for Kunjerab (4,730 m) and Shandur (3,750 m) appear to Despite the limitations discussed, there is clear support this suggestion. The correlation analysis evidence that the existing climate network in the has been repeated for other seasons. The winter Himalaya can provide some useful guidance on period, October–December, shows lower correlation trends in climatic variables that are relevant not only coefficients and may be influenced to a greater 9 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains extent by local temperature inversion effects at which precipitation measured at valley floor level valley stations. The station at Astore (2,394 m) has (where annual precipitation is generally less than the most consistent correlation with other stations 200 millimeter (mm)) is representative of variations for both the October–December and April–June and trends at higher altitudes (where precipitation periods. may reach 1,500 mm) (Wake 1987). Again, the evidence lies in correlation between precipitation What further evidence is there that variations in measured at valley level and the runoff from low-elevation valley temperatures are representative catchments where the flow originates from snow of variations at high elevations? The evidence lies accumulated during the previous winter season. not just in correlation with high-level measurements Figure 2.2 shows the relationship between winter of temperature (which are limited) but also in precipitation at Astore and seasonal runoff for the correlation with runoff from catchments where the Astore at Doyien (mean catchment elevation flow originates from melt of high-level permanent 3,921 m) and the Jhelum at Kohala (mean snowfields and glaciers (Archer 2003; Fowler and catchment elevation 2,629 m) stations. Archer 2006). Figure 2.1 shows the relationship between temperature at valley stations at Gilgit and The use of low-level precipitation as an index of Skardu and seasonal runoff for the Hunza and Shyok precipitation at higher elevations does not in itself catchments, both with mean catchment elevations provide knowledge of the actual precipitation at over 4,400 m. higher elevations. Such information is essential, for example, in assessing year-to-year changes 2.3.2 Precipitation in glacier mass balance. Some guidance on the progress and retreat of snow cover is provided A problem similar to the correlation problem with by remote sensing observations of snow-covered temperature emerges with respect to the extent to area from the NASA Moderate Resolution Imaging Figure 2.1 Seasonal Temperature and Runoff, June–August, at Two Locations in Pakistan a. Gilgit and Hunza River at Dainyor Bridge b. Skardu and Shyok River at Yugo 30.0 26.0 y = 0.006x + 22.836 y = 0.0253x + 19.923 29.0 R2 = 0.6025 25.0 R2 = 0.5257 Seasonal Mean temperature (°C) Mean Temperature June-August 28.0 24.0 Skardu June-August 27.0 23.0 26.0 22.0 25.0 21.0 24.0 20.0 23.0 19.0 22.0 18.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 50.0 100.0 150.0 200.0 Seasonal runoff (mm) June-August Seasonal runoff (mm) June-August Source: Fowler and Archer 2006. 10 Climate Monitoring Figure 2.2 Preceding Seasonal Precipitation (October–March) at Astore and Runoff (July–September) at Two Locations a. Astore River at Doyien b. Jhelum River at Kohala 1000 700 Astore at Doyien Runoff (mm) Jul-Sep y = 1.181x + 279.08 y = 0.705x + 220.55 900 R2 = 0.5257 600 R2 = 0.4355 800 700 500 Runoff at Kohala 600 400 500 300 400 300 200 200 100 100 0 0 0 100 200 300 400 500 0 100 200 300 400 500 Astore Precipitation Oct-Mar Astore Precipitation Oct-Mar Source: Archer and Fowler 2005; Archer and Fowler 2008. Spectrometer (MODIS) instrument. However, actual and topographic variables including elevation, measurements of precipitation are quite inadequate slope, orientation, and exposure. Such analysis is not (or not made available) above 2,500 m and new. Before the computer era, graphical methods virtually absent over 5,000 m in the northwestern were used to demonstrate that rainfall in the Rocky Himalaya. In Nepal, in the Khumbu valley, a high- Mountains in the United States of America was elevation network of six automatic weather stations influenced by elevation, slope, rise, orientation, and has been installed over the past 10 years, which exposure, which together accounted for 85 percent measure total precipitation as well as other standard of variation. Establishment of such relationships meteorological variables.2 They range in elevation using either ground-based or remotely-sensed data from 2,660 to 5,050 m above sea level near the may provide key links to a realistic assessment of Pyramid Laboratory-Observatory. However, the total areal averages of variables in rugged mountain topography. Existing evidence is given here mainly number of such high-level stations is still inadequate. with respect to precipitation and temperature in the HKH. Other components of energy exchange are 2.4 Environmental Features considered briefly below. Affecting Variations in Climate and Glacier Mass Balance 2.4.1 Precipitation Lapse rates of temperature with altitude are a The general perception is that orography provides fundamental physical aspect of climate. Basista, the necessary uplift for moisture-laden currents Bell, and Meentemeyer (1994) developed statistical striking against a mountain range, resulting in relationships between mean annual precipitation copious rainfall, mainly on the windward side and 2 EVK2CNR website: http://www.evk2cnr.org/cms/en/home.html. 11 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains increasing with altitude. However, in high-mountain even approximately in balance) that significantly areas, orographic uplift and convergence is higher precipitation occurs at higher altitudes. balanced with increasing altitude by the decreasing Nevertheless, the evidence base is still very weak, capacity of the atmosphere to hold moisture at based on very limited time series or locations, and it lower temperatures and the smaller remaining requires strengthening. depth of the air column. Thus, beyond a certain point, precipitation will decrease with altitude. In A primary reason for the contrast between the the HKH, an early study by Hill (1881) suggested Karakoram and Himalaya is probably that the that rainfall in the northwestern Himalaya increases Karakoram is affected to a much greater extent by with elevation up to about 1,200 m and decreases the winter and spring westerlies, which are high- thereafter. Dhar and Rakhecha (1981) found that level meteorological systems compared to the maximum rainfall occurred in the foothills of the low- to middle-level monsoons. The Karakoram Nepal Himalaya at an elevation of 2,000–2,400 m. appears to share the increasing precipitation with altitude regime with the Greater Himalaya In the Indus and Jhelum basins, the highest totals (Singh, Ramashastri, and Kumar 1995), where the are at the comparatively low-level southern foothill monsoon contributes only 35 percent of the annual stations, which receive significant totals in both precipitation total. summer and winter. The main influence is the sheltering effect of mountain barriers. Thus, there is 2.4.2 Temperature a sharp northward decrease in the monsoon rainfall originating from the southeast in the Hindu Kush Since few measurements of temperature are made from July to September, falling to as low as 5 percent at high altitudes, extrapolation must be made from of the annual total at Chitral. measurements at lower elevations using lapse rates of temperature. The accuracy of such extrapolation Limited evidence from the Karakoram suggests a is critical for determining the extent, magnitude, and quite different relationship between precipitation and duration of the melt of snow and glaciers over a altitude than in the southern ranges of the Himalaya. range of elevations. Environmental lapse rates refer First, the valley floors at 1,000 to 1,500 m are arid, to the actual change of temperature with altitude with annual totals generally less than 200 mm. for a stationary atmosphere (that is, the temperature Studies by Jacobsen (1997) in the Yasin valley, a gradient); they depend on the saturation of the air tributary of the Gilgit River, showed increasing mass and can vary from day to day and season to precipitation up to 4,400 m (636 mm). Cramer season. Where measurements are made at ground- (1997) similarly found rainfall up to 720 mm in the based stations, additional variability arises from Bagrot valley at 4,120 m. Evidence at higher ground cover, including snow and ice, and the elevations comes from the analysis of firn ice. potential for temperature inversions at valley stations. Shi and Wang (1980) concluded that precipitation above 5,000 m on the Batura Glacier could exceed Linear regression of seasonal temperatures with 2,000 mm, while Wake (1989) suggested typical station altitude in the upper Indus (Archer 2004) annual accumulation rates of 1,500 to 2,000 mm at gave marginally higher lapse rates during the 5,500 m in the central Karakoram. In addition, spring and summer (0.75°C per 100 m) and hydrological studies of the runoff rates from lower during fall (0.65°C per 100 m). Similar Karakoram glaciers show annual rates of over monthly analysis of lapse rates for maximum and 1,000 mm, which again implies (if the glaciers are minimum temperatures enabled freezing levels to 12 Climate Monitoring Figure 2.3 Estimates of Monthly Freezing Level and Seasonal Mean Daytime Land Surface Temperature Lapse Rates for Upper Indus Basin a. Elevation of freezing level for monthly maximum b. Daytime land surface temperature versus altitude and minimum temperatures based on station data, by season, 2000–09, based on MODIS remote 1994–2000 sensing data Variation in monthly maximum and minimum Seasonal mean of Daytime Land Surface Temperature (MODIS Terra) vs Elevation, NW UIB, Period means (2000-2009) by season 7000 7000 6000 6000 Elevation (m above sea level) Maximum 5000 5000 Elevation (m) 4000 Minimum 4000 3000 3000 2000 2000 1000 0 1000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec -35 -25 -15 -5 5 15 25 35 45 Land Surface Temperature (°C) Tday_DJF Tday_MAM Tday_JJA Tday_SON 97% tile//500mb 600mb 700mb 5% tile Source: Archer 2004; Forsythe et al. 2010. be calculated for each month, as shown in Figure in Figure 2.3.b (Forsythe et al. 2010). Note that the 2.3.a. From a hydrological point of view, the upper MODIS land surface temperature data are based zone at elevations above the 0°C isotherm for on “clear sky conditions,” and do not incorporate maximum temperature is one of continuous frost, data from overcast periods; and they are based on where precipitation falls as snow and where there is separate passes at approximately 11:00 (daytime, virtually no contribution to river runoff. However, it as shown in Figure 2.3.b) and 23:00 (nighttime, provides nourishment to lower zones through snow not shown). Examination of the daytime land avalanche and glacier flow. The middle zone is one surface temperature lapse rate seems to indicate with frequent freeze–thaw cycles, where precipitation differentiated behavior depending on the snow may fall as rain or snow, melt of lying snow occurs cover or melt state. For example, the summer (June– during daylight hours, and refreezing occurs at night. August) and fall (September–November) seasons – In the lower zone with continuous above-freezing corresponding to melting and snow-free conditions temperatures, precipitation is expected to fall as rain – demonstrate largely identical vertical patterns, and melt is continuous, though enhanced during a key feature of which appears to be a relatively daylight hours. constant value (approximately 7°C per 1,000 m) below 4,500 m above sea level. In contrast, the Satellite-based MODIS land surface temperature winter (December–February) and spring (March– data provide an opportunity to evaluate continuous May) seasons – corresponding to freezing and snow- changes in temperature and lapse rate through the covered conditions – have a steadily reducing lapse full elevation range of a catchment area, as shown rate with increasing altitude. 13 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains A quite different picture of lapse rate variation melt is primarily due to radiant heat under cloud- emerges from studies in the Sutlej basin (Figure 2.4) free conditions. It has been investigated rigorously (Wulf, Bookhagen, and Scherler, in review). Both simply because it is the only variable with long and MODIS and ground-based data from the Sutlej continuous records. In practice, temperature has valley suggest that the lapse rate is at a minimum been found in many applications to perform as during the summer. Wulf, Bookhagen, and Scherler well in modeling streamflow as using the full set of suggest that this is because the upper Sutlej basin variables for the energy budget (Bergström et al. on the plateau of southwestern Tibet heats up much 1992; WMO 1986). However, application of the full more during the summer, thus lowering the lapse energy budget may prove necessary to solve some rate. The Sutlej also has a much stronger monsoon of the difficult problems of the HKH. Some of these influence than the upper Indus. questions are as follows: How is energy partitioned These analyses suggest that patterns of lapse rate between melt and evaporation or sublimation? vary not only with altitude and with season but also in What is the source of the increasing trend in quite different ways in different reaches of the HKH. diurnal temperature range experienced across Verification of lapse rates with ground-based data will the Himalaya (and shared with other parts of the be an essential component of successful modeling of Indian subcontinent) but contrasting with decreasing the melt runoff process, both for the upper Indus and diurnal temperature range elsewhere in the world? for catchments elsewhere in the Himalaya. In addition, radiation variables may be more readily adjusted than temperature for aspect, slope, and 2.4.3 Energy Balance Variables exposure where distributed modeling is based on a digital elevation model. Air temperature is an imperfect indicator of the heat budget at the snow and ice surface, especially where Some climate stations, for example the high-level automatic weather station in northern Pakistan, now Figure 2.4 include a full range of solar radiation, wind, and Annual Variation of the Temperature Lapse Rate humidity measurements, which should be thoroughly for the Sutlej River Valley investigated for spatial variations when they are made available. 10 MODIS lapse rate 2.5 Monitoring and Analysis Needs Temperature Lapse Rate (oC/km) 9 30-day moving average 8 The summary of previous analysis across the HKH, and particularly in the upper Indus basin, indicates 7 that some of the stated objectives can be plausibly satisfied with existing data in some parts of the 6 region. In all cases, the release of more up-to- weather station date data and data collected but withheld would 5 lapse rate provide a more secure basis for such analysis. This is particularly true in the case of meeting the needs 4 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec of water resources management using monthly and Month seasonal data in the upper Indus basin, where the Source: Modified after Wulf, Bookhagen, and Scherler, in review, WRR. main source of moisture for summer melt originates Note: The weather station location is a point measurement, while the MODIS data include all measurements within the Sutlej River basin. from winter and spring precipitation and where 14 Climate Monitoring climate variables are well correlated spatially. temperature lapse rates. Remote sensing provides Analysis could be carried out with data currently information on snow-covered area but little reliability available to establish whether such relationships in assessing snow water equivalent. The relationship apply in the eastern and central Himalaya, where between high- and low-level precipitation is likely to the principal source of precipitation, both for vary from year to year and from region to region. direct runoff and for accumulation as snow, is the There seems to be little alternative but to attempt summer monsoon. Flow and flood forecasting need to make the measurements, mainly of precipitation estimated catchment data at a short time interval, falling as snow, at the elevations where glaciers are preferably a day or less. While previous analyses being fed. indicate that temperature and precipitation are well correlated spatially and altitudinally on a seasonal Snow measurement offers the most serious basis, it is not clear whether these relationships challenges of all climatic variables in obtaining break down at shorter time intervals. This could be reliable estimates, either as falling snow or as snow investigated with current data, but results would on the ground, even in accessible environments be more reliable if high-elevation data, currently with good power and communication services. withheld, could be made available. More serious difficulties arise in attempting to make automatic measurements in remote environments The more difficult problems arise in satisfying the with limited power sources. The following is an objectives associated with climatic controls (both assessment of whether and how these difficulties can energy and moisture) on glacier mass balance be overcome. and trends. Glacier mass balance depends both on precipitation inputs and on energy outputs; 2.5.1 Using Existing Climatological Data trends could result from secular changes in either moisture or energy. Given the above analysis, it is A principal objective in climate monitoring is possible that the combination of temperature data to assess trends and periodicities in climatic measured at lower elevations and remotely sensed variables – or more generally, nonstationary or data could be used on a monthly or seasonal basis nonhomogeneous aspects of a climatic record to make first estimates of conditions at the elevations – either existing or eventually acquired. Such (3,500–5,500 m) where the greatest contribution assessment requires a long record period and will be of melt to runoff occurs. Direct measurements both long delayed if it is to depend on future monitoring. of temperature and other energy balance variables Therefore, the best use must be made of existing at these elevations will be needed to validate or records and an attempt should be made to ascribe improve on these estimates. causes of trends or periodicities observed. It is of particular interest to determine whether observed The greatest problem arises in the assessment changes are the result of: (a) widespread climatic of precipitation in the headwaters of glaciated phenomena such as the growth in greenhouse gases catchments. The analysis for the Indus River suggests or the influence of oceanic–atmospheric processes that precipitation measured at valley level can be such as the El Niño Southern Oscillation and the used as an index of runoff from catchments fed by Pacific Decadal Oscillation; or (b) the result of melt of seasonal snowfall. However, this is merely an observational changes or local land use influences. index; it does not indicate the actual precipitation The second group of factors is unlikely to have any received at different elevations. There is no physical significant effect on trends in glacier mass balance. basis for an extrapolation such as that provided by Awareness of these sources of nonhomogeneity or 15 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains nonstationarity is important, not only for the analysis The India Meteorological Department and Pakistan of existing data but also for the design, location, and Meteorological Department took over responsibility management of an improved network. for the networks after independence, though with a break of several years at some mountain stations. The climatological network is long established As far as can be ascertained, they continued the on the Indian subcontinent, with some records instrumentation and practices previously adopted of precipitation and temperature extending back with well-trained observers and periodic inspections. to the mid-19th century. For example, there is a It is assumed that inspection reports and changes in precipitation record for Leh in Ladakh from 1868 location and instrumentation practice were held in a and temperature from 1882. Prior to 1890, there central office. were large differences in the methods of measuring rain (time and type of rain gauge). The colonial Separate networks were established by hydropower government of India established a uniform system for and irrigation agencies to serve their own specific rainfall measurement in 1890 covering areas now in purposes. It cannot be assumed that these records Bangladesh, India, and Pakistan. were maintained to a similar high standard. Rainfall listings in annual yearbooks are for the Other daily climate variables, including temperature Indian subcontinent and cover principal stations for the pre-independence period for the whole in what are now Bangladesh, India, and Pakistan. Indian subcontinent, were published by the India Among the information provided for each year is a Meteorological Department in books, each covering comparison with the average monthly rainfall from a six-month period with one page for each day of the previous record. The average monthly rainfall record. It is assumed copies of these reports are held and the number of rainy days in each month were by the India Meteorological Department in Pune. An calculated for all stations with at least five years incomplete set is held at the Pakistan Meteorological of rainfall data but, in the majority of cases, the Department in Lahore. Given that a number of averages in 1891 extended back over 20–30 years papers have been published in which temperature and were listed in columns entitled “normal data.” data back to 1901 are analyzed (for example, Presumably all these data are still held at the India Kumar, Kumar, and Pant 1994), it can be assumed Meteorological Department as paper files, probably that monthly maximum and minimum temperatures now digitized back to 1900 or to the beginning of must have been held at many locations for over 100 records around 1860. years. The United Kingdom Meteorological Office holds 2.5.2 Assessing Stationarity and Homogeneity copies of India Meteorological Department rainfall of Records reports dating from 1891 to 1947. Lists of rainfall stations by province for areas covering locations A homogeneous climate time series is defined as now in Pakistan and neighboring areas of India were one where variations are caused only by variations in extracted for three snapshot years – 1891, 1920, weather and climate (Conrad and Pollak 1950). As and 1946. For Jammu and Kashmir, nine stations Peterson et al. (1998, p. 1493) note: are listed for 1891, 40 in 1920, and 46 in 1946. The earliest records included Gilgit, Leh, Srinagar, [M]ost decade- to century-scale time series and Skardu. Punjab, initially a much larger province, of atmospheric data have been adversely had 196 stations in 1891 and for a province impacted by inhomogeneities caused by, for reduced in size in 1946 had 206 stations, several of example, changes in instrumentation, station which were on the mountain fringe. moves, changes in the local environment such 16 Climate Monitoring as urbanization, or the introduction of different (and possibly other stations) now use thermistors observing practices like a new formula for instead of liquid-in-glass thermometers for maximum calculating mean daily temperature or different and minimum temperatures. Quayle et al. (1991) observation times. If these inhomogeneities used station metadata in the United States to are not accounted for properly, the results show significant changes, especially in maximum of climate analyses using these data can be temperature, since thermistors can respond more erroneous. quickly to rapid fluctuations in temperature. Data from the HKH are equally subject to the Inspection of the precipitation records for Gilgit possibilities of these nonclimate-related changes, as and Skardu shows that for the period 1928–47, the examples below demonstrate. daily rainfall totals in the archive were entered to the nearest 0.1 inch (2.5 mm), then to 0.01 inches Observational and Entry Errors during 1893–1927 and after 1947. In a study of Gilgit rainfall, Cramer (1997) found that, although Even the best records contain some observational there were 95 rain days in the year, 39 days had less or entry errors, perhaps arising from illegible than 0.2 mm and a further 21 days had less than handwritten records. For example, a misplaced 1.0 mm of rainfall, and together these contributed decimal point can affect the extreme statistics of just over 10 percent of the annual total. Thus during an entire record. Problems are multiplied where 1928–47 rainfalls of less than half 2.5 mm would observers are poorly educated or trained, especially have been rounded down to zero. The recording at remote sites where the prime purpose of the method is, therefore, a source of heterogeneity in organization is not meteorology. Thus the author the precipitation record. It is believed that the total has observed on the Indian subcontinent more annual rainfall will be underestimated, perhaps by as than one hill station where the observer ignored much as 10 percent. the minus sign in low temperatures! Such errors are often easily identified (minimum temperature Changes in Station Location greater than maximum), but others are more difficult, for example, those arising from a Many principal climate stations are now located at faulty thermometer. airports. Clearly they cannot have been at the same site through their entire 100-year record period. Changes of Instrument or Measurement Metadata in the form of station records, where Practice available, should be checked for the impact of such site changes. Kumar, Kumar, and Pant (1994) note that the instrument housing for thermometers in India Changes in Station Environment changed in 1926, when thatched sheds were replaced by Stevenson screens. Presumably this Station location changes can cause sharp affected all early records. Since the radiation discontinuities, while change in the environment properties of the housing might have changed, it was around the station can cause gradual biases in the reported that the India Meteorological Department data. Factors may be very local or arise from more took overlapping observations to estimate the bias. widespread phenomena and include: The earlier records were then adjusted. Further details of these adjustments are noted in Kumar and • Growth of vegetation or nearby new buildings Hingane 1988, mainly for the effects of urban heat affecting exposure of the station; islands on temperature. Automatic weather stations • Effects of new irrigation in the vicinity of the 17 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains station (affecting temperature as well as humidity discontinuities in records is double-mass analysis. and evaporation); A double-mass curve analysis plots the cumulative • Effects of the urban heat island with growing sum of the candidate station against the cumulative urbanization; and sum of a nearby station. Most double-mass plots • The growing influence of the Asian “brown are roughly linear, so a sudden change to a new cloud.” slope indicates a discontinuity. A problem with the method is that it is impossible to determine whether While mountain stations are less vulnerable to the indicated discontinuity occurred at the candidate some of these sources of bias, they are not entirely or nearby station. To correct for this problem, immune. For example, Kathmandu and Srinagar Rhoades and Salinger (1993) used plots of parallel may well be affected by their own urban heat island, cumulative sums at several nearby stations at the while all stations on the southern foothills are same time. affected to some extent by the Asian brown cloud. The use of data from several neighboring 2.6 Identification of and Adjustment stations to develop a reference series is integral for Bias to many methods (Potter 1981; Alexandersson 1986). However, in some cases, where changes A whole branch of statistical science has grown up to instrumentation and other elements have all around the identification of inhomogeneities in time been made at the same time (such as the change series data. Many of the techniques now in use have in instrument shelters in 1926), neighboring been developed over the last 20 years in order to station data cannot provide insights into those ensure that estimates of the effects of greenhouse inhomogeneities. gases on global temperature are not misrepresented. The statistical procedures are described briefly Methods are also available for detecting below, based mainly on review papers: Peterson et inhomogeneities in single stations (Zurbenko et al. al. 1998 for monthly or annual resolution series, 1996; Rhoades and Salinger 1993; Potter 1981). and Wijngaard, Klein Tank, and Können 2003 for Alexandersson (1986) developed the standard daily series. normal homogeneity test for a single break in a time series. A technique based on multiple linear Generally, a combination of statistical methods regression was developed by Vincent (1998) to and methods relying on metadata information identify steps and trends in temperature series. The is considered to be most effective to track down technique systematically divides the tested series inhomogeneities. Where metadata are not into homogeneous segments. Adjustments are available (as has so far been the case on the Indian applied to bring each segment into agreement with subcontinent) statistical methods alone can be the most recent homogeneous part of the series applied but with lower confidence. Wijngaard, Klein Tank, and Können (2003) applied a sequence of four tests to individual sites. The four 2.6.1 Statistical Methods test methods selected to test the homogeneity in the time series are the standard normal homogeneity test To isolate the effects of station discontinuities from (SNHT) (Alexandersson 1986); the Buishand range regional climate change, many techniques use data test (Buishand 1982); the Pettitt test (Pettitt 1979); from nearby stations as an indicator of the regional and the Von Neumann ratio test (Von Neumann climate. A long-established method for detecting 1941). The first three tests are capable of locating 18 Climate Monitoring the year where a break is likely, and are referred spring precipitation. If these records were found to as location-specific tests. Wijngaard, Klein Tank, to be unbiased and also representative of higher and Können (2003) also found that although elevations, then retreating glaciers might be more discontinuities were difficult to detect in annual readily explained by changes in precipitation rather means of temperature, the diurnal temperature than temperature. range proved a more robust means of detecting changes due to station or instrument relocations, 2.7 Monitoring Network Components changing observing practices and measuring techniques. Archer (2001) also found that diurnal Climatological information is lacking principally for temperature range could be used to detect a step elevations greater than 2,500 m above sea level, change in the temperature record for Gilgit. Peterson but particularly at levels above the equilibrium line et al. (1998) provides a description of the detection of glaciers, where ground-based measurements are and homogenization techniques that have been used extremely limited. Satellite-based data can fill some in both country-specific and global datasets. of the gaps at higher elevations but some variables (notably the water equivalent of precipitation falling 2.6.2 Investigation of Regional Consistency as snow) remain largely outside the current scope of satellite remote sensing capabilities. Therefore Apart from these objective tests, an appreciation ground-based information is still essential both for its may be gained of the reliability of trend assessments own sake and as a means of verifying satellite-based by a preliminary investigation of the consistency of measurements. However, the extreme environment records across a region, researching the following places severe constraints on instrumentation, power questions: Are seasonal totals (rainfall) and averages demands, and communication of data, as well as (temperature) highly correlated? Are the observed on installation and management of such networks. seasonal trends consistent between neighboring Manually maintained climate stations may be stations and across regions? Are there stations that possible in some locations up to the altitudes of are poorly correlated and have inconsistent trends? the highest villages (about 3,000 m) but, above Such an analysis can be particularly useful for that altitude, automatic weather stations will almost temperature that is spatially conservative and can be invariably be required. The absence of continuous seasonally correlated over a wide region. routine manual measurements places further limitations on the reliability and completeness of Indian mountain temperature records show strong data. Since measurements are required in relation seasonal consistency in the direction of trend to glacier mass balance, the siting of stations on between 1950 and 1991, which, surprisingly, during glaciers of variable mobility creates further difficulties the spring and summer months is a downward trend in measurement. but is predominantly upward in winter months, more in line with expectation (Shrivastava 2011). 2.7.1 Automatic Weather Stations The summer result is at least curious, given the widespread observation of retreating glaciers Many thousands of automatic weather stations, and the belief that it is caused by global warming manufactured and supplied by companies, are in initiated by an increase in greenhouse gases. use around the world. Companies generally provide A similar analysis applied to the same stations a standard complete weather station that includes a with respect to rainfall trend shows consistently mast from which to attach the sensors. For standard falling summer precipitation but consistently rising meteorological stations, sensors include temperature 19 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains and relative humidity, solar energy flux, possibly soil Figure 2.5 temperature, an anemometer and wind vane, and Kunjerab Automatic Weather Station at 4,733 m a tipping bucket or weighing rain gauge. Additional above Sea Level in Hunza Tributary of the Upper Indus in Pakistan sensors can be attached that measure barometric pressure, additional measures of solar radiation or sunshine duration, soil moisture content, and evaporation (measured by pressure transmitter in an evaporation pan). Standard stations come with associated data loggers and software. Logger memory can be uprated if many sensors are attached or if the logger memory is downloaded infrequently (remote sites). Power requirements are generally low for a standard station and may be supplied by standard AA battery Source: Photo by D. Archer. cells. However, there are options for operation via power mains (with adaptor), rechargeable lead-acid batteries with or without solar panels for recharging. This description is primarily for stations operating in Different communication techniques are used, a temperate environment. Where weather is extreme, particularly the Argos and Iridium satellite it may be preferable to custom-build an automatic systems, but also Inmarsat satellite system, GSM, weather station, selecting the most robust sensors radio communication and Bluetooth. All have for the site. Power may have to be uprated with solar advantages and disadvantages related to power panels or a wind turbine, for example, to heat a rain demands, costs involved, distance over which they gauge to prevent bridging of snow, to ventilate the can be communicated, amount of data that can temperature screen, or for communications. Some be transferred, one- or two-way communication new technologies for snowfall measurement may possibilities. For example, the Argos system is have such power demands that their installation very reliable but is expensive and limited in the may still be impractical without a power main. The amount of data to be transferred, and only data automatic weather station at Kunjerab in the Hunza retrieval is possible. Iridium, Inmarsat, and the GSM basin at 4,733 m above sea level is shown in system have two-way communication possibilities Figure 2.5. but are more power demanding, with the latter two having limited communication ranges. Radio 2.7.2 Communications communication is only used for short line of sight distances. In general new measuring systems are Options are generally available for local download mostly using the Iridium satellite system. using a portable personal computer or a hand-held data capture device, or for remote download by Meteor burst communication has been used landline telephone, GSM mobile phone link, satellite effectively for the automatic weather station link, or meteor burst technology. Reijmer (2011, p. operated in northern Pakistan by the Water and 13) summarizes typical communications used from Power Development Authority (WAPDA) and Pakistan remote mountain sites as follows: Meteorological Department. 20 Climate Monitoring 2.7.3 Measurement of Snow measured snowfall is generally deficient (Goodison and McKay 1978). New sensors offer the possibility Since precipitation contributing to glacier mass of more reliable measurement of snowfall, for balance or forming the seasonal winter snowpack example, the Campbell Scientific PWS 100 system, falls as snow, the accurate measurement of which uses an optical system and a laser-based snow is an essential component of glacier and light to accurately measure the velocity and size catchment water balances. However, measuring of precipitation. The use of two separate detectors snow has greater problems than measuring liquid and laser Doppler anemometry techniques allows precipitation. Snow is hard to catch, it is readily the device to distinguish falling snow from blowing redistributed on the ground, it melts differentially, snow by assessing only the downward component. and its presence restricts access to measurement Precipitation rate and accumulation values can be points. For hydrological and glaciological purposes, calculated automatically. Similarly, the OTT Parsivel the most important information required is the is a laser-based disdrometer for comprehensive volume of snow as water equivalent – the depth measurement of all types of precipitation using of water that would result from melting. Both the shadowing effects they cause when they pass measurement of snow as it falls and measurement through a laser band. A signal processor uses the on the ground are fraught with difficulties that are raw data to calculate the type of precipitation as further compounded where automatic measurement well as the amount and intensity of the precipitation. on remote sites is required. Both these instruments have a high energy demand and have so far been applied mainly where mains Falling Snow power is available. Further developments and testing may prove a valuable addition to measurements in For the measurement of snowfall, the standard extreme environments. practice in many countries and with respect to automatic weather station sensors is to use a Snow on the Ground conventional rain gauge where amounts and rates of fall are determined by a tipping bucket or a weighing The difficulties in accurately measuring falling snow mechanism. Such conventional gauges suffer and the subsequent differential redistribution and seriously from the effects of wind eddies created melting of the snowpack have focused attention by the gauge itself and by nearby obstructions, on the measurement of snow on the ground. Snow including interception of drifting or blowing snow, depth measurements do not in themselves provide bridging of the gauge orifice, or occasionally a useful hydrological measure, owing to the wide complete burial (Archer 1998). Wilson (1954) range of snow density, unless density can be demonstrated the effect of wind speed on gauge separately assessed. catch and showed, for example, that with a wind speed of 11 m per second the deficiency is around Where manual measurements are possible, 60 percent. For regular manual measurement in the standard practice is to measure snow water countries with regular high snowfall, special shields equivalent using a snow core sampler at a repeated are fitted to gauges to minimize the effects of series of points along a snow course. Averaging turbulent eddies, including the flexible Alter shield in of multiple points enables account to be taken the United States, the Tretyakov shield in the Russian of variability due to vegetation, drifting, and Federation, and the Nipher shield in Canada. Even obstructions. Few such measurements have been with these modifications, investigations show that made in the HKH, but De Scally (1994) used a 21 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains limited set of such measurements (now discontinued) 5,500 m elevation. Few such wide plateau reaches on the Kunhar River tributary of the Jhelum River exist at elevations above the ELA, but where they and obtained high correlation coefficients between do exist, they should be exploited to assess annual annual maximum snowpack water storage and precipitation accumulations and to compare with annual runoff. There seems to be potential for more totals derived by other methods. widespread use of such manual measurements at levels up to 3,000 m above sea level. Such 2.8 Recommendations measurements are likely to prove more useful in catchments fed by winter snowfall than in Himalayan Good climate data provide the basis for analysis catchments fed mainly by monsoon precipitation that relating to climate change and water resources falls as rain to a much higher elevation. management and are essential for policy making, design, and operation of water resources systems. The snow core sampler has the advantages of The following recommendations summarize and are mobility and low cost but has the disadvantage that derived from the previous discussion. The first set it cannot be adapted for remote and automatic of recommendations relates primarily to historical readings. For this purpose, a number of devices climate records whose acquisition and use has been have been developed. Ultrasonic distance- limited. However, such data need careful scrutiny to measuring sensors placed over the snowpack are ensure their reliability and homogeneity. widely used as part of an automatic weather station assembly for continuous recording of snow depth at 2.8.1 Data and Metadata Acquisition and a point, but it requires independent measurement of Validation Recommendations snow density to convert depth to water equivalent. • Acquisition and analysis of existing climate data Snow pressure pillows provide such information by should take priority over further development of detecting the pressure of snow on an antifreeze- the monitoring network; filled pillow. Differences in sequential measurements • Attempts should be made to extend the readily indicate rates of snow water equivalent accumulation available climate data (from about 1960–95) to and ablation. Again, such measurements are not include early pre-independence records from the mid-19th century. Hard copy records and perhaps without problems: the possibility of accumulation digitized records are available at government of material other than snow, development of ice meteorological departments, and some records lenses that form bridges over the pillow, and the are held at the United Kingdom Meteorological effects of differences in pillow albedo and roughness Office. Also, local sources, such as monasteries, from surrounding vegetation in shallow packs should be asked for climate records; may produce unrepresentative measurements. In • The data should also be updated to the present, addition, these are only single point samples of especially as a basis for validating satellite water equivalent and may not be representative of monitoring records, some of which may cover the surrounding area. only the last decade. Of particular importance are records from high-altitude automatic weather A unique approach was taken by Wake (1989), who stations, some of which now have records in used glaciochemical methods to distinguish annual excess of 15 years; accumulations in firn ice; he suggested typical • It is assumed that inspection reports, and hence annual accumulation rates of 1,500–2,000 mm at station histories, were made for every station. 22 Climate Monitoring These should be requested and made available indicate that some of the stated objectives can as a basis for checking the homogeneity of be plausibly satisfied with existing data in some records. Some sources of metadata contain parts of the region. In all cases, the release of large amounts of irrelevant information, more up-to-date data and data collected but making extracting what is useful or relevant withheld would provide a more secure basis time consuming and tedious. Digitization of for such analysis. This is particularly true in the metadata should be considered, as it could case of meeting the needs of water resources ultimately offer researchers access to station management using monthly and seasonal data history information without the expensive burden in the upper Indus basin, where the main source of performing a station-by-station search of moisture for summer melt originates from through paper archives; winter and spring precipitation and where climate • A preliminary appraisal of the data might include variables are well correlated spatially; cross-correlation of seasonal totals, averages, • Analysis could be carried out with data currently standard deviations, and assessment of trends available to establish whether such relationships in the raw data. This will provide a basis for apply in the eastern and central Himalaya, where identifying anomalous, though not necessarily the principal source of precipitation, both for incorrect, records; direct runoff and for accumulation as snow, is the • A program of homogeneity testing should then summer monsoon; follow, using agreed-upon and consistent tests. • Flow and flood forecasting need estimated Where bias correction has occurred in the past, catchment data at a short time interval, an attempt should be made to identify both raw preferably a day or less. While previous analysis and corrected data; indicates that temperature and precipitation • Trend analysis of the adjusted data should then are well correlated spatially and altitudinally on proceed; a seasonal basis, it is not clear whether these • If correction is necessary, the corrected data can relationships break down at shorter time intervals. be applied to catchment and glacier modeling; This could be investigated with current data, • Existing methods for the creation of gridded but more reliably if data from high elevations, climatic databases are severely limited currently withheld, could be made available; in the Himalayan region by the sparse, • For energy inputs to high-level snow and unrepresentative, ground-based climate network glaciers, it is possible that the combination of with which to validate models and interpolation temperature data measured at lower elevations methods. Such databases should be used with and remotely sensed data could be used on a care; and monthly or seasonal basis to make first estimates • Whilst numerous targeted climate research of conditions at the elevations 3,500 to 5,500 m, projects have been carried out in the HKH, where the greatest contribution of melt to runoff there is a need for continuing routine analysis, occurs. Direct measurements of temperature and especially relating to water resources, which other energy balance variables at these elevations requires coordination with hydrological databases will be needed to validate or improve on these often managed by different agencies. estimates; and • The greatest problem arises in the assessment of 2.8.2 Climate Analysis Recommendations precipitation contributing to glacier mass balance in the headwaters of glaciated catchments. • In summary, the previous analyses across the Monitoring and instrumentation for this purpose HKH, and particularly in the upper Indus basin, are noted below. 23 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 2.8.3 Monitoring and Instrumentation References Recommendations Alexandersson, H. 1986. “A Homogeneity Test • Owing to the extreme conditions, automatic Applied To Precipitation Data.” International monitoring stations will provide the main source Journal of Climatology 6 (6): 661–75. of continuous climate data at elevations above 3,000 m above sea level, and often at lower Archer, D.R. 1998. “Snow Measurement.” In levels also; Encyclopedia of Hydrology and Water Resources, • Where there are permanent villages at high eds. R.W. Herschy and R.W. Fairbridge, 618–21. altitudes, local personnel should be trained to Dordrecht: Kluwer Academic Publishers. run manual weather stations, check automatic Archer, D.R. 2001. The Climate and Hydrology of weather stations, and carry out snow surveys, Northern Pakistan with Respect to the Assessment among other tasks; of Flood Risk to Hydropower Schemes. • The operation and comparison of sensors Unpublished report. Lahore: GTZ/WAPDA/VSO. for a wide range of climatic variables should Archer, D.R. 2003. “Contrasting Hydrological be reviewed with respect to operation in Regimes in the Upper Indus Basin.” Journal of extreme conditions. In its 2011 workshop, the Hydrology 274 (1–4): 198–210. International Arctic Science Committee made recommendations on the use of automatic Archer, D.R. 2004. “Hydrological Implications of measuring systems on glaciers; Spatial and Altitudinal Variation in Temperature in • Standard bucket type precipitation gauges the Upper Indus Basin.” Nordic Hydrology 35 (3): seriously and variably underestimate snowfall; 213–27. their results should be used with caution in glacier Archer, D.R., and H.J. Fowler. 2004. “Spatial mass balance studies; and Temporal Variations in Precipitation in the • At suitable sites, snow survey courses should be Upper Indus Basin: Global Teleconnections and established as a means of validating automatic Hydrological Implications.” Hydrology and Earth measures of snow water equivalent; System Sciences 8 (1): 47–61. • New sensors for measuring falling snow (and other precipitation) using the shadowing effects Archer, D.R., and H.J. Fowler. 2005. “The Use of caused by precipitation when it passes through a Exploratory Data Analysis with Ground-based laser band offer the potential for more accurate Data to Assess Climate Runoff Links in the measurement both at manned and automatic Upper Indus Basin.” In Landschaftsökologie und sites. The operation and power demands should Umweltforschung, 48, ed. A. Herrmann, 75–82. be reviewed to determine whether they could be Brunswick. adopted for use in the HKH; and Archer, D.R., and H.J. Fowler. 2008. “Using • Snow profiling at suitable sites above the ELA Meteorological Data to Forecast Seasonal should be carried out to determine historical Runoff on the River Jhelum, Pakistan.” Journal of annual (and seasonal) snowfall. Hydrology 361 (1–2): 10–23. Basista, A., G.D. Bell, and V. Meentemeyer. 1994. “Statistical Relationships between Topography and Precipitation Patterns.” International Journal of Climatology 7 (9): 1305–15. Bergström, S., C.E. Bøggild, K. Einarsson, Y. Gjessing, N.R. Saelthun, T. Thomsen, B. 24 Climate Monitoring Vehviläinen, and K. Sand. 1992. Snow Eurasian Snow Cover within the Meteo-France Modelling, Water Resources, Climatic Change. GCM.” Climate Dynamics 12 (7): 449–66. SINTEF Report, Norwegian Hydrotechnical Forsythe, N., C.G. Kilsby, H.J. Fowler, and D.R. Laboratory. Archer. 2010. “Assessing Climate Pressures on Blanford, H.F. 1884. “On the Connexion of Glacier-Melt and Snowmelt-Derived Runoff in the Himalayan Snowfall and Seasons of Drought Hindu Kush-Karakoram Sector of the Upper Indus in India.” Proceedings of the Royal Society of Basin.” In Proceedings of the British Hydrological London 37: 3–22. Society Third International Symposium: Role of Hydrology in Managing Consequences of Bleasdale, A. 1965. “Rain Gauge Networks a Changing Global Environment, Newcastle, Development and Design with Special Reference United Kingdom, July 19–23, 2010. to the United Kingdom.” In Proceedings of the World Meteorological Organization/International Fowler, H.J., and D.R. Archer. 2006. “Conflicting Association of Hydrological Sciences Symposium Signals of Climatic Change in the Upper Indus on the Design of Hydrological Networks, IASH Basin.” Journal of Climate 19 (17): 4276–93. Publication 67: 46–54. Goodison, B.E., and D.J. McKay. 1978. “Canadian Buishand, T.A. 1982. “Some Methods for Testing Snowfall Measurements: Some Implications for the Homogeneity of Rainfall Records.” Journal of the Collection and Analysis of Data from Remote Hydrology 58 (1–2): 11–27. Stations.” Proceedings of the Western Snow Conference 46: 48–57. Conrad, V., and C. Pollak. 1950. Methods in Climatology. Cambridge, MA: Harvard University Hill, S.A. 1881. “The Meteorology of North-West Press. Himalaya.” Indian Meteorological Memorandum I (IV): 377–429. Cramer, T. 1997. “Climatic Gradients in the Karakoram and Their Effects on Natural Jacobsen, J.-P. 1997. “Investigations into the Vertical Vegetation.” In Perspectives on History and Temperature and Precipitation Gradients in Two Change in the Karakorum, Hindukush and Test Areas in Northern Pakistan.” In Perspectives Himalaya, ed. I. Stellrecht and M. Winiger, 265– on History and Change in the Karakorum, 77. Pakistan–German Research Project, Culture Hindukush and Himalaya, ed. I. Stellrecht and Area Karakorum Scientific Studies 3. Köln: M. Winiger, 146–61. Pakistan–German Research Rudiger Köppe Verlag. Project, Culture Area Karakorum Scientific Studies 4 (1). Köln: Rudiger Köppe Verlag. De Scally, F.A. 1994. “Relative Importance of Snow Accumulation and Monsoon Rainfall for Kumar, K.R., and L.S. Hingane. 1988. “Long Term Estimating the Annual Runoff, Jhelum Basin, Variations of Surface Air Temperature at Major Pakistan.” Hydrological Sciences Journal 39 (3): Industrial Cities of India.” Climatic Change 13 199–216. (3): 287–307. Dhar, O.N., and P.R. Rakhecha. 1981. “The Effect Kumar, K.R., K.K. Kumar, and G.B. Pant. 1994. of Elevation on Monsoon Rainfall Distribution in “Diurnal Asymmetry of Surface Temperature the Central Himalayas.” In Proceedings of the Trends over India.” Geophysical Research Letters International Symposium on Monsoon Dynamics, 21 (8): 677–80. 253–60. Cambridge, United Kingdom: Peterson, T.C., D.R. Easterling, T.R. Karl, P. Cambridge University Press. Groisman, N. Nicholls, N. Plummer, S. Torok, I. Douville, H., and J.F. Royer. 1996. “Sensitivity of Auer, R. Boehm, D. Gullett, L. Vincert, R. Heino, the Asian Summer Monsoon to an Anomalous H. Tuomenvirta, O. Mestre, T.S. Szentimrey, 25 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains J. Salinger, E.J. Førland, I. Hanssen-Bauer, H. Von Neumann, J. 1941. “Distribution of the Ratio Alexandersson, P . Jones, and D. Parker. 1998. of the Mean Square Successive Difference to the “Homogeneity Adjustments of In Situ Atmospheric Variance.” Annals of Mathematical Statistics 13 Climate Data: A Review.” International Journal of (4): 367–95. Climatology 18 (13): 1493–517. Wake, C.P. 1987. “Snow Accumulation Studies Pettitt, A.N. 1979. “A Non-Parametric Approach to in the Central Karakoram.” In Proceedings of the Change-Point Detection.” Applied Statistics the Eastern Snow Conference (North America), 28 (2): 126–35. 44th Annual Meeting, June 3–4, 1987, 19–33. Fredericton, Canada. Potter, K.W. 1981. “Illustration of a New Test for Detecting a Shift in Mean in Precipitation Series.” Wake, C.P. 1989. “Glaciochemical Investigations Monthly Weather Review 109 (9): 2040–45. as a Tool to Determine the Spatial Variation of Snow Accumulation in the Central Karakoram, Quayle, R.G., D.R. Easterling, T.R. Karl, and P.Y. Northern Pakistan.” Annals of Glaciology 13: Hughes. 1991. “Effects of Recent Thermometer 279–84. Proceedings of the Symposium on Snow Changes in the Cooperative Station Network.” and Glacier Research Relating to Human Living Bulletin of the American Meteorological Society Conditions held at Lom, Norway, September 4–9, 72 (11): 1718–24. 1988, ed. B. Wold. Reijmer, C.H. 2011. “Workshop Summary and Wijngaard, J.B., A.M.G. Klein Tank, and G.P . Recommendations.” In Workshop on the Use of Können. 2003. “Homogeneity of 20th Century Automatic Measuring Systems on Glaciers, ed. European Daily Temperature and Precipitation C.H. Reijmer. IASC Workshop, March 23–26, Series.” International Journal of Climatology 23: 2011, Pontresina, Switzerland. 8679–92. doi: 10.1002/joc.906. Rhoades, D.A., and M.J. Salinger. 1993. “Adjustment Wilson, W.T. 1954. “Discussion of Precipitation at of Temperature and Rainfall Records for Site Barrow Alaska Greater than Recorded (by R.F. Changes.” International Journal of Climatology Black).” Transactions of American Geophysical 13: 8899–913. Union 35: 203–7. Shi, Yafeng, and Wenying Wang. 1980. “Research WMO (World Meteorological Organization). on Snow Cover in China and the Avalanche 1986. Intercomparison of Conceptual Models of Phenomena of Batura Glacier in Pakistan.” Snowmelt Runoff. Operational Hydrology Report Journal of Glaciology 26 (94): 25–30. 23. Geneva: WMO. Shrivastava, S.K. 2011. Evaluation of Temporal WMO (World Meteorological Organization). 2010. and Spatial Climatic Variability over Indian Guide to Climatological Practices, 3rd Edition. Himalaya. PhD thesis. Roorkee: Indian Institute of WMO 100. Geneva: WMO. Technology. Wood, M. 2007. The Story of India. London: British Singh, P., K.S. Ramashastri, and N. Kumar. 1995. Broadcasting Corporation Books, Random “Topographic Influences on Precipitation House. Distribution in Different Ranges of the Western Zurbenko, I., P.S. Porter, S.T. Rao, J.Y. Ku, Himalayas.” Nordic Hydrology 26 (4–5): 259– R. Gui, and R.E. Eskridge. 1996. “Detecting 84. Discontinuities in Time Series of Upper Air Data: Vincent, L. 1998. “A Technique for the Identification Development and Demonstration of an Adaptive of Inhomogeneities in Canadian Temperature Filter Technique.” Journal of Climate 9 (12): Series.” Journal of Climate 11 (5): 1094–1104. 3548–60. 26 Glacier Mass Balance Monitoring 3. Glacier Mass Balance Monitoring Monitoring glacier hydrology presents a unique set • All-year conditions and cycles in glacier basins, of challenges compared to other phases of water notably the largely neglected winter and and energy balance, especially under the conditions “shoulder” seasons; and in the HKH region. Glacier monitoring usually has • Glacier movement and thermal regimes, especially three main objectives: to determine mass balance; in relation to the predominance of sliding motion to track fluctuations in ice extent; and to identify and evidence of chronic flow instabilities that are hazards. This chapter describes what is needed to most sensitive to conditions controlling thermal, monitor mass balance in relation to water supplies meltwater, and drainage variability. from snow and ice and their sustainability in the face of climate change and increasing demand. These realities raise important challenges for monitoring. Some data on these factors exist The challenges of the HKH environments and for a few glaciers, but there are no systematic diversity of societal conditions are well known. Not observations across the region. Obviously, neglect only are the mountains the highest, steepest, most is partly due to the difficulties of observing many of rugged, and extensive on Earth, but there is also these phenomena. Globally, monitoring for mass enormous variability of conditions and types of ice balance and glacier hydrology has deliberately mass and differences in glacier responses to global avoided glaciers with these characteristics. Their climate change across the region. In addition, importance is rarely recognized in approaches planners must keep in mind the diversity of societies imported from other regions. This chapter describes and cultures in the mountains, development priorities how these aspects of glaciers intervene to generate and security issues in the countries concerned, and distinctly different spatial and temporal patterns institutional arrangements. in the region, creating severe problems for widely used notions such as the snowline, ELA, and This chapter describes the following phenomena that accumulation–area ratio. play distinctive roles in glacier mass balance in the HKH: The chapter concludes with a provisional list of • High-elevation snowfall and its quantities, suggestions concerning direct monitoring, including gradients, and variability in glacier source zones, setups for training and experience of personnel, and as well as the differing climatic and seasonal research to address important issues. The neglected regimes; conditions seem mainly to require observing systems • Redistribution of snow by avalanche and wind; that work with and help test and support indirect • Avalanche-fed glaciers, predominant in the HKH and remotely sensed approaches. Actual monitoring but virtually unexamined and unresearched; networks for the glaciers themselves may not be a • Debris covers in ablation zones, including the feasible goal, but much can be done to improve and extent and relative roles of different thicknesses, integrate direct glacier observations into broader especially in the rather neglected areas of thin systems, and any successful program will need and scattered, ablation-enhancing materials, field competence and observations. This raises which take up more of ablation zones, are larger special questions of equipment and instrumentation, sources of water yields, and are climatically more personnel, training, and safety, as noted by Young sensitive than heavy mantles; and Hewitt (1993): 27 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains The greatest value scientifically would be derived The same is likely to apply in the strategies that from a concentration on continued measurement may be recommended for the HKH and countries of ablation in the ablation areas of the glaciers. It concerned, although the focus here is on issues is here that very high specific yields are combined for direct monitoring. It seems fair to say that best with relative ease of access. Within these zones practices – the core concepts and standards for the strategy should be to further refine balance glacier monitoring – have been developed mainly gradients and how those gradients change under through direct approaches or tested in relation to different meteorological conditions. This should be them. These have their problems everywhere, notably effected on areas of bare, relatively smooth ice at all relating to logistics, cost, instrumentation, and safety. elevations up to the equilibrium line. In the high mountains, constraints on access to many parts of the glaciers, instrumentation, and safety are 3.1 Monitoring Approaches especially severe. Approaches to glacier monitoring may be broadly Mass balance is the main and preferred approach defined as follows: to glacier hydrology. It looks at the relations of inputs or nourishment and losses for the whole • Direct measurement of ice mass variables and glacier. With mountain glaciers, the usual concerns environmental controls in glacier basins, which are with snowfall accumulation on the upper basin requires field programs and instrumentation on and surface ablation in the lower. The HKH raises and beside the ice; unusual problems in this respect. In determining • Indirect measurement, usually based on climatic fluctuations, ideally the whole glacier should be conditions and water discharge measured outside considered, including its thickness and area–altitude glacier basins themselves. Standard weather distribution. In practice, this has rarely been done stations and river gauging stations are the main in the HKH. Nearly all studies relate to terminus basis for longer-term analyses, as well as possibly fluctuations. Reliance on these to track glacier snow surveys. The quality of the results depends changes is unusually problematic in the region. upon assumptions, correlations, and calibration with respect to what is happening at the glaciers; In the HKH, physical constraints along with economic • Remote monitoring using terrestrial, airborne, and pressures and institutional priorities have led to a satellite-based sensors that can determine glacier preference for alternatives to direct measurement. and off-ice parameters. Ground-truthing is a Most estimates for the region to date use indirect and critical concern; and remotely sensed data and model assumptions. With • Models of glacier systems that provide data on a few exceptions, described below, inputs are usually relations and outputs derived from numerical inferred, not measured, and attributed strictly to or analog and related variables. Most have “precipitation.” Also, with a few important exceptions, been developed for well-established cases and direct observations of outputs are confined to termini: tested for them in terms of the first three types sometimes lower ablation zone areas, more often of measurements. A basic issue, again, is how from streamflow data measured below the glaciers well model results address critical conditions and themselves. Nearly all direct observations available correlate with direct measurements in the HKH. are for glaciers close to transmountain highways, on trails to highly valued mountaineering areas such Each of these approaches has its advantages and as Mount Everest or K2, or on pilgrimage routes. problems. In most places today, where water supply, This raises problems of representativeness and even glacier change, or hazards are concerned, some concentrated human influences on glacier ablation, combination of the four approaches is in use. such as black carbon from fuel burning. 28 Glacier Mass Balance Monitoring There is an emerging awareness, however, that 3,000 km from west to east, or almost 35° of checking assumptions and results against what longitude. The glaciers of Nepal, Bhutan, and Sikkim happens on the ground is essential, given the recent are 6°–7° further south than those in Karakoram and reports, exaggerated or misread, of “disappearing Hindu Kush; 660–770 km further into the tropics, glaciers” and imminent water crises. While not as well as being much further east. A predominantly entirely unwarranted, much commentary has been southeast to northwest trend introduces considerable based on poor or absent evidence and on poorly climatic and ecological diversity but the strongest constrained assumptions (Raina 2009; Cogley gradients tend to be across the grain of the 2011; Scherler, Bookhagen, and Strecker 2011). mountains – north to south or, in the western part, Glaciers with characteristics that prevail in the HKH northeast to southwest. have been systematically avoided and excluded. Conversely, programs are lacking in the HKH not Some estimates put the number of glaciers in the only because of economic and cultural conditions, HKH region at around 50,000 (Williams and or even the extreme terrain, but also because the Ferrigno 2010). The Karakoram, feeding the glaciers and environments of interest present serious Indus and Yarkand Rivers, has the most extensive challenges for conventional approaches. Security cover, with over 16,000 km2 of perennial snow issues have also been a major impediment. An and ice. The second largest is in and around the approach is needed that, while recognizing existing Nyainqentanglha Range in southeastern Tibet, best practices, is informed by the conditions in the draining mainly to the Tsangpo-Brahmaputra River. HKH and, where necessary, open to other methods. Before looking specifically at this parameter, the Most of the ice cover consists of valley glaciers. geography of HKH glacierized areas and programs Some small ice caps occur on peaks rising from already investigating glaciers will be described. the Tibetan plateau, and countless hanging, slope, and cirque glaciers on the lesser ranges. The larger 3.2 High Asian Context valley glaciers comprise a major share of the ice volumes, however. In the Karakoram, the largest The broader geographic setting is variously referred 15 glaciers comprise about half the total cover; to as the Greater Himalaya region, the Tibetan the largest 50, almost three quarters. They include plateau and adjoining regions (Tandong 2007), or a majority of the largest glaciers in the region and more recently the “Third Pole” (Qiu 2008). As well the largest outside high latitudes (Hewitt 2011). as the main Himalaya, Karakoram, and Hindu Kush, The size of glaciers, their proportions in ice masses the region includes the Pamir, Kunlun, Tien Shan, of different sizes, and their distributions should and east Tibetan ranges and the Tibetan plateau. be considered when assessing their usefulness The Greater Himalaya region supports over a dozen for monitoring. To date, this consideration of major concentrations of glaciers at high elevations. distribution has not happened, although the various The total perennial snow and ice cover is thought to inventories available may help in determining exceed 100,000 km2. monitoring networks and procedures. The region that is the focus of this study, the HKH, The near total neglect of the smallest glaciers in is the arc of high mountains rimming the southern most of the region raises some questions about margins of the Tibetan plateau. They are the source how to proceed. First, although individually small, of the headwaters of the three river basins of in several of the ranges of interest they represent primary interest – the Indus, Ganges, and Tsangpo- considerable amounts of snow and ice in total, as Brahmaputra. The high mountains form a belt, rarely much as 6,000 km2 in the upper Indus basin and more than 200 km wide, that stretches nearly even more in the upper Tsangpo-Brahmaputra. 29 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains They add up to much more snow and ice than in countries and the rivers draining the catchments some better-known and more intensively studied cross several countries. The Indus glaciers and areas such as the European Alps or New Zealand meltwaters originate within the boundaries of Alps. Second, if seemingly minor in countrywide four countries and drain to the lowlands of three. overviews and regional water supplies, smaller Glaciers of the Ganges basin are found in three glaciers are more critical for the populations living countries and the tributaries drain through four. The in the mountains and for ecosystems. Small ice Tsangpo-Brahmaputra glaciers are found in four masses on lesser ranges and interfluves, along with countries and drain through five. seasonal snowmelt in their basins, tend to offer a more manageable local water supply, for example, The largest quantities and relative shares of glacier in small-scale hydropower developments. They meltwater involve the main stems of the Indus and can also be more sensitive to climate variations. Tsangpo-Brahmaputra. It is noteworthy that both Third, these are glaciers closer in size to those used mainly derive from trans-Himalayan glacier systems. worldwide as reference or benchmark glaciers to Most glaciers draining to the Indus do so through the track mass balance and climate-related glacier northwestern Himalayan or Nanga Parbat syntaxis, change. and most draining to the Tsangpo-Brahmaputra do so through the eastern Himalayan or Namche Barwa A major problem for monitoring is that the small ice syntaxis – some of the most geologically active, steep, masses are strongly influenced by local terrain and and high-relief terrain on Earth, which is also being by topoclimatic and other conditions. Ruggedness closely scrutinized for planned hydroelectric projects. and the glacier altitude relative to the regional snow line mean that many are largely or wholly fed For the Indus, the main glacier area and mass is by avalanches or affected by wind redistribution in the Karakoram Mountains, most of it controlled of snow. Such conditions have been scrupulously by Pakistan but disputed and claimed by India. avoided when selecting benchmark glaciers, raising Other significant glacier areas are under Indian problems for any plans to establish them in the HKH; control, mainly in the eastern Karakoram, Ladakh, whether small glaciers can ever be representative of and Zanskar ranges, many of which are territories their region is in doubt. The larger glaciers are more disputed by Pakistan. China controls a small area of important in regional hydrology and may provide a the Aksai Chin region, with some glaciers, notably better basis for assessing regional conditions. They the great Rimo system, also disputed by India. do present other large logistic and safety problems, Finally, most of the Hindu Kush and Hindu Raj as covered in the next section. glaciers drain to the Kabul River through Afghanistan but originate in the far northwest of Pakistan. 3.3 National and Transboundary Issues The river returns to Pakistan through the densely populated and strategically located Peshawar basin. Most of the waters from glaciers of the HKH originate within the boundaries of one country but The Tsangpo-Brahmaputra glaciers are largely in flow into and across one or more others. The larger territory controlled by China in the southern and populations and extent of dependence on waters eastern Tibetan ranges, notably the Nyainqentanglha originating from glaciers occur in the surrounding Mountains. Smaller but important glacier areas lowlands, not necessarily in the same country as the occur in Bhutan and Sikkim. India also has some glacier headwaters. This applies to the three major Himalayan tracts and minor glaciers within the rivers of interest: the Indus, Ganges, and Tsangpo- Tsangpo-Brahmaputra, but its situation is otherwise Brahmaputra. For each, the catchment areas similar to Bangladesh in receiving the glacier waters in which the glaciers are found straddle several from other countries. 30 Glacier Mass Balance Monitoring The glacierized areas of the Ganges basin are 3.4 Glacier Inventories and comparably large in total, but more scattered by Reference Materials mountain range, countries, and several major tributaries. The waters from them are dwarfed Several attempts have been made to assemble by runoff from heavy rainfall in the foothills and comprehensive inventories identifying the extent and plains. Nevertheless, on certain tributaries, they locations of glaciers in the region. It is important are critical for water and power developments to recognize that there are some limitations to the in the mountainous headwaters. They assume organization of regional inventories in terms of special importance in years of weak or failed glacier conditions. The inventories seem to include monsoons, at times of flood risk, and for perennial snow and ice areas rather than strictly groundwater recharge in the vast fans where the glacier cover. The areas inventoried combine rivers leave the mountains. permanent snow and glaciers above climatic snowlines or firn limits, as well as active glacier ice From a monitoring perspective, these observations below them. raise two distinct issues: transboundary sharing   of water and information. Also at issue is how A recent revision of the Karakoram inventory based far strictly glacial conditions are of interest on higher-resolution satellite imagery suggests when the water supply from them is attenuated that the figures consistently and considerably and modified in the key lowland areas. In the overestimate the actual glacier areas. Active glacier past, awareness of what is needed to effectively ice comprises less than 50 percent of perennial monitor the glaciers has been lacking and has snow and ice (Hewitt 2005, 2011). Mayer et al. received limited interest and funding compared to (2006) came to a similar conclusion for Baltoro other priorities. Glacier. The inventories have nothing to say about the huge avalanche- and wind-driven redistribution Each country faces different challenges, options, of snowfall in off-ice areas and to the glaciers, which and likely solutions with regard to glacier probably involve two thirds or more of all the snow monitoring. The objective givens of glacial that ends up as glacier ice. hydrology and geography require sharing and cooperation where major components of water No inventory appears to have taken into account supply involve transboundary flows. It is difficult the considerable amounts of dead ice in areas of to plan for effective monitoring for larger water glacier retreat since the Little Ice Age, as well as resources concerns and to track climate change in thousands of active rock glaciers. These two without information sharing between countries in accumulations could amount to some thousands of the region. However, such sharing also offers a km2 of near-surface ice in the HKH. Although they potential source of secrecy and conflict. are minor sources of water outside the mountains because of slow ablation rates, locally they are From a broad perspective too, glaciers and important for water resources within the mountains, aspects of their hydrology are of most immediate and in relation to climate change, ecosystem and greatest concern for people living in the stresses, and cryosphere hazards. HKH areas within each country. They may have very different interests and needs to those in In addition to glacier inventories, other inventories downstream countries, and may have broader have been made of glacial lakes, outburst floods, national and international concerns. and permafrost. These help provide a perspective 31 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains on the region and in considering priorities and Pakistan, as well as China which has headwaters for options for monitoring. However, the criteria used glacier-fed HKH rivers. may not be appropriate for mass balance concerns, water yields, and responses to climate change, or 3.5.1 India as guides to monitoring. The inventories provide necessary and useful statistics for national and Since 1978, the Geological Survey of India has cryosphere background, but may also create an carried out glacier mass balance and hydrological impression of comprehensive and coordinated studies for the Indian Himalayan states of Arunachal knowledge of what lies behind the distributions and Pradesh, Himachal Pradesh, Jammu and Kashmir, forms so classified, when this hardly exists. Sikkim, and Uttarakhand (Kaul 1999). An inventory of 5,243 glaciers was compiled with a total area of Glacier investigations in the region, especially the 37,959 km2 (Raina and Srivastava 2008). The data more intensive and scientific ones, have been done were based on Survey of India topographic maps, mainly by expeditions, which go back almost two aerial photographs, and satellite images, but there centuries. They can be the best or only sources of has been very limited field verification. The highest observations on the glaciers when reconstructing concentration of glaciers was found in Jammu and recent glacier change or compiling inventories of Kashmir, the lowest in Arunachal Pradesh. events for risk assessment (Mason 1954; Dainelli 1959; Bhambri and Bolch 2009). Institutions Bhambri and Bolch (2009) provide a critical assuming responsibility in this field will need access to assessment of glacier mapping in the Indian comprehensive bibliographic resources to track down Himalaya; they used available topographic sheets data and results from past research. Nevertheless, the to reconstruct glacier fluctuations. Raina (2009) sources suffer from patchiness in space and time, as provides a critical assessment of the state of well as a range of different national languages and knowledge of Indian glaciers. He summarizes the scientific traditions (Hewitt 2007a). Reconstructing observational base and identifies where there was a conditions from the older literature and datasets field measurement component to check estimates. certainly can be of value, but is complicated by how While his research supports the conclusion of dramatically glacier science changed and advanced generally declining glacier covers and volumes, it after the 1950s. A compilation of key references on challenges the popular view of “rapidly disappearing the glaciers, possibly inventories of the best available glaciers.” topographic maps of glaciers, and satellite coverage – detailed in the recent United States Geological Raina provides a table of “net mass balance” results Survey (USGS) Satellite Image Atlas of Glaciers in millions of cubic meters (m3) for seven glaciers and of the World – would provide a useful research for various years. However, he then points out that: source (Williams and Ferrigno 2010). Sources cited in this chapter focus on studies dealing with direct Hardly any information is available measurements and assessments of mass balance. regarding winter precipitation/accumulation. Even during summer months, though 3.5 Past and Present Monitoring Efforts meteorological stations had been established in the Region at each and every glacier under observation, no data about snow precipitation were Monitoring efforts comprising glacier mass balance available. Lack of these data has been a and hydrological studies have taken place in the major constraint in evaluating a specific factor HKH since at least the 1970s. Recent efforts are that leads to fluctuation in glacier regimen described here by country for India, Nepal, and (Raina 2009, p. 17). 32 Glacier Mass Balance Monitoring Apparently some surveys were undertaken to rectify 3.5.3 Nepal the lack of snow data, and satellite imagery is now being used to demarcate snow cover. Nevertheless, Between 1999 and 2004, the International Center he identified a critical problem, which is one reason for Integrated Mountain Development (ICIMOD) and why different strategies are needed for the HKH. its partner institutes prepared an inventory of glaciers in selected parts of the HKH region. The inventory 3.5.2 China was based on the 1:50,000 scale topographic maps published by the Survey of India in the 1970s, or China itself is outside the terms of reference for Landsat satellite images from 2000 for areas for this study, but is important because of active and which no topographic maps were available. The growing investigations of glaciers in High Asia as study identified more than 15,000 glaciers with a well as in all three of the basins of interest, some total area of some 33,000 km2 (Ives, Shrestha, and of which have headwaters in Chinese-controlled Mool 2010). territory. ICIMOD is conducting a Cryosphere Monitoring Chinese scientists and institutions have conducted Program with funding from Norway.3 This is a five- many field-based investigations in the past two or year program that includes a mass balance study of three decades. A detailed glacier inventory was two glaciers, Rika Samba and Yala. Both catchments published in 2000 in 22 separate documents, based will be equipped with permanent weather and on World Glacier Inventory guidelines. It identified hydrometric stations, with plans to measure each almost 46,300 glaciers with a total area of 59,406 glacier once every year. Short-term intensive km2 (Chao-hai et al. 2000). The Institute of Tibetan field missions are planned to collect relevant Plateau Research and the Cold and Arid Regions glaciohydrological data, including ablation on Environmental and Engineering Research Institute debris-mantled ice, with the aim to determine spatial (CAREERI) of the Chinese Academy of Sciences and vertical variations in precipitation. The program compiled a glacier inventory (Shih et al. 2010). intends to build the capacity of ICIMOD’s Nepalese In 2009, CAREERI published a map to a scale partners in glaciology and glaciohydrology. It of 1:2,000,000 of the glaciers and lakes of the includes a master’s level course that has already Tibetan plateau and adjoining regions (Tandong started in Kathmandu University, intended to enroll 2007). CAREERI is reported to be in the process of five students every year for five years. Most of this is compiling a second, updated and improved national yet to be carried out; it would be important to know inventory. what will be measured, the elevation ranges to be covered, instrumentation (how the considerable A major initiative is under way under the title “Third problems of winter weather and maintenance will be Pole,” focusing on climate change and its impacts managed), and integration with indirect and remote on snow, glaciers, and permafrost. Leadership, approaches. significant participation, and funding come from China. Glacier investigations throughout the A French team from the Institut de Recherche pour Greater Himalaya region are being actively pursued, le Développement, Grenoble, has been monitoring including transects and other approaches within Mera Glacier for last two years, conducting the three basins of interest here. This may initiate measurements to estimate mass balance. In 2011, or assist in national and transboundary monitoring they also started monitoring a debris-covered glacier systems and information sharing. – Changri Nup – in the Khumbu area of Mount 3 Arun Shestra, personal correspondence, November 2011. 33 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Everest. Details of the work are not known, nor is it to climate change, for which a proposal was known whether these glaciers will be representative prepared for the federal government and an of larger mass balance and water resources approach made to the World Bank for funding. estimation. In addition, the Pakistan Snow and Ice Hydrology 3.5.4 Pakistan Project includes a program to monitor the Pasu Glacier, Hunza, in collaboration with the Pakistan There has been a surge of interest in the upper Meteorological Department. An automatic river level Indus glaciers in recent years, mainly in relation station has been installed at the outlet of the glacier, to threats of “disappearing glaciers” and water and the use of tracer technology to measure discharge resources demands. Various centers, cells, institutes, from the glacier is being investigated. The Pakistan and government departments have begun using Meteorological Department has installed an automatic remote sensing, GIS, and satellite imagery. They weather station near the terminus of the glacier and are mainly concerned with mapping of snow and plans to install another at a higher elevation. ice and cataloging the glaciers. Several institutions have taken initiatives to develop field programs and 3.6 Current State of initiate or expand instrumentation near glaciers in Direct Glacier Monitoring the upper Indus basin. Most of this is at an early stage of development. A number of efforts have been made to begin or extend and upgrade glacier observations in the HKH The Hydrology Research Division of WAPDA is unusual region, which, although promising, are at a very in that it has glacier and snow and ice hydrology early stage of development. Critical questions need investigations going back to the early 1980s, and even to be addressed regarding whether what is already some snow survey activity in the 1970s. WAPDA is known offers a sound basis on which to determine current engaged in the following activities:4 the best way forward. • Keeping the facilities provided under the Pakistan Almost all data currently gathered and estimates Snow and Ice Hydrology Project within acceptable for mass balance depend mainly or wholly on limits and generating seasonal and 10-day flow indirect data or extrapolation from terminus changes forecasts for the Indus at Tarbela River, Jhelum and snowline estimates (Ren et al. 2006; Raina River at Mangla, and Kabul River at Nowshera; 2009). Some major efforts are under way to exploit • Analyzing the data collected from the upper Indus improved quality and frequency of satellite imagery basin high-altitude network to understand the to assess total ice mass changes (Bolch, Pieczonka, climatic behavior of the hydrological active zone and Ben 2011; Scherler, Bookhagen, and Strecker through use of the unit hydrographs; 2011; Gardelle, Berthier, and Arnaud 2012; Benn • In collaboration with ICIMOD, training an et al. 2012; Kayashta and Harrison 2008). engineer in modeling for the Hunza basin by using the topographic kinematic approximation Only a handful of individual research projects have and integration (TOPKAPI) model; and been conducted on one or two glaciers for which • Most importantly, establishing a center within mass balance estimates do not depend entirely on WAPDA to monitor upper Indus basin glaciers indirect, remote, and modeling resources (Fujita for water resources management in relation et al. 2006; Smiraglia et al. 2008). A very few 4 Daniyal Hashmi, personal correspondence, 2012. 34 Glacier Mass Balance Monitoring automatic weather stations are located near and In its two major databases for mass balance and along the same glaciers. In general, however, the for fluctuations, the World Glacier Monitoring sort of ground control that has been essential to Service has no HKH glaciers (WGMS 2009). In his check and calibrate remote sensing elsewhere is pioneering study of global changes, Oerlemans lacking in the HKH. Prior to satellite coverage, most (2001) cited one glacier from the region, Minapin of the HKH glacierized area lacked any modern in the Karakoram, but it is predominantly avalanche observation capabilities. nourished and has only intermittent observations which suggest it is a surge type glacier. Like the Various studies have pieced together dispersed nearby Pasu Glacier, the main attraction of the reports from visitors to the glaciers for a century or Minapin for use as a source of observations is more, mainly to track terminus positions (Mason its road accessibility. However, it is questionable 1930; Bhambri and Bolch 2009; Shroder and whether established global categories and Bishop 2010). Inventories of terminus fluctuations assumptions for benchmark glaciers are appropriate have been used to provide regional overviews for or useful in the HKH for reasons addressed in the recent decades (Raina 2009). Longer-term regional next section. assessments of glacier hydrology and glacier change are largely based on indirect and model approaches Useful observations of Karakoram glaciers go (Kayashta and Harrison 2008; Bhambri et al. 2011). back at least 150 years, much longer than for the As yet there are no provisions for regular assessments western United States or Canada, for instance. or updates, forecasts, and information sharing. Observations in the region for the International Glacier Commission started to be compiled more In a few places in the HKH and for barely a handful than a century ago (Mason 1930; Korzhenevsky of glaciers, the full range of critical variables has 1930; Raina 2009). Activities to determine mass been measured, as described in the next section. balance and glacier hydrology in the Karakoram, for Rarely are they at all sufficient to produce plausible example, were resumed by major expeditions within mass balance estimates or to track the sources of a decade after the Second World War (Untersteiner glacier fluctuations. Few of the best datasets on 1957; Wiche 1959; Schneider 1969; Batura Glacier the glaciers extend beyond a few weeks of summer Investigation Group 1979). However, the work never and only rarely for more than one or two successive progressed to fully fledged or continuous monitoring. years. The few instrument stations in glacier basins are in off-ice locations, and none has as much as a Second, as noted, more is now being done in decade of continuous measurements. and by organizations in the countries concerned than at any time in the past. India and Pakistan, More specifically, there are no reference or for example, have established hydrological and benchmark glaciers in the HKH (Fountain et meteorological networks that, in terms of the length al. 1997; Bolch 2011). Such glaciers are the of record, forecasting capabilities, and quality of bedrock of global mass balance and climate personnel, should be the envy of most countries. change monitoring. In all cases, they are derived Historically, however, very little of this involved or from programs of direct measurements of main seemed to need monitoring of the high-mountain mass balance variables over many years and, for areas or glaciers. The focus has been on rainfall– the global set, over several decades (UNESCO runoff studies and river and groundwater hydrology 1998; Haeberli 2011). Such glaciers are generally in the foothills and lowlands. The field is benefitting expected to be representative of the region from the increasing numbers of professionals with concerned, although the fact that most are quite an essential requirement for this work – interest and small ice masses makes this doubtful. experience in mountains and with snow and ice. 35 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 3.6.1 Elements of Mass Balance in the HKH These relations are also the basis for expecting that any advance of the glacier terminus will reflect an Mass balance is an accounting approach, ultimately increase in the net balance. Above the terminus, about volumes or water equivalent amounts, and there will be thickening with positive balance or ratios between processes that add mass to a glacier thinning with negative, and somewhat faster or and those that remove it. Measurements are made slower rates of movement, respectively. In such or calculated for an annual cycle or “budget year.” cases, the main complication is a certain lag time, Among others, Paterson (1994) gives a detailed according to the size and morphology of the glacier, account of terms, relevant phenomena, and as adjustments pass through the whole. If the lag equations. times can be determined, terminus changes ought to be useful in tracking changes in annual budgets. Although new techniques and directions continue to be developed, a fairly well-defined conventional The conventional procedure needs to be picture has emerged from mass balance studies clearly understood. It arose out of direct glacier worldwide, providing a basis for most actual measurements, primarily in the European Alps monitoring and for indirect approaches and and the mountains of western North America, that assumptions. The focus has been on the balance has become the basis for the general analysis of between snowfall and accumulation in the upper the advance or retreat of the glacier terminus. glacier area, and ablation and glacier ice melting Moreover, mass balance estimates for the HKH from the surface on the lower part (Haeberli 2011). have used or assumed the conventional terms: that Mass may be added or subtracted in other ways, but is, nearly all studies and estimates for the HKH existing studies of mountain glaciers largely concern use the concepts developed in the Alps or North these two. They lead to a spatial division of glaciers America (Harper and Humphrey 2003; Raina into accumulation zones, where there is an annual 2009). When models or concepts are imported net addition of snowfall, and an ablation zone, with from other regions, the limitations need to be net losses. In valley glaciers an ELA separates the clearly understood. two, where inputs are exactly balanced by losses. Typically, the ELA is found to occur at or very close to Figure 3.1 shows the main zonal, vertical, and the firn limit on the glacier where the seasonal snow mass balance regimes of valley glaciers, based is completely removed, exposing glacier ice. “Firn” on Karakoram examples. While interrelated, the is snow that survives on the glacier through the end different area–altitude conditions respond to of the budget year. The firn limit at the end of the climate and climate change in distinctive ways. ablation season is commonly seen as part of the They vary from glacier to glacier and regionally, climatic snowline – the highest retreat of seasonal and all are factors that monitoring projects must snow cover in the mountain area. Physically, ice address. It will be seen that the conventional movement from the accumulation to the ablation approach, focused on snowfall accumulation zone maintains balance. Glacier movement volumes and surface ablation zone losses, and assuming and throughput are expected, therefore, to increase consistent relations between them, greatly towards the ELA and decrease below it down to the oversimplifies the high Himalayan picture. terminus. This is observed in most glaciers with well- established mass balance monitoring. It can be seen This sketch (Figure 3.1) of a Karakoram glacier how, if firn limits snowlines, and ELAs are closely illustrates many of the features of glaciers in the related, they offer a very convenient way to estimate circum-Tibetan Mountains of Asia. These glaciers mass balance. Deriving it from direct measurements are commonly composed of “cold” snow and ice in is a much more laborious task. the accumulation zone, with temperatures constantly 36 Glacier Mass Balance Monitoring Figure 3.1 Main Zonal, Vertical, and Mass Balance Regimes of Valley Glaciers Horizontal Regimes Mean Mean Net Accumulation Mixed ACC/ABL Net Ablation Annual Annual Air Temp Preciptation Elevation (m above (oC) sea level) Low Velocities Avalanche 7000 (except icefalls) High Velocities -25 Accum Medium Velocities Avalanching -20 6000 ow Cold -15 Vertical Regimes m n Zone of cu t S Surface Net Free Ac rec zing Temperatures (No net ablation) maximum Ice Fall Di No Sli -10 accumulation 5000 ding Clean ELA Dusty Ablation 1-6 weeks -5 Fre Ice ez July-August e/ 4000 M 0 elt Ablation Debris Net Melting 2-8 months Covered +5 Ice 3000 Bans al Sli ding +10 1000 2000 mm (we) Source: Hewitt 2007b. Note: ACC = accumulation; ABL = ablation; mm (we) = millimeters water equivalent. below the melting point of ice, and “warm” ice • Direct snowfall is not the main form of input to below, where seasonal melt and runoff occur. These glacier mass balance; two zones are separate by an ELA at approximately • Most valley glaciers in the HKH have a very 5,000 m that often coincides with the altitude of restricted accumulation zone and many have maximum annual snow accumulation, and the zone none at all; of maximum glacier surface area. The duration of • These glaciers are mainly or wholly avalanche the ablation season increases downward from the nourished; ELA, where there is minimal ablation, to the terminus • Avalanches and wind action intervene to where ablation may persist for several months each redistribute most of the snow that feeds glaciers; year – the “ablation gradient.” • ELAs rarely lie close to the snowline and, usually, are found hundreds of meters below it. To the In fact, the great majority of glaciers in the HKH extent they can be determined at all, they have depart more or less from the conventional picture. complicated geometries; The few exceptions can be valuable for establishing • Debris covers in ablation zone areas intervene to comparative and baseline data, as outlined regulate patterns and rates of melting. Extensive, below. In most cases, however, other and different heavy mantles suppress ablation, but even more conditions intervene to govern mass balance in this extensive areas with thin ones enhance it; region. They lie behind the fairly complicated picture • Mass balance gradients differ markedly from most in Figure 3.1 and are as follows: in the literature; 37 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains • Glacier movement rates rarely follow the Three very different types of glacier can be conventional pattern. Local sections of recognized in terms of nourishment: acceleration and deceleration are found within accumulation and ablation zones. Velocities • The snow-fed or “Alpine” type glaciers have further complicate relations to mass balance direct snowfall as the dominant input, large because they fluctuate widely on timescales from accumulation zones, and, usually, well-defined minutes to decades; and firn limits and ELA. Almost all mass balance • Sliding and block motion are predominant in studies assume this type; most of the valley glaciers. Such movement is • The avalanche-fed or “Turkestan” type glaciers sensitive to and varies with meltwater and ice are nourished more or less entirely by avalanches thermal conditions; it may or may not reflect mass of snow and ice from higher areas. Firn basins balance. The large elevation span of glaciers are small, absent, or not connected to the main and the role of avalanches and icefalls in the ice network. Main ice streams commence below rapid downslope movement of mass introduce the perennial snow zone (Figure 3.2). In the complications in vertical relations of mass HKH, countless small ice masses are of the “fall” balance and the relations (as yet unexamined) to glacier type, sustained by avalanching well below the rates of transformation to glacier ice and the the snowline; and thermal state of snow and ice. • The wind-fed type glaciers depend largely or wholly on snow redistributed and carried to them The founding documents and definitions of mass by wind action. Most are small, mainly high- balance work clearly recognized that such conditions elevation cornice apron and niche glaciers, or create distinct constraints on budgets, spatial and below the snowline on lee slopes. temporal patterns, and glacier dynamics (Meier 1962; Kasser 1967; Paterson 1994). To date, these Virtually all valley glaciers for which well-established factors have not been considered very important. mass balance records are available would fall Mass balance monitoring has tried to avoid into the Alpine class. The problem is that the other situations or glaciers where these factors exist. An two types are predominant in the HKH. In the assessment is needed as to what the implications are Karakoram, three quarters of the larger glaciers are for appropriate and effective monitoring in the HKH. largely or wholly avalanche fed (Hewitt 2011a). Most small and minor ice masses are predominantly wind 3.6.2 Accumulation and Source Zones fed or avalanche fed. However, while it is helpful to identify the distinct type of glacier, in many if not [T]he redistribution of snowfall by avalanching from most HKH glaciers all three forms of nourishment steep slopes, and wind scouring from exposed are found to occur. One form of nourishment may areas, can result in accumulation patterns that be predominant, but only a few glaciers are purely differ markedly from original climatically controlled of one type (Figure 3.3). distributions (Benn and Evans 1998, p. 79). As noted, the accumulation zone is usually defined as An unknown but large part of high-altitude snowfall an area of the glacier itself on which snow falls and is redistributed and modified by wind action. It serves survives from year to year as firn, to be transformed at to strip, redistribute, deposit, and compact snow. It depth into glacier ice (Paterson 1994). Some glaciers affects glacier conditions to a greater or lesser extent in the HKH have extensive accumulation zones. at all elevations, but especially around the interfluves They can be invaluable for monitoring strategies, as and the uppermost parts of glacier basins. A critical described below. However, the major inputs sustaining function is to prepare and feed the avalanche the ice in most glaciers are not direct snowfall. cascade to the ice streams that commence far below. 38 Glacier Mass Balance Monitoring Figure 3.2 Figure 3.3 Typical Avalanche-fed Glacier: Bazhin Glacier, Avalanche-nourished Sumaiyar Bar Tributary of Nanga Parbat East Face Barpu Glacier, Central Karakoram Source: Hewitt, July 2010. Note: The glacier begins in huge avalanche cones. There is virtually no continuous ice stream linking the ablation zone to the perennial snow and ice zone. The more obvious features are the cornices and wind slab areas. Also important is snow that drops out of the wind field on lee slopes and into gullies and couloirs where avalanches begin. The indirect evidence of the role of wind is not supported by any quantitative information, because so much occurs in the steepest parts of the basins and along narrow interfluves. In any case, avalanches and wind action intervene Source: Hewitt, 1986. between snowfall and ice to determine the spatial Note: Taken in August, the photo indicates that avalanching is an all-year condition. Note the extent of wind redistribution of snow at higher elevations. There patterns of inputs to the glaciers, their timing, is about 3,000 m of relief in the photo, from the foreground at 4,800 m to Mount and their characteristics. Most of this intervention Malubiting (7,458 m) in the top right background. between climate and glacier ice occurs above about 4,000 m and, in many parts, above 5,500 m down. The upper parts of glacier basins, rather elevation. Now, according to Paterson (1994, than firn basins, comprise largely rock walls too p. 55): “Climatic factors influence accumulation on steep for snow to remain on them. Much or all avalanche-fed glaciers only in so far as the size and of the snowfall is redistributed in time and space frequency of avalanches is climatically controlled.” before reaching the glaciers. Wind action serves mainly to redistribute snow laterally, compact However, nowhere has recognition been it, and help prime the avalanche cascade. An accompanied by research. Some of the constraints unknown but large fraction of all the avalanched and consequences for mass balance can be snow descends 1,000–1,500 m to reach the outlined; the neat separation of accumulation and glacier surfaces, where it becomes incorporated ablation zones in the conventional scheme breaks into glacier ice. In some ways, it is like transposing 39 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains the snow through some 10–15º of latitude. In a 3.6.3 High-Elevation Snowfall at budget year, it is likely that avalanches transfer Biafo Glacier, Central Karakoram quantities of mass downslope in glacier basins equivalent to and, in many cases, greater than that Biafo Glacier is an Alpine type glacier, which transported by glacier movement. is atypical for the region. However, it has vast, relatively gentle, and accessible accumulation Some nourishment – or in the strictly avalanche- basins between 4,700 m and 6,000 m elevation nourished (Turkestan) glaciers, all nourishment (Figure 3.4). They offer an opportunity to investigate – occurs in what is conventionally called the high-altitude snowfall at elevations where otherwise ablation zone. In the ablation zones, avalanched avalanching and wind redistribution prevail and, snow survives from year to year below where the since it is located in the heart of the Mustagh conventional snowline or firn limit would occur. Karakoram, it should offer insights for the most Significant ablation can take place side by side heavily glacierized areas. Snow accumulation was with significant inputs. Mass is added as well as observed there in the 1980s to identify sources of lost. The ELA is not at elevations where firn limits or precipitation, and seasonal, elevation- and storm- snowlines are found, but usually hundreds of meters related variables (Wake 1987; SIHP 1990). Such below them. The great effort that has been work has not been repeated. expended to determine snowlines and ELAs in the HKH – generally placed between 4,500 and Snowfall was measured between 4,800 and 6,500 m – looks at elevations where avalanche 5,800 m, the elevation zone comprising 70 percent and wind redistribution of snow are at their most of Biafo’s main, connected glacier system (Hewitt et frequent. This also challenges the assumption al. 1989; Hewitt 2005). Methods employed snow that these phenomena will respond to climatic pits and drill cores to establish vertical profiles, and temperature change in any obvious, direct way. to retrieve samples for snow chemistry and isotope analyses. Snow samples were taken systematically Wind and avalanches not only redistribute snow in time and space, they also alter its character. Figure 3.4 Wind-packed snow is much denser than snowfall. Biafo Glacier Accumulation Zone: Source of Snow Avalanche-deposited snow is usually even denser, Pit and Drill Core Samples often close to the densest firn. This speeds up transformation to glacier ice. Avalanches can carry large amounts of debris eroded from slopes and tend to be much dirtier than regular snow. They involve frictional and compressive warming, possibly melting, as they pass to lower elevations. It seems important to establish the above points. Nevertheless, a grasp is also needed of what the primary inputs of high-altitude snowfall involve. Some sense of this can be gained from work carried out in the Karakoram that has implications for the whole region. The following study of Biafo Glacier is Source: Hewitt, July 2010. the only known relatively comprehensive profile yet Note: View from Hispar Pass; middle ground is about 5,000 m elevation; Baintha Brakk peak in the background is 7,285 m. Snow pit and drill core samples were available for high-altitude snowfall as a contribution obtained from the gentler, more sheltered areas well removed from avalanched to mass balance. slopes (Wake 1987, 1989). 40 Glacier Mass Balance Monitoring by depth and weighed to determine their density Figure 3.5 and water equivalent. When compiled for an Snowfall (Water Equivalent) from Selected Sites on entire profile, estimates can be made of annual or Biafo Glacier and Adjacent Basins, 1983–88 budget year accumulations. Chemistry and isotope 6000 work helped to determine seasonal contributions and variability in more precise ways, and to Snow pit and trace precipitation sources. Some comparative borehole stations observations were made in the more limited firn } ‘Shark Col’ (5,660 m) basin areas of the Hispar and Khurdopin Glaciers. } ‘Upper Khurdopin’ (5,520 m) Elevation above sea level (m) Measurements confirmed the relatively heavy 5500 } ‘Hispar Dome’ (5450 m) snowfall at these elevations. Averages at all sites exceeded 1,000 mm water equivalent in a budget year, an order of magnitude greater than precipitation records for valley weather stations > ‘Approach Gl.’ (5,100 m) below 3,500 m. Yearly estimates ranged from a low of 850 mm (we) at one site to more than 2,300 mm at another. 5000 { Hispar Pass (5,060 m) > Lukpe Lawo (4,950 m) > ‘Whaleback’ Gl. (4,900 m) > Hispar East (4,850 m) The zone of maximum precipitation is of special interest. The data suggest it occurs above 4,900 m, hence it is entirely in the glacier > ‘EQ Line’ (4,650 m) accumulation zone. The broadest elevation range of concurrent data, 1984–86, put the highest 4500 inputs at sites between 4,900 and 5,100 m, with 0 0.5 1.0 1.5 2.0 2.5 some decline indicated above and below. Results in Accumulation water equivalent (m) other years and sites leave a distinct possibility that Key (balance year): maximum precipitation occurs at or continues up 1987-88 1986-87 1985-86 1984-85 1983-84 to 5,600 m, possibly higher. The greatest amounts, Source: Hewitt 2005 (after Wake 1987). recorded on the upper Khurdopin Glacier at 5,520 m, may involve “overcatch” – the result of wind-blown snow carried across the watershed. The Summer snow samples also include a monsoon more exposed “Shark Col” (Wake 1987) may be component in every year, indicated by isotopes, or subject to wind stripping (Figure 3.5). chemical signatures, coming from the vast lowland agricultural areas of the subcontinent (Wake 1987). Sources of precipitation show how glacier nourishment and health relate to climate systems Seasonal incidence of snowfall is another critical and may be affected by global climate change. concern. The measurements put a little over half Chemical signatures in snow at three higher the average high-altitude snow accumulation in Biafo sites show, as expected, that winter snowfall winter – but almost half in summer. The data are well comes from westerly sources; the Atlantic Ocean, constrained at the highest Biafo sites by stratigraphy, the Mediterranean, Black Sea, and Caspian Sea chemistry, and isotope signals. On average, (Wake 1987, p. 109). In late spring and early 45 percent of accumulation came in May– summer, Arabian Sea moisture may be drawn into September (Wake 1989). Summer contribution the Karakoram by the same westerly frontal storms. varied from 30 percent of annual precipitation to as 41 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains much as 78 percent, in contrast to valley weather Figure 3.6 stations lower down where it ranges from 20 to Accumulation Profile Exposed in a Crevasse, 45 percent (Fowler and Archer 2006). Biafo Glacier Considerable variability is seen from year to year in winter and summer contributions, according to air mass source, and at different elevations and sites. The snow pits at Biafo also confirmed field experience of the overwhelming role of a few large snowstorms, evident in the incidence of thick, uniform bands of snow, as illustrated in Figure 3.6. Heavy snowfalls were recorded high up in summer when little or none appeared in base camp records at 4,080 m or at valley weather stations. Before considering the mass balance implications, ablation conditions in the HKH must be appreciated. 3.6.4 Ablation in the HKH In most mountain glaciers, ablation losses are dominated by surface melting below a certain elevation and are controlled by energy exchanges with the atmosphere. The HKH is not unusual in this regard, but important departures from the Photo: Hewitt, 1985. Note: Although chemical signatures were spoiled, the profile indicates the conventional picture are noted. On the one hand, contrasting clean winter and dirty summer layers. The scale of summer inputs is surface melting is overwhelmingly driven by received masked by intermittent melting and refreezing in ice layers. Year-to-year variability and the role of single storm events are evident. solar radiation; on the other, various conditions intervene to modify where and how it operates, in particular the following: cloudy versus sunny days, and the incidence of snowfall in the ablation zone. • Surface albedo. This varies seasonally and following new snowfalls, but the most important With respect to the energy relation of ablation, it difference is between relatively clean ice and should be noted that even off-ice weather stations debris-mantled ice; beside the glaciers, or the few of these that have • Debris thickness. In addition to the effect on been maintained, do not exactly represent on-ice albedo, there is a critical thickness at which conditions (Figure 3.7). In the main ablation season, ablation rates are the same as for clean ice, for similar surface conditions, ablation rates are above which they are progressively less and nearly the same over most of the ablation zone. below which they are greater. There is also a Differences between conditions and water yields at critical thinness, usually of a few mm, where dust different elevations mainly relate to the length of or dirt veneers support the highest ablation rates; ablation season and debris covers. The most critical and conditions that intervene between atmosphere and • Summer weather. Surface melting is affected ablation relate to the extent and different thicknesses by summer weather, especially the numbers of of debris covers. 42 Glacier Mass Balance Monitoring Figure 3.7 Ablation Season Weather Observations for On-ice and Off-ice Stations at Same Elevation and 1.5 km Apart at Baintha Profile, Biafo Glacier, 4,050 m, 1986 26 24 22 20 18 16 14 Temperature 12 10 8 6 4 2 0 -2 -4 -6 Cloud Cover 1/100 m 25 30.1 5 10 15 20 25 June July Day Off-ice station Off-ice station Max Max Mean Mean Min Min Source: Based on unpublished Snow and Ice Hydrology Project data. 3.7 Debris-covered Glaciers absorption of debris is offset by lower conductivity through it to the ice surface. For progressively thinner Supraglacial debris has a major influence on covers, ablation tends to increase to a second ablation as a result of the combined influences critical thickness, usually a few mm, at which the of solar radiation and large but highly variable highest rates occur – commonly 1.5–2 times greater amounts of debris delivered to the glaciers by than for clean ice, and in some cases up to four avalanches, rock falls, wind-blown dust, and icefalls. times greater (Adhikari, Marshall, and Huybrechts Unlike most aspects of mass balance, ablation of 2009). Ablation rates again decline below this debris-mantled ice has been much investigated in critical thickness, as more of the ice is exposed. the HKH, although almost all the work has focused Scattered particles will enhance ablation compared on heavy debris covers (Muller 1958; Mattson and to clean ice. Rates vary somewhat with surface Hewitt 2005; Shroder and Bishop 2010; Scherler, roughness, the weathering that occurs with rapid Bookhagen, and Strecker 2011). ablation, and surface slope and local topography. With debris that exceeds about 10 cm, ablation rates The research has revealed critical thicknesses of decline to a fourth critical thickness, at which the debris that affect ablation. At a certain thickness, penetration of warmth approaches zero in a given ablation rates are about the same as for clean year and little or no ablation can occur. In detail, ice, roughly 5–6 centimeters (cm). The higher heat the ablation rates under apparently heavy debris 43 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Figure 3.8 Figure 3.9 Debris Cover on Ablation Zone of Baltoro Glacier, Light, Scattered Debris, Upper Baltoro Glacier, Central Karakoram, June Representative of about Two-thirds of the Ablation Zone, July Photo: Hewitt, 2005. Photo: Hewitt, 2005. cover vary considerably. Impressions are deceptive: in total. In the Karakoram, they average about what looks “heavy” is rarely more than 50 cm on two-thirds of ablation zone areas. They tend to be average and, except in the most slow-moving ice, at a higher elevation with a shorter ablation season; is highly variable in thickness and constantly shifting nevertheless, hydrographs show they are far more around (Hewitt 2005; Mihalcea et al. 2006). For important for total ablation losses, glacier health, mass balance, it is important to know the distribution and water supply. The thinner covers are also more of different debris covers. Generally the presence of sensitive to weather and climate change. relatively thin covers (of, say, less than 10 cm) is hard to discern within the heavy covers or to discriminate 3.8 Water Yield from Glaciers within the critical thicknesses of about 3–10 cm, where ablation rates can be as high as bare ice, The greatest interest in HKH glaciers is as a source or higher, and may be changing with quite small of water. It is important, therefore, to be clear about increases or decreases in thickness. Debris covers which water yields are being measured or estimated. vary considerably over much of the ablation zones The main, if not exclusive, focus of conventional (Figures 3.8 and 3.9). No satisfactory method has mass balance studies is ablation of glacier ice and yet been devised to determine debris thicknesses and ablation zone outputs. For the heavily glacierized types, or effects on ablation rates, except by direct areas of the HKH, this is surely the largest source of observation. It would be beneficial to find better water. When glacier ice is exposed at the surface, ways to identify and classify surface conditions on ice under the energy conditions that apply in summer from satellite imagery, which does show something between, say, 3,500 and 5,500 m elevation, of the diversity involved. ablation yields can far exceed precipitation. A common range for specific losses is 4–8 m water Heavy debris protects the ice in the lowest, warmest equivalent in a year. At a site at 2,900 m on Batura areas with the longest ablation seasons. It is widely Glacier, ablation loss exceeded 18 m in a year believed this phenomenon is decisive for glacier (Batura Glacier Investigation Group 1979). This behavior, responses to climate change, and water reflects a long ablation season, with some ablation supply, but the view is difficult to accept. The on 311 days of the year and a thin debris cover that areas of clean, dusty, and dirty ice are larger increases ablation rates. 44 Glacier Mass Balance Monitoring Of course, other sources of water exist in glacier Glacier ice exposed to ablation explains why, in basins. If gauging stations are at or very close to a much of the HKH, 80–90 percent of water yields glacier terminus, it is still necessary to determine the from glacier basins occur in two to three months relative contributions of several sources to isolate the of the year. In this sense, it is the “golden egg” of mass balance component. The various sources can glacier hydrology and constitutes the so-called include the following: Himalayan “water towers.” However, how the egg is peeled, so to speak, is important in year-to-year • Ablation of glacier ice; and longer variations, something in which the other • Melting of seasonal snowfall, on the glacier factors are critical. Unlike rainfall–runoff or snowfall– ablation zone and ice-free areas; snowmelt relations, the roles of annual precipitation • Melting of the seasonal freeze–thaw carapace and temperatures, so often the best indirect guides, that develops on subtropical glacier surfaces; are buffered by intervening conditions in glacier • Rainfall within the basin; and basins and can even seem unrelated. Each ablation • Rock glaciers and degrading permafrost in season unfolds according to how these various tributary valleys within glacier basins. factors work together. Only ablation and melting of the carapace pertain For peak yields, area–altitude relations are critical. entirely to the glacier and its mass balance, At lower elevations, heavy debris covers suppress while seasonal snowmelt and rainfall involve ablation throughout the year. This adds to the precipitation in on-ice and off-ice areas. The scale importance of what happens in middle to upper of contributions from glacier basins in the HKH ablation zone areas. In these areas, for ice ablation probably follows the order of the list, subject to to start, seasonal snowfalls and the cold season variations in different parts of the region, at different freeze–thaw carapace must first be removed to times of the year, and with climate change. While expose glacier ice with thin or no debris covers. ice ablation tends to be a much larger factor Hardly any data have been collected on this from larger glacierized areas, the other sources phenomenon, but experience at Biafo suggests these nevertheless involve amounts and variations that are not trivial aspects, especially in year-to-year equal or exceed those implied by climate changes variations (see section 3.14). over a decade or more. It should be added that conditions on small or minor Discharge hydrographs most directly reflect seasonal glaciers and areas where they prevail are likely to conditions in ablation zones, moderated by the be different yet again. The larger yields from their length of the ablation season, which varies vertically basins are likely to come from seasonal snowfall, depending on the following factors: directly or as redistributed by wind and avalanches. Because smaller ice masses generally occur with the • Annual migration of above-zero temperatures lower mountain watersheds, seasonal snowmelt and with elevation; rainfall assume a much greater significance in water • Area of ice exposed to ablation, usually supply and climate responses. Their hydrology differs increasing through the middle ablation zone; accordingly. • How long glacier ice is exposed, which tends to decline with elevation and according to seasonal Another difficult issue for water resources assessment snowfall and the extent of the freeze–thaw based on river flows is that peak yields from the carapace; HKH glaciers coincide with the summer monsoon. • Distribution of debris mantles and the spectrum of In the three major river basins of interest this is when thick versus thin covers; and rainfall runoff dominates flows of most sections, • Ablation season weather. especially towards the heavily populated lowlands. 45 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 3.9 Glacier Regimes Karakoram was treated as having the same. As described above, however, high-altitude summer Glaciers in world mountain regions have been snowfall is almost equal to that in winter. It might classified and investigated in terms of nourishment, suggest an intermediate type but involves an seasonal conditions, mass balance gradients, activity accumulation regime at least as distinctly different indices, and thermal regime or morphological types. as the inner tropics type, which in some respects Few other areas confront us with or prepare us for it resembles. Alternatively, the Karakoram annual the diversity of types in the HKH. For example, the budget can identify a distinctive, fourth regime – a criteria that identify glaciers with distinct climatic year-round accumulation and summer ablation regimes elsewhere – from polar and temperate to type, which also results in two different outer equatorial, or in maritime versus continental – are all tropics types. found within this one region. As the discussion has shown, accumulation, Mass balance regimes refer to the annual pattern even ignoring the intervening roles of wind and of gains and losses identified with the seasonal avalanche, is the part of HKH mass balance least incidence of accumulation and ablation. They relate well understood or measured. Glacier basins in to broad patterns of controls over glacier behavior all areas receive winter snowfall for which there is and water yields, and some potential responses to very little information and for which measurement is climate change. They vary substantially across the sorely needed. It may be seriously underestimated HKH, with important implications for monitoring. for high elevations in the summer accumulation Much has been written about the “summer areas. Records for weather stations are usually accumulation regime” of the Greater Himalaya, far away from the glacier areas, mainly at lower where summer monsoon snowfall dominates primary elevations; and the data of primary interest, snowfall, inputs (Ageta 2001). However, regimes change from are difficult to collect as snow gauges give the least west to east and north to south, partly as a function reliable of precipitation data. of latitude, but mainly through relations to different moisture-bearing air masses and the Tibetan plateau. Regime classes emphasize annual patterns. Of Globally, several distinct mass balance regimes have equal or greater importance in the HKH are spatial been identified for valley glaciers, mainly on the patterns, especially the vertical organization of mass basis of latitude or zonal climates of the following balance. types (Kaser and Osmaston 2002, p. 25): 3.10 Mass Balance Gradients • Inner tropics type, with two-season or year-round accumulation and year-round ablation; Mass balance gradients reveal the amount and • Outer tropics type, with summer accumulation shares of inputs or outputs occurring at different and strong summer and weak or no winter elevations. In the conventional scheme, they are ablation; and shown to vary systematically with elevation (Benn • Mid-latitudes type, with winter accumulation and and Evans 1998, p. 78–79). Ablation losses are summer ablation. shown as greatest at or near the terminus, declining upwards to a well-defined ELA. Accumulation The Greater Himalaya range from Kashmir to increases progressively above that. The gradients Sikkim can be placed in the outer tropics type with, usually appear as straight lines or nearly so. as noted, a summer accumulation regime. The Differences between glaciers and regions are Hindu Kush and Pamirs have a mid-latitude regime reflected in their slope, or highest versus least dominated by winter snowfall. In the past, the ablation and accumulation values. The highest are 46 Glacier Mass Balance Monitoring typically identified with more humid, usually maritime 3.11 Verticality glaciers. A narrower spread reflects drier, colder, more continental settings. In the high mountains, mass balance is as much about spatial as seasonal regimes. The latter depend Mass balance gradients for Karakoram glaciers mainly on regional climates, the former on altitude, hardly fit the conventional picture, with considerable elevation range or relief, topography, and orientation diversity of profiles. Even Biafo, an Alpine type of glacier basins. In the main concentrations glacier that might be expected to have a fairly of glaciers of the HKH, the larger masses have conventional profile, does not. The very low value exceptional elevations and relief (vertical span). near the terminus is mainly an effect of debris cover and ablation increases slowly through the The Karakoram glacial zone spans more than lower 15 km of the main glacier (compare Inoue 6,300 m vertically, that is, from the summit of K2 1977). Ablation expands to a maximum in the (8,610 m) to the lowest glacier termini in the Hunza middle ablation zone and declines sharply through valley, which reach down to 2,300 m. No individual the upper ablation zone. Through the lower basin spans the entire whole range, but several accumulation zone, inputs increase sharply but are basins span over 5,200 m, and at least 40 more expected to decline at the highest elevations. An than 3,000 m. Two of the larger Everest region S-shaped profile results, much different from the glaciers, the Khumbu in Nepal and the Kangshung near-straight ones typical of Alpine type glaciers on the Tibetan side, commence on Mount Everest elsewhere. (8,848 m), higher than any Karakoram glacier, but terminate at 4,800 and 4,560 m, respectively, a Missing from the Biafo curve are the roles of wind range of just over 4,000 m. For many of the larger redistribution and avalanche inputs. Since, in Karakoram glaciers, available relief is the main a conventional approach, mass balance inputs constraint of elevation range, whereas mass balance apply only where the glacier itself is and in effect is more critical for the Everest glaciers. The latter, to the main connected glacier, most sources in the unlike many Karakoram glaciers, terminate where perennial snow zone of avalanche-fed glaciers valleys continue steeply down, as do glaciers on the are excluded from mass balance calculations and west and south faces of Nanga Parbat. The well- gradient. Where the accumulation zone would known Rakhiot Glacier, which descends from the normally be, wind and avalanche movement of summit of Nanga Parbat (8,125 m) to 3,070 m, has snow and ice prevail. In avalanche-fed glaciers a vertical span of almost 4,500 m. The terrain drops the S-shape remains but is unlike anything in the over 7,000 m in 21 km. However, the terminus lies conventional picture. Most or all net inputs occur some 2,000 m above the Indus River, to which the below snowlines or firn limits, so that both positive valley falls steeply. and negative curves are in the ablation zone. Altitude itself is a factor in that, for any given part Virtually no work has been done on just how and of the HKH, the higher the mountains the greater where inputs occur from snow and ice avalanches the ice cover, and the further ice streams extend and from wind-redistributed snow. Clearly, some downslope. For mass balance, glacier dynamics, of the most widely applied notions relating to mass and responses to climate change, it is also necessary balance need reconsideration. At the heart of the to address the extreme range of environments that problem is how elevation, hypsometry, steepness, the elevation spans imply. The scope of avalanching, and ruggedness intervene to reconfigure the icefalls, and meltwater drainage accelerate processes determining the relation of the mountain linkages between elevation zones with very “climate” to glaciers and glacier dynamics. different conditions. The relations to mass balance 47 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains parameters involve the various controls in which The scope of such verticality relations in the HKH – elevation, hypsometry, and topography (steepness, notably compared to those of the mountain glaciers ruggedness, basin orientation) are important. whose mass balance guides current practices and analyses of global conditions – necessitates attention The term “verticality” encompasses these various to conditions and events in certain elevation zones. aspects, while helping to emphasize how they work By way of illustration, on-ice and glacier nourishment together. Verticality encompasses issues “relating to conditions in the upper Indus basin vary with or composed of elements at different levels” and “of, elevation as follows: constituting, or resulting in, vertical combination”.5 As such, it speaks to conditions addressed in the • 4,500–8,000 m, where avalanching redistributes “vertical zonation” of Klimek and Starkel (1984), snowfall over the off-glacier terrain and wind the “altitudinal belts” of Ives, Messerli, and Spiess action plays a dominant role around interfluves (1997), and the “elevation effect” and “altitudinal and an important one in the preparation of organization” of Hewitt (1993, 2005). However, avalanches; in addition to the virtue of brevity, verticality • 4,800–6,000 m, the zone of maximum directs attention to the spectrum of spatial– precipitation and for accumulation in Alpine type physical relationships in which gravity is of primary glaciers; significance, drawing together conditions such as: • 3,800–5,500 m, where avalanche deposits provide most or all nourishment of Turkestan type • Vertical gradients. Environmental conditions glaciers; can change with height, notably temperature, • 3,500–4,500 m, the most extensive areas of pressure, and humidity; exposed glacier ice in ablation zones, where • Area–altitude distributions. The extent dust and thin dirt enhance ablation and from and share of conditions or features vary with which annual ablation losses and water yields are elevation, such as debris mantles or seasonal greatest; and snow cover; • 2,300–4,000 m, where the lowest ice tongues • Altitudinal zones. Conditions and processes terminate and most ice areas are protected can be concentrated in certain elevation zones, by heavy supraglacial debris, giving them a including ablation and accumulation zones; they comparatively small role in water yields and are generated in different ways, but with similar conservative response to climate change. effects to “zonal” climates, in different latitudes or according to distance from the ocean; Several other conditions will now be considered • Aspect. While orientation of mountain slopes as controls or aspects of glacier behavior that may is not itself about verticality, variations between affect or reconfigure mass balance parameters. The slopes change and increase with available relief factors are glacier motion, thermal classes, and and steepness, giving aspect a diversified and neglected seasons. intensified role; and • Vertical cascades. Connections occur upslope 3.12 Glacier Motion and downslope in which slope angles and steepness are key; mass balance and climatic The balance between inputs to and losses from responses relate to the vertical moisture and glaciers is maintained through ice movement. debris cascade and valley wind systems. Glacier motion is a crucial part of mass balance, 5 http://dictionary.reference.com/. 48 Glacier Mass Balance Monitoring but because conventional studies see it as a direct occurs, there is a boxlike velocity profile across the function of accumulation and ablation, it appears glacier and a well-defined line of shear at the ice straightforward. Velocities and throughputs are margin. In the larger glaciers, extensive ice stream shown as increasing in the accumulation zone sections appear to move and respond to changes towards maxima around the ELA, and decreasing as slablike units. They seem related to the sharp with ablation losses towards the terminus. HKH fluctuations in velocities observed in virtually all glaciers depart from this picture in various ways. timescales from minutes to decades. These suggest chronic instabilities, probably related to steepness, Three main types of motion have been established basal sediments, and thermal complexities, as for glaciers: internal creep, basal sliding, and well as meltwater availability. Such aspects of ice through deformation in soft, subglacial sediments. dynamics are important because they intervene to Glacier ice responds to applied stress above a create irregularities in response to changes in mass certain limit by permanent deformation. Microscopic balance and suggest why terminus fluctuations melting and recrystallization in the ice structure are may be unreliable indicators. In the HKH, there involved, and microshearing. The resulting internal may be little or no consistency between terminus creep is the characteristic form of motion, or “glacier advance and retreat in adjacent glaciers, or for flow.” Movement can also occur through basal those of apparently identical characteristics. It also sliding, usually considered important only in “warm” directs attention to another distinctive feature and or temperate ice, not frozen to the bed. Sliding complication in HKH ice masses, the thermal is substantially influenced by meltwater. Rates of state of ice. sliding movement are usually greater in the ablation season, a response to higher meltwater availability. 3.13 Thermal Classes Observations from HKH glaciers generally show winter movement rates are 20–50 percent less Two main thermal types of glacier recognized are than summer. warm (or temperate) and cold. Where ice is at or very close to pressure melting point throughout, The third component of ice motion can arise from glaciers are called “warm;” their surface layers deformation in soft, subglacial sediments. It is may become subzero for a time in winter. In “cold” thought to apply mainly, if not only, to unfrozen glaciers, all or most of the ice is below the pressure bed and sediment conditions, and to occur where melting point. There are two subcategories or basal sediments are in the sand to clay grades, or polythermal types: polar, where geothermal or where coarser particles are supported by a fine- frictional heating develops a zone of “warm” ice grained matrix. The quantities of debris suggest this at the bed; and subpolar, with a period of surface could be important in the HKH. Many lower glacier warming and ablation in summer (Benn and tongues sit on thick ramps of moraine with fines-rich Evans 1998). horizons. It has also been proposed that surge type glaciers, common in the Karakoram and Pamirs, As the terms suggest, these various thermal types may involve soft sediment deformation (Jiskoot have been named or attributed to latitudinal zones 2011). However, there are no actual measurements or zonal climates. However, in the HKH, a complete to establish the presence and extent of this type of range of possible thermal types is present. The movement. In many if not most HKH glaciers, sliding extreme cold at high altitude means glaciers have is the greater component of motion and it is in this properties identified with polar or subpolar types. context that block movement phenomena have Smaller, south-oriented glaciers and the lower been widely reported (von Klebelsberg 1925–26; tongues of many large glaciers have the warm Finsterwalder and Pillewizer 1939). Where sliding ice and unfrozen beds of the temperate thermal 49 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains type. This supports Hambrey’s (1994) view that most substantial fluctuations not driven by climate the thermal categories of polar and subpolar but related to other controls over mass balance are “misleading and best avoided.” For similar relations, most likely involving thermal instabilities reasons, in the HKH context, the same applies to (Hewitt 2007b). temperate, tropical, and subtropical categorizations – geographic designations for thermal phenomena 3.14 Neglected Seasons that are not confined to these regions or any particular region. In the Karakoram, for example, Mass balance is generally measured over a budget despite the limited number of observations, year and used to track annual changes. However, it is evident that the larger glaciers and many almost all investigations in glacier basins of the HKH intermediate ones are of mixed thermal regimes. have taken place in July and August, a very few as early as June or as late as September. For water The importance for mass balance concerns is how supply, summer appears the critical time. However, the thermal regime affects the following processes: can it be assumed that nothing important happens in the other nine to10 months? • The rate of deformation, with “warm” ice deforming faster than “cold” and affecting Only a handful of expeditions have left any kind velocities of ice flow; of instrumentation or sites to track events through • Meltwater production and distribution, depending the rest of the year; fewer still remained there to on glacier temperatures; observe what happens. Of course, the importance • Cold ice requires energy to raise it to the of winter snowfall in accumulation is acknowledged. melting point before ablation begins. There is no Avalanching varies seasonally in given elevation discharge in winter. Meltwaters tend to spread zones, but how it does so is little understood. Its inputs over to the margins of the glacier and travel to glacier masses continue throughout the year, but mainly in surface drainage lines; are affected by the vertical shifts in temperatures. • Warm ice develops subsurface drainage lines and, for glaciers that are warm throughout, Almost wholly ignored is whether anything meltwater more readily penetrates to the bed and significant happens to ablation zones outside the there is some discharge in winter; ablation season. New snowfall and snow covers • Type of motion: cold ice is frozen to the bed, and through the winter must be considered. Some there is usually no sliding motion; and important developments certainly occur in the • Warm ice is unfrozen, so that sliding can “shoulder seasons,” October–November and comprise a more or less large fraction of total April–May. The migration of a zone of frequent movement, and the amount of moisture reaching freeze–thaw cycles is crucial. Freeze–thaw and wind or generated at the bed influences this. action affect snow and the buildup of a carapace of icy layers and refrozen and wind-packed snow. Pertinent here is Paterson’s (1994, p. 337) Before ablation in the mass balance sense begins, observation that “only in temperate glaciers is the these layers must also be degraded. Slush flow effect of a climate change restricted to a change in activity can be important as a pre-ablation process, mass balance.” Changes in ice thermal regimes, or following the retreat of the snowpack on the glacier meltwater availability, can also lead to changes in up to and beyond the firn limit and onto the higher glacier behavior and extent without mass balance avalanche cones. change. The exceptional number of surging glaciers in some HKH ranges suggests mixed and thermally A similar case could be made for looking at two unstable conditions at the bed. Surges are the other largely neglected phenomena that affect mass 50 Glacier Mass Balance Monitoring balance relations. One concerns detached ice masses • Short-term observational projects of key variables at high elevation and ice avalanches from them. The and zones on glaciers to calibrate datasets other is the role of icefalls, which especially affect coming from existing, accessible, long-term middle and upper ablation zones. However, these are hydrometeorological stations and to provide all in the realm of detailed glacier processes, poorly ground control for remotely sensed parameters understood and raising questions of whether and just from glacier surfaces; how far monitoring should consider them or simply • More rigorous testing or calibration of the continue to ignore them. relations of such permanent meteorological and river gauging networks as indicators of glacier 3.15 Discussion contributions; and • Identification of better-located and more Attempts to derive mass balance estimates and representative stations. changes have been based largely on temperature and precipitation data extrapolated from weather Global benefits and efficiencies can be derived from stations outside the glacier zones, or climate models, information sharing, coordinated monitoring, and or sometimes on assumptions about snowlines cooperation across relevant administrative and state and ELAs. Conditions known to influence mass boundaries, such as the following: balance in the HKH, but largely lacking in direct measurements, include high-elevation snowfall, • Coordinated analyses of satellite imagery and avalanche and wind redistribution of that snow, ground-truthing for both broad glacier-related avalanche-fed glaciers, all-year conditions and parameters and glaciers of special concern; cycles in glacier basins, glacier thermal regimes, and • Coordination through joint projects with glacier movement. institutions of higher education and research in the country concerned; 3.15.1 Field Programs and Instrumentation • International collaboration on glacier and related hydrometeorological research projects to address Two strategies that are the norm in regions with outstanding scientific and technical questions well-established monitoring may not work in the regarding the glaciers; and HKH: a set of benchmark glaciers or a glacier • International collaboration on courses, diplomas, network. Both imply mass balance monitoring for higher degrees, and visiting professionals in whole glaciers. The former has succeeded mainly glacier-related scientific fields. by choosing small, relatively simple glaciers that nonetheless seem representative for the region. It is 3.15.2 Personnel and Safety doubtful this approach can work in the HKH, which is an environment that requires strategic engagement In each country, direct observations will require one between field and indirect approaches. Glaciers or more teams trained and permanently ready to would need to be chosen for their suitability for work in glacierized areas, whether to repeat and training, ground control, historical reconstruction update key observations or to investigate hazardous of glacier change, and experimental efforts. The situations and events. None of this is likely to value must be decided of the following suggested happen or be successful without a core of personnel approaches: experienced in mountain environments, usually with mountaineering and winter skills, and enthusiastic • Identification and setup of selected glacier about the environment and the work. In this context, basins for experiments, testing of instruments, important and special problems of personnel safety procedures, and training; and training must be considered. 51 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Looming over any plans for direct monitoring References of glaciers is the fact that field activities are constrained by often large problems of funding, Adhikari, S., S.J. Marshall, and P. Huybrechts. personnel, training, instrumentation, and safety. 2009. “A Comparison of Different Methods of Working conditions on glaciers generally require Evaluating Glacier Response Characteristics: constant attention to safety – in the preparations, Application to Glacier AX010, Nepal, Himalaya.” equipment, work schedules, and health care The Cryosphere 3: 765–804. http://www.the- and fitness of personnel involved. In most of the cryosphere-discuss.net/3/765/ 2009/tcd-3-765- HKH, safety needs and imperatives are unusually 2009.pdf. high. While it is unlikely that any monitoring programs will extend to the highest and steepest Ageta, Y. 2001. “Study Project on the Recent areas, there is hardly anywhere that teams will not Shrinkage of Summer Accumulation Type Glaciers in the Himalayas, 1997–1999.” Bulletin of need to acclimatize to working altitudes and be Glaciological Research 18: 45–49. equipped with appropriate clothing and gear for harsh weather and traveling over difficult terrain. Bajracharya, S.R., P.K. Mool, and B.R. Shrestha. Specific kinds of safety issues arise with crevasses, 2007. Impact of Climate Change on Himalayan avalanches, icefalls, and accumulation zones Glaciers and Glacial Lakes: Case Studies on where there is deep snow. Such matters should be GLOF and Associated Hazards in Nepal and addressed in planning and training with the help Bhutan. Kathmandu: ICIMOD. of qualified and experienced mountaineers. Batura Glacier Investigation Group. 1979. “The Everywhere that well-developed snow and ice Batura Glacier in the Karakoram Mountains and Its Variations.” Scientia Sinica 22 (8): 958–74. monitoring and glacier investigations exist, the people who play the largest part have two distinct Benn, D.I., T. Bolch, K. Dennis, J. Gulley, A. but necessary attributes. First, there are the Luckman, K.L. Nicholson, D. Quincey, relevant qualifications in technical, engineering, S. Thompson, R. Toumi, and S. Wiseman. 2012. or scientific fields. Second, there is mountain “Response of Debris-Covered Glaciers in the experience, at least basic mountaineering Mount Everest Region to Recent Warming, skills, and enthusiasm for working in mountain and Implications for Outburst Flood Hazards.” environments. Personnel who have worked Earth-Science Reviews 114 (1–2): 156–74. doi: everywhere that well-developed snow and ice 10.1016/j.earscirev.2012.03.008. monitoring and glacier investigations exist have Benn, D.I., and D.J.A. Evans. 1998. Glaciers and two necessary attributes: relevant qualifications in Glaciation. London: Arnold. technical, engineering, or scientific fields; and at least basic mountaineering skills and enthusiasm Bhambri, R., and T. Bolch. 2009. “Glacier Mapping: for working in mountain environments. For A Review with Special Reference to the Indian anyone not so equipped, working on glaciers Himalayas.” Progress in Physical Geography 33 can seem extremely difficult, frightening, and (5): 672–704. dangerous, in which case the turnover will be high and the work may be done improperly Bhambri, R., T. Bolch, R.K. Chaujar, and S.C. or not at all. Kulshreshtha. 2011. “Glacier Changes in the Garhwal Himalayas, India 1968–2006 Based on Remote Sensing.” Journal of Glaciology 57 (203): 543–56. 52 Glacier Mass Balance Monitoring Bolch, T. 2011. “Benchmark Glaciers.” In Gardelle, J., E. Berthier, and Y. Arnaud. 2012. Encyclopaedia of Snow, Ice and Glaciers, ed. V.P. “Slight Mass Gain of Karakoram Glaciers in the Singh, P. Singh, and U.K. Haritashaya, 95–98. Early Twenty-First Century.” Nature Geoscience 5: Dordrecht, Netherlands: Springer-Verlag. 322–25. Bolch T., T. Pieczonka, and D.I. Ben. 2011. “Multi- Haeberli, W. 2011. “Glacier Mass Balance.” In decadal Mass Loss of Glaciers in the Everest Area Encyclopaedia of Snow, Ice and Glaciers, ed. V.P. (Nepal Himalaya) Derived from Stereo Imagery.” Singh, P. Singh, and U.K. Haritashaya, 399–405. The Cryosphere 5: 349–58. Dordrecht, Netherlands: Springer-Verlag. Chao-hai, L., S. Ya-feng, Z. Wang, and X. Zi-chu. Hambrey, M.J. 1994. Glacial Environments. London: 2000. “Glacier Resources and Their Distributive University College London Press. Characteristics in China: A Review on ‘Chinese Glacier Inventory’.” Journal of Glaciology and Harper, J.T., and N.F. Humphrey. 2003. “High Geocryology 22 (2): 106–12. Altitude Himalayan Climate Inferred from Glacial Ice Flux.” Geophysical Research Letters 30 (14): Cogley, J.G. 2011. “Present and Future States of 1764. doi: 10.1029/2003GL017329. Himalayan and Karakoram Glaciers.” Annals of Glaciology 52 (59): 69–73. Hewitt, K. 1993. “The Altitudinal Distribution of Karakoram Geomorphic Processes and Dainelli, G. 1959. Esploratori e Alpinisti nel Depositional Environments.” In Himalaya to Caracorum. Turin: Unione Tipografico-Editrice the Sea: Geology, Geomorphology and the Torinese. Quaternary, ed. J.F. Shroder Jr., 159–83. New York: Routledge. Dyurgerov, M.B., and M.F. Meier. 1997. “Mass Balance of Mountain and Sub-polar Glaciers: A Hewitt, K. 2005. “The Karakoram Anomaly? Glacier Global Assessment for 1961–1990.” Arctic and Expansion and the ‘Elevation Effect,’ Karakoram Alpine Research 29 (4): 379–91. Himalaya, Inner Asia.” Mountain Research and Development 25 (4): 332–40. Finsterwalder, R., and W. Pillewizer. 1939. “Photogrammetric Studies of Glaciers of High Hewitt, K. 2007a. “Rediscovering Colonised Asia.” Himalayan Journal 11: 107–13. Landscapes: The First Europeans at the Mustagh Pass, Karakoram Himalaya, Inner Asia.” In The Fountain, A.G., M.J. Hoffman, F.D. Granshaw, Exploitation of the Landscape of Central and Inner and J. Riedel. 2009. “The ‘Benchmark Glacier’ Asia, ed. M. Gervers, U. Bulag, and G. Long, Concept: Does It Work? Lessons from the North 41–67. Toronto Studies in Central and Inner Asia Cascade Range, USA.” Annals of Glaciology 50: 9. Toronto: University of Toronto, Asian Institute. 163–68. Hewitt, K. 2007b. “Tributary Glacier Surges: An Fowler, H.J., and D.R. Archer. 2006. “Conflicting Exceptional Concentration at Panmah Glacier, Signals of Climatic Change in the Upper Indus Karakoram Himalaya.” Journal of Glaciology 53 Basin.” Journal of Climate 19 (17): 4276–93. (181): 181–88. Fujita, K., L.G. Thompson, Y. Ageta, T. Yasunari, Hewitt, K. 2011a. “Glacier Change, Concentration Y. Kajikawa, and T. Takeuchi. 2006. “Thirty- and Elevation Effects in the Karakoram Himalaya, Year History of Glacier Melting in the Nepal Upper Indus Basin.” Mountain Research and Himalayas.” Journal of Geophysical Research Development 31 (3): 188–200. http://dx.doi. 111: 6. org/10.1659/MRD-JOURNAL-D-11-00020.1. 53 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Hewitt, K. 2011b. “Glaciers of the Karakoram Kick, W. 1989. “The Decline of the Last Little Ice Age Himalaya.” In Encyclopaedia of Snow, Ice and in High Asia Compared with That in the Alps.” In Glaciers, ed. V.P. Singh, P. Singh, and U.K. Glacier Fluctuations and Climate Change, ed. J. Haritashaya, 429–36. Dordrecht, Netherlands: Oerlemans, 129–42. Dordrecht: Kluwer. Springer-Verlag. Klimek, K., and L. Starkel. 1984. Vertical Zonality Hewitt, K., C.P. Wake, G.J. Young, and C. David. in the Southern Khangai Mountains (Mongolia). 1989. “Hydrological Investigations at Biafo Geographical Studies 136. Wrocław: Zakład Glacier, Karakoram Himalaya: An Important Narodowy im. Ossolinskich. Source of Water for the Indus River.” Annals of Glaciology 13: 103–8. Korzhenevsky, N.L. 1930. Glacier Inventory of Central Asia [Katalog Lednikov Sredney Asii]. Inoue, J. 1977. “Mass Budget of Khumbu Glacier.” Tashkent. Seppyo Special Issue 39: 15–19. Kotlyakov, V.M., A.M. Dyakova, V.S. Koryakin, V.I. Ives, J.D., B. Messerli, and E. Spiess. 1997. Kravtsova, G.B. Osipova, G.M. Varnakova, et “Mountains of the World: A Global Priority:” In al. 2010. “Glaciers of the Former Soviet Union.” Mountains of the World: A Global Priority, ed. B. In Satellite Image Atlas of Glaciers of the World: Messerli and J.D. Ives. New York: Parthenon. Glaciers of Asia, eds. R.S. Williams Jr. and J.G. Ferrigno, 1–93. Washington, DC: United States Ives, J.D., R.B. Shrestha, and P.K. Mool. 2010. Geological Survey, U.S. Government Printing Formation of Glacial Lakes in the Hindu Office. Kush-Himalayas and GLOF Risk Assessment. Kathmandu: ICIMOD. Mason, K. 1930. “The Glaciers of the Karakoram and Neighbourhood.” Records of the Geological Jiskoot, H. 2011. “Dynamics of Glaciers.” In Survey of India 63 (2): 214–78. Encyclopaedia of Snow, Ice and Glaciers, ed. V.P. Singh, P. Singh, and U.K. Haritashaya, 245–56. Mason, K. 1954. Abode of Snow. London: Dalton. Dordrecht, Netherlands: Springer-Verlag. Mattson, L.E., and J.S. Gardner. 1989. “Energy Kaser, G., and H. Osmaston. 2002. Tropical Exchanges and Ablation Rates on the Debris- Glaciers. Cambridge, United Kingdom: Covered Rakhiot Glacier, Pakistan.” Zeitschrift Cambridge University Press. für Gletscherkunde und Glazialgeologie 25 (1): 17–32. Kasser, P. 1967. Fluctuations of Glaciers, 1959–1969. Brussels, Belgium: International Mayer, C., A. Lambrecht, M. Belo, C. Smiraglia, and Association of Hydrological Sciences. G. Diolaiuti. 2006. “Glaciological Characteristics of the Ablation Zone of Baltoro Glacier, Kaul, M.K. 1999. Inventory of the Himalaya Karakoram, Pakistan.” Annals of Glaciology 43: Glaciers: A Contribution to the International 123–31. Hydrological Programme. GSI Special Publication 34. Kolkata: Indian Geological Survey. Meier, M.F. 1962. “Proposed Definitions for Glacier Mass Budget Terms.” Journal of Glaciology 4: Kayastha, R.B., and S.P. Harrison. 2008. “Changes 252–61. of the Equilibrium-Line Altitude Since the Little Ice Age in the Nepalese Himalaya.” Annals of Glaciology 48: 93–99. 54 Glacier Mass Balance Monitoring Mihalcea C., C. Mayer, G. Diolaiuti, A. Lambrecht, Shih, Y., S. Liu, D. Shangguan, D. Li, and B. Ye. C. Smiraglia, and G. Tartari. 2006. “Ice Ablation 2006. “Peculiar Phenomena Regarding Climatic and Meteorological Conditions on the Debris- and Glacial Variations on the Tibetan Plateau.” Covered Area of Baltoro Glacier, Karakoram, Annals of Glaciology 41: 106–10. Pakistan.” Annals of Glaciology 43 (1): 292–300. Shih, Y., D. Mi, T. Yao, Q. Zeng, and C. Liu. 2010. Muller, F. 1958. “Eight Months of Glacier and Soil “Glaciers of China.” In Satellite Image Atlas Research in the Everest Region.” In The Mountain of Glaciers of the World: Glaciers of Asia, ed. World, ed. M. Barnes, 191–208. New York: R.S. Williams Jr. and J.G. Ferrigno, 127–66. Harper and Brothers. Washington, DC: United States Geological Survey, U.S. Government Printing Office. Oerlemans, J. 2001. Glaciers and Climate Change. London: Taylor and Francis. Shroder, J.F., and M.P. Bishop. 2010. “Glaciers of Pakistan.” In Satellite Image Atlas of Glaciers of Paterson, W.S.B. 1994. The Physics of Glaciers, 3rd the World: Glaciers of Asia, ed. R.S. Williams edition. Oxford, United Kingdom: Elsevier. Jr. and J.G. Ferrigno. Washington, DC: United Qiu, Jane. 2008. “China: The Third Pole.” Nature States Geological Survey, U.S. Government 454: 393–96. doi: 10.1038/454393a. Printing Office. Raina, V.K. 2009. Himalayan Glaciers: A State-of- SIHP (Snow and Ice Hydrology Project). 1990. Art Review of Glacial Studies, Glacial Retreat Snow and Ice Hydrology Project: Final Report, and Climate Change. New Delhi: Ministry of Volume IV. WAPDA/WLU/ICIMOD, Cold Regions Environment and Forests. Research Centre, Wilfrid Laurier University, Canada. Raina, V.K., and D. Srivastava. 2008. Glacier Atlas of India. Bangalore: Geological Society of India. Smiraglia, C., C. Mayer, C. Mihalcea, G. Diolaiuti, M. Belo, and G. Vassena. 2008. “Himalayan- Ren, Jiawen, Zhefan Jing, Jianchen Pu, and Xiang Karakoram Glaciers: Results and Problems in Qin. 2006. “Glacier Variations and Climate the Study of Recent Variations of Major Non- Change in the Central Himalaya over the Past polar Glaciers.” In Terra Glacialis, Special Issue: Few Decades.” Annals of Glaciology 41 (1): Mountain Glaciers and Climate Changes in the 218–22. Last Century, ed. L. Bonardi, 149–64. Milan: Servizio Glaciologico Lombardo. Scherler, D., B. Bookhagen, and M.R. Strecker. 2011. “Spatially Variable Response of Himalayan Tandong, Yao. 2007. Map of the Glaciers and Lakes Glaciers to Climate Change Affected by Debris on the Tibetan Plateau and Adjoining Regions. Cover.” Nature Geoscience 4: 156–59. doi: Xi’an, China: Xi’an Cartographic Publishing 10.1038/ngeo1068ce. House. Schneider, H.J. 1969. “Minapin: Gletscher und UNESCO (United Nations Educational, Scientific Menschen im NW-Karakorum.” Die Erde 100: and Cultural Organization). 1998. Into the 266–68. Second Century of Worldwide Glacier Monitoring: Prospects and Strategies, ed. W. Haeberli, M. Shi, Yafeng. 2001. “Estimation of the Water Hoelzle, and S. Suter. Studies and Reports in Resources Affected by Climatic Warming and Hydrology 56. Paris: UNESCO. Glacier Shrinkage Before 2050 in West China.” Journal of Glaciology and Geocryology 23 (4): 333–41. 55 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Untersteiner, N. 1957. “Glazial-Meteorologische Wiche, E. 1959. “Klimamorphologische Untersuchungen im Karakoram.” Archiv für Unterschungen im Westlichen Karakoram.” Meteorologie, Geophysik und Bioklimatologie Verhandlungen des Deutschen Geographentages Serie B, 8 (1): 1–30 and (2): 137–71. 32: 190–203. von Klebelsberg, R. 1925–26. “Der Turkestanische Williams, R.S. Jr., and J.G. Ferrigno, eds. 2010. Gletschertypus.“ Zeitschrift für Gletscherkunde Satellite Image Atlas of Glaciers of the World: 14: 193–209. Glaciers of Asia. Washington, DC: United States Geological Survey, U.S. Government Printing Wake, C.P. 1987. Spatial and Temporal Variation Office. of Snow Accumulation in the Central Karakoram, Northern Pakistan. Master’s thesis. Waterloo, Young, G.J., and K. Hewitt. 1993. Canada: Wilfrid Laurier University. “Glaciohydrological Features of the Karakoram Himalaya: Measurement Possibilities and Wake, C.P . 1989. “Glaciochemical Investigations as Constraints.” In Snow and Glacier Hydrology: a Tool for Determining the Spatial and Seasonal Proceedings of the International Symposium, Variation of Snow Accumulation in the Central Kathmandu, Nepal, November 16–21, 1992, Karakorum, Northern Pakistan.” Annals of ed. G.J. Young, 273–83. IAHS Publication 218. Glaciology 13: 279–84. Wallingford, United Kingdom: International WGMS (World Glacier Monitoring Service). 2009. Association of Hydrological Sciences. Glacier Mass Balance Bulletin. Bulletin 10 (2006– 2007). Zurich, Switzerland: WGMS. 56 Mountain Hydrology 4. Mountain Hydrology 4.1 Background to Mountain model, the inputs typically are not simply entered Hydrology into the model but must themselves first be modeled. This necessity arises from the fact (discussed earlier) Mountain hydrology is defined here as the that measurement of precipitation and energy inputs methodologies associated with the monitoring can often not be made where they are most needed and measurement of the water balance of the but only where they are technically feasible – usually HKH mountain catchment basins. Traditionally, in the accessible mountain valleys. Thus, what in the hydrological monitoring undertaken for purposes standard models is merely a processing of inputs of water resources planning or management has (typically a simple or weighted areal averaging of been based on “rainfall–runoff” or “black box” point measurements) becomes input modeling in correlation modeling, in which input, as measured mountain hydrology. Its aim is to estimate, from the precipitation, is correlated with output, as measured scarce and ineffectively located point measurements streamflow, to provide an estimate of the timing of precipitation and energy components, the areal and volume of streamflow from a basin. This type of and elevation distributions of: (a) precipitation modeling produces useful information for engineers amounts; (b) precipitation form; and (c) energy (or at and water managers concerned with lowland rivers least temperature). originating in the mountain basins. This modeling approach, however, provides relatively little insight The hydrological regime of the HKH mountain into questions about the relative contributions of catchment basins is not well studied or understood. rain, snow, or glacier melt to streamflow volumes, or The countries of the Indus and Ganges basins the role of glacier retreat and climate change in the customarily treat all data describing streamflow as streamflow regimes of the great rivers of South Asia, confidential. While this practice varies among the a major ongoing debate. countries of India, Nepal, and Pakistan, none of the three have produced readily available, digitized Two important features distinguish mountain monitoring records of streamflow that would hydrology models. The first is that the input is not permit assessments of the hydrological variability identical to precipitation, as is the case in other of the catchment basins. Because much current hydrological models, but also includes energy concern is centered on the role of the glaciers in inputs as an indispensable component. While in the streamflow production and the potential impact of classical rainfall–runoff model, energy inputs are glacier retreat for the region, discussion of the role required only to determine the evapotranspiration of glaciers in the mountain hydrology is included output from the basin, in mountain hydrology here (see chapter 3 for a full discussion of glaciers). models energy inputs are essential to determine the Given the general lack of analysis, however, the “active” portion of the total precipitation input, that primary emphasis in this chapter is to illustrate the is, to separate the liquid part, which immediately various hydrological relationships along the east– contributes to runoff, from the solid part, which west transect of the HKH Mountains. remains temporarily inactive in storage. The volume and timing of runoff as measured in The second feature, less obvious but perhaps more the annual hydrograph are indicators of the nature, important, is that, unlike in a standard hydrological timing, and volume of the water and energy budgets 57 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains of a basin. In the absence of an extensive climate The literature provides only the most cursory analyses monitoring network, the hydrograph provides the of the hydrological regimes of the HKH Mountains, most direct link to any assessment of the effect of and the number of serious studies is quite limited. climate change on streamflow from a catchment Drawing a clear distinction between the runoff basin by reflecting variations in the climate-related volumes resulting from snowmelt and glacier melt is water and energy exchange processes. The precise difficult. If data reported in the literature are correct, form of the annual hydrograph will be strongly the primary zone of meltwater from both sources is influenced by location, as well as the time period maximized at around 4,000–5,000 m as a result in which data are collected. Data based on daily of a combination of maximum terrain surface area, values in close proximity to a glacier or the snowline maximum glacier surface area, and maximum snow during the melt season will reflect the diurnal freeze– water equivalent deposition that occurs there. This is thaw cycle that is characteristic of both snowmelt the altitudinal zone generally reached by the upward and glacier melt during summer months, while those migration of the freezing level during the months of at increasing distances will be increasingly subdued July and August, which is also the time of maximum as dilution from other sources – seasonal snowmelt, runoff. The transition from snowmelt to ice melt rainfall – with distance downstream in a basin. that occurs during this period may be a result of a transition from snowmelt to glacier melt as the winter The Himalaya is characterized by a complex snow disappears. Additional studies of the factors three-dimensional mosaic of meteorological and determining the volumes of snowmelt and ice melt hydrological environments, in a geography ranging would be required to permit a definitive distinction from tropical rain forests to arctic deserts and in between the two. Even the most cursory analyses, an altitudinal range of more than 8,000 m. A few however, demonstrate the increasing importance reliable maps of the region exist and essential of snowmelt in the extreme western portions of climate and hydrological data are often not readily these mountains as the summer season monsoon, available. With the lack of a basic understanding dominating the hydrology of the eastern Himalaya in of runoff sources and timing in the rivers of South Nepal, is replaced by the summer melting of snow and Central Asia, usage issues related to the water deposits resulting from winter westerlies in the Hindu budget, such as glacier retreat, cannot be resolved. Kush ranges of northern Pakistan. And the general unavailability of data describing the hydrology, climate, and topography makes it From the standpoint of studies and monitoring of difficult to apply hydrological concepts and models components of the mountain water budget, the developed for mountain catchments in Europe or major concern of the region should be water, not North America. glaciers. A factor in the hydrology of the Himalayan mountain chain, glaciers may have a profound The hydrometeorological environments of the HKH impact on life at the scale of mountain villages between eastern Nepal and Afghanistan are defined but are less important factors at the scale of the primarily by two major seasonal air masses – the major river basins of the region. In the recent past, southeast monsoon in the eastern and central emphasis has been placed on the relationship Himalaya and winter westerlies in the Hindu Kush between climate change and glacier retreat in the and Karakoram – interacting with the 8,000 m of mountains of Asia, with limited attempts to define the mountain relief. This interaction involves primarily climate of these mountains, or on the relationship variations of water and energy transfer with respect between glaciers and climate there. Recent to the topographic variables of altitude, aspect, and discussions have accepted the conventional wisdom slope as both air mass properties and mountain that the mountain climates and glaciers of the terrain vary from east to west. Himalaya may be described by the Alpine models 58 Mountain Hydrology with which Western science is most comfortable, Himalaya questionable. At the scale of the global and that the scales of global circulation models and circulation model, it is generally not possible to mountain catchment basins are compatible. Both undertake more than two-dimensional analyses of these interpretations are most likely in error. spatial climate variation, involving only latitude and longitude. At the scale of the mountain catchment Both the Indus and the Ganges are transnational basin, an analysis of the relationship between rivers. The Indus River has headwaters in Tibet climate and any water budget variable will require a (China), India, Afghanistan, and the northwest three-dimensional model that includes altitude. territories of Pakistan. The Ganges River arises primarily in India, with an estimated 10 percent 4.2 Monitored Streamflow contributed by the central Himalaya, 50 percent of the HKH Mountains from the lowlands to the south of the mountains, and 40 percent from Nepal. Uses of this water are Ultimately, the concern for the future of the glaciers those customary everywhere: hydroelectric energy of the upper Indus basin is a concern for the water generation, multipurpose reservoir management, resources the river represents. As a component of irrigation, urban and industrial water supplies, and the hydrology of the mountain headwaters of the recreation. Primary problems are contamination, mountain basins, it can be expected that changes flooding, and drought. in the glaciers will be reflected in changes in the volume and timing of runoff from the mountain No single monitoring network exists that will provide basins. In seeking answers to the concerns related to useful input to the management of each of these the potential impact of climate change and glacier uses or problems. Each will have differing needs in retreat on runoff from the mountain basins, very terms of variables to be monitored, and the scale little use has been made of the existing records of appropriate to the problem. measured runoff from these basins. Much of the ongoing debate concerning the role 4.2.1 The Indus River of glaciers in the volume and timing of the flow of Asian rivers has emphasized air temperature at The Indus River is transnational, with headwater the expense of any serious consideration of the tributaries in the countries of Afghanistan, China total water and energy budgets of the Himalayan catchments. Given the relative lack of climate (Tibet), India, and Pakistan. The river originates studies of large mountain ranges, climate-glacier north of the Great Himalaya on the Tibetan plateau. relationships have been based on models at the The main stem of the river runs through the Ladakh scale of the global circulation models, in many cases district of Jammu and Kashmir and then enters the driven by satellite-derived data, at a comparable northern areas of Pakistan (Gilgit-Baltistan), flowing scale. between the western Himalaya and the Karakoram Mountains. Along this reach of the river, streamflow A typical global circulation model has a grid cell volume is increased by gauged tributaries entering size of 105 km2 at 40°N latitude. At the same time, the main river from catchments in the Karakoram most Himalayan catchment basins are on the order Mountains – the Shyok, Shigar,6 Hunza, Gilgit, and, of 103 km2. This great disparity in scales makes in the western Himalaya, the Astore (Young and the findings of recent comparisons of climate and Hewitt 1993), as well as ungauged basins on the hydrological models for the catchment basins of the north slope of the western Himalaya. Immediately 6 The gauging station for the Shigar basin has reportedly been discontinued (personal communication, D. Archer, 2010). 59 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains north of Nanga Parbat, the westernmost of the Figure 4.1 high peaks of the Himalaya, the river turns in a Mountain Catchment Basins of the Indus River southerly direction and flows along the entire length of Pakistan, to emerge into the Arabian Sea near the port city of Karachi in Sindh. Tributaries to this reach of the river from the western Himalaya are the Jhelum, Chenab, Ravi, and Sutlej Rivers, from the Indian states of Jammu Kashmir and Himachal Pradesh, and the Kabul, Swat, and Chitral Rivers from the Hindu Kush Mountains. The total length of the river is about 3,180 km, and its total drainage area exceeds 1,165,000 km2. The river’s estimated annual flow at the mouth is 207 cubic kilometers (km3), making it the 21st largest river in the world in terms of annual flow. Figure 4.1 is a digital elevation model of the gauged Source: B. Krumwiede, in Alford 2011. mountain catchment basins of the upper Indus River, Note: The speckled blue area is the approximate area of glaciers and perennial snowfields. in which the highest altitudes are dark brown and the lowlands are green. Gauging sites are shown as red circles, and the general area of glaciers is shown as Table 4.1 shows the general topographic and stippled blue and white. The geomorphometric data hydrological characteristics of the gauged basins of on which this map is based – latitude, longitude, the upper Indus River. and altitude – provide the terrain variables used in the calculation of the water and energy budgets of The general hydrology of the lower Indus basin is Karakoram glaciers. assumed to be reasonably well understood as a Table 4.1 Descriptive Statistics of the Basins Considered in the Study River Basin Gauge site Area, km2 Specific Annual Average runoff, mm streamflow, altitude, m million m3 Indus Astore Doyen 3,988 1,291 5,184 3,981 Gilgit Gilgit 12,680 692 8,777 4,056 Hunza Dainyor 13,732 761 10,448 4,516 Shigar Shigar 6,922 917 6,350 4,611 Shyok Kiris 33,350 312 10,705 5,083 Indus Besham 166,096 440 71,679 4,536 Chitral Chitral 12,504 712 8,118 4,120 Jhelum Dhangalli 27,122 1,075 29,156 2,628 Chenab Aknoor 22,422 1,222 27,424 3,542 Source: Alford 2011. 60 Mountain Hydrology result of a network of gauging stations, reservoirs Figure 4.2 such as the Tarbela and Mangla, and irrigation Diversity of Annual Streamflow from Catchments in barrages on the piedmont immediately south of the the Upper Indus Basin, One Year mountains. While this network provides data on which to base management decisions concerning 300 mm Hunza 600 mm Batura C H INA Upper Hunza water uses in the lower basin, the hydrology of 200 PA K ISTAN 100 100 mm 320mm INDIA the upper basin remains largely undefined. Under 0 1050 mm 0 J 1974/5 D 400 J 1970 D normal circumstances, this is not a problem. In 200 200 Gilgit KA recent years, however, concerns regarding the mm 100 a RA 1570 Gilg it Hunz possible impacts on rivers such as the Indus as a 730 0 mm 0 mm ilgit K J 1974/5 D N J 1970 D G O Braldu result of climate changes in the mountain headwater Indus R RA Skardu M o Shy regions have become increasingly alarmist (see Tarbela 200 mm Indus Ind ok R Upper us for example Rees and Collins 2006; IPCC 2007, reservoir 100 0 370 100 mm mm Indus 240 R R Leh 0 chapter 10.6), with only minimal data from the mm Islamabado J 1970 D J 1970 D mountain headwaters for support. Without a better understanding of the timing and sources of runoff -32oN from the catchment basins of the upper Indus, the 100 km nature and severity of any climate change impacts 72oE cannot be assessed with confidence. Source: Goudie et al. 1984. The general outlines of the hydrology of the upper Note: This pattern illustrates the fact that major sources of runoff in the upper Indus basin are quite localized and average values applied to the entire basin will have Indus basin have been defined in recent studies by little relevance in assessing the potential impacts of climate change. foreign geoscientists (for example, Goudie et al. 1984; Young and Hewitt 1993). Streamflow gauging stations are maintained by Pakistan’s WAPDA • The two principal sources of runoff from the (Figure 4.2). upper Indus basin are: (a) winter precipitation, as snow, which melts the following summer; 4.2.2 Upper Indus Basin Hydrology and (b) glacier melt. Winter precipitation is most important in producing seasonal snow runoff The characteristics of upper Indus basin hydrology volume; summer temperature most affects glacier are as follows: melt volume; • Variability in the main stem of the Indus, based • The mean annual flow of the upper Indus basin is on the record from Besham, has ranged from approximately 72 km3 from the main stem above approximately 85 percent to 140 percent of the Tarbela reservoir, 29 km3 from the Jhelum basin, period of record mean of 72 km3; and 27 km3 from the Chenab basin, a total of 128 • The wide diversity of hydrological regimes in km3. The total estimated flow of the Indus River at the mountain basins complicates the problem its mouth is slightly more than 200 km3 annually; of relating streamflow timing and volumes to a • The total surface area of the main stem of uniform climate change; and the Indus above Tarbela is approximately • The mountain headwaters of the Indus River 166,000 km2, with an estimated glacier area of contribute approximately 60 percent of the mean approximately 17,000 km2. The other glacierized annual total flow of the river, with approximately basin, the Chenab in the western Himalaya, has 80 percent of this volume entering the river a surface area of 22,500 km2 and a glacier area system during the summer months of June– of 2,700 km2; September. 61 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 4.2.3 The Nepal Himalaya from the glacierized basins, and approximately 3 percent of the total annual streamflow from Three major tributaries of the Ganges have Nepal into the Ganges basin. This ice melt headwaters primarily in the Nepal Himalaya: the presumably will enter the normal hydrological Karnali, Narayani, and Sapta Kosi Rivers. The total budget of the respective basins, some fraction average annual gauged streamflow volume from the becoming storage, some fraction evaporating, three rivers is approximately 145 km3 (Nepal DHM and the remainder becoming runoff (Alford et al. 1988). Total streamflow from Nepal, including that 2009); from the ungauged lowlands to the south of the • In the eastern and central Himalaya (Nepal mountains, has been estimated at 200 km3 annually and India to Himachal Pradesh), the volume of (Sharma 1983). The topographic and hydrological runoff will be determined by the strength of the characteristics of the Nepal gauged basins with a southeast monsoon (which will determine the glacier cover are shown in Table 4.2. orographic rainfall pattern) and the area–altitude distribution (hypsometry). Most precipitation will An analysis of the existing hydrological and be convective, on windward slopes, reflecting glaciological data for the Nepal Himalaya (Nepal the importance of aspect in determining runoff DHM 1988; Mool et al. 2001) indicates that volume and timing. Timing will be determined by glaciers are not a major factor in determining the arrival – and persistence – of the southeast the volume of flow in the rivers of South Asia. For monsoon. The predictability of the southeast glacierized catchments of Nepal that also had monsoon has declined during the last decade hydrometric data available, the approximate findings and the monsoon regime has become more were: erratic; • The eastern HKH basins of Nepal contribute • Calculated total ice melt volume from glaciers in 40 percent to the total flow of the Ganges River these basins was approximately 4.5 billion m3, as measured at Farraka, approximately 500 km3; approximately 5 percent of the total streamflow runoff from the Indian Himalaya contributes Table 4.2 Descriptive Statistics of the Glacierized Catchment Basins of the Nepal Himalaya River Basin DHM Area, km2 Ave. Qb, m3/s qb, mm Qv, million ID# altitude, m m3 Karnali Bheri 270 13,677 4,400 435 1,116 13,718 Narayani Kali Gandaki 420 6,553 3,200 267 1,270 8,420 Marsyangdi 439 4,781 4,200 212 1,737 6,686 Budhi Gandaki 445 3,707 5,400 169 1,182 5,048 Trisuli 447 3,623 5,200 173 1,382 5,456 Sapta Kosi Tama Kosi 647 2,382 4,900 145 1,661 4,573 Likhu Khola 660 1,297 3,500 57 2,184 1,798 Dudh Kosi 670 4,515 4,400 223 1,715 7,033 Tamor 690 6,330 2,600 336 1,879 10,596 Total 46,164 63,328 Source: Alford et al. 2009; data from Nepal DHM 1988. Note: DHM = Nepal Department of Hydrology and Meteorology (ID no.); Qb = streamflow volume, m3/s; qb = specific runoff, mm; Qv = annual streamflow volume, million m3. 62 Mountain Hydrology 10 percent, and 50 percent comes from runoff Figure 4.3 originating south of the river. The Indian portion Recession Curves for Glacierized Basins of the of the HKH Mountains contributes approximately Karakoram, Based on Mean Monthly Data for 50 km3 to the estimated total flow of 200 km3 of July–December the Indus River (Sharma 1983); Recession Curves • Current monitoring in Nepal involves weather 350 stations and streamflow gauging at altitudes generally below 1,000 m above sea level. No 300 public data are available for Indian HKH basins Specific Runoff, mm 250 (Bookhagen and Burbank 2010); and • It is estimated that runoff from the glaciers of the 200 Nepal portion of the Himalaya contributes less than 5 percent of the total flow of the Ganges 150 River. No public data are available for Indian 100 HKH basins (Alford et al. 2009). 50 4.2.4 Recession Flows 0 0 2 4 6 8 The recession curve is a characteristic feature of Months a hydrological basin, reflecting the relationship Chitral Shyok Shigar between inputs, in this case snow and ice melt, that Astore Gilgit Hunza determine the nature of peak flow; and storage, Source: Alford 2011. which determines the slope of the curve between the peak flows, and the annual low flow. The recession curve represents a powerful forecast tool, Figure 4.4 once the hydrological characteristics of the basin Recession Curves for Besham, Based on Mean have been determined. In the upper Indus basin, Monthly Data for July–December these characteristics are determined primarily by Recession Curves – Besham the relative importance of snow and ice melt and 12000.0 summer season temperatures. There is minimal postmelt storage in any of the gauged basins for 10000.0 which data are available. Recession curves for glacierized basins of the Karakoram, based on mean 8000.0 monthly data for the months July to December, are shown in Figure 4.3. Two basins, the Gilgit and m3/s 6000.0 Astore, with a glacier cover of about 10 percent, have a single summer peak in July, followed by 4000.0 recession to a base flow level. The remainder, with glacier covers of 20–50 percent, has a flood peak 2000.0 persisting through July and August, on average. 0.0 A comparison of recession curves for a single Jul Aug Sep Oct Nov Dec Months hydrometric station – Besham Qila, on the Indus Max Mean Min River main stem, immediately above the Tarbela Source: Alford 2011. reservoir – is shown in Figure 4.4. The Besham 63 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains recession curves for years with maximum, mean, of the HKH Mountains, resulting in decreasing and minimum volumes of flow are typical of most runoff values. Recent studies have considered the basins in the western HKH Mountains. For all entire transect, using satellite imagery such as flows exceeding the long-term mean value, the MODIS in the extreme western portion (for example, peak flow occurs in July, presumably reflecting the Immerzeel, Beek, and Bierkens 2010), or using period during which the maximum surface area the Tropical Rainfall Monitoring Mission (TRMM) of snow-covered terrain is producing meltwater. in the eastern and central portions of the range The peak flow occurs in August during years of (for example, Bookhagen and Burbank 2010). The lower-than-average flows, presumably as a result values of specific runoff – from the data maintained of a decreased winter snowpack, and a greater by Pakistan’s WAPDA in the west, and the Nepal dependence on glacier melt runoff. In all cases, Department of Hydrology and Meteorology in the the bulk of the annual streamflow volume occurs east – serve as a useful check on these estimates in a period of about 90 days, and base flows are based on data from satellite imagery. The sudden independent of the volume of the peak flow. change from the monsoon-driven hydrological regimes of the eastern basins of the Nepal and 4.2.5 East–West Variation in Runoff much of Indian portions of the HKH, to the snowmelt and glacier melt hydrology of the upper Indus basin, It is generally recognized that there is an east–west can be seen in the change between the Astore basin, gradient of water exchange in the catchment basins located on the eastern flank of Nanga Parbat, the westernmost Himalayan peak, and the Shyok basin, to the north in the eastern Karakoram Range (Figure 4.5). Figure 4.5 East–West Variation in Specific Runoff in HKH 4.2.6 Altitudinal Gradients of Runoff 2500 While the negative orographic air temperature lapse rate characterizing the HKH Mountains has 2000 been accepted without question (see chapter 2), the existence of an orographic gradient defining Specific runoff, mm/yr the water budget of the mountain catchments has 1500 received much less consideration. It is becoming increasingly apparent, however, that the areal 1000 distribution of both precipitation and runoff may vary widely over the surface of a Himalayan 500 catchment basin. While the variation may be of minor importance in determing water supply for the adjaceent lowlands, for an analysis of water budget 0 or climate-related questions, such as the role of git nza us yok e ab eri ki i i i si si la or k gd sul tor Ko Ko da da ho Tam glaciers in streamflow formation, it is paramount. Ind en Bh Gil an Tri Hu Sh As an an LK D. T. Ch rsy KG BG While Figure 4.6 is based primarily on data from the Ma Basin Nepal portion of the Himalaya, a comparison of the Source: Data from WAPDA files, transcribed by Archer, 2010; and Nepal DHM 1988. trend shown is essentially duplicated by data from Note: The figure shows east–west variation in mean annual specific runoff, in millimeters, along a transect extending from the Tamor basin in eastern Nepal to the Karakoram at altitudes above 4,000 m the Gilgit basin in the upper Indus basin. (Alford 2011). 64 Mountain Hydrology Figure 4.6 conditions might be considered simple models in Regional Orographic Runoff Gradient for the themselves, as they have a predictive capacity. Himalaya Based on Data from Glacierized and The number of examples that follow are by no Nonglacierized Basins means exhaustive. Mean Annual Specific Runoff with Altitude 4.3 Assessing Comparative 7000 Contribution to Streamflow 6000 Because runoffs resulting from glacier melt, from Mean Basin Altitude (m) 5000 melt of seasonal snow, and from monsoon rainfall generally coincide in time, it has proved most 4000 difficult to establish the comparative contributions 3000 from each and how each of them differ across the HKH. Majeed, Hussain, and Asghar (2010) used a 2000 simple technique of hydrograph separation to assess the direct rainfall contribution to inflow of the Indus 1000 to Tarbela reservoir, a catchment of some 200,000 0 km2. They found that significant monsoon rainfall 0 500 1000 1500 2000 2500 2000 only occurs in the lowest part of the catchment, and Mean Specific Runoff (mm) the streamflow response shows sharp runoff spikes Source: Data from Nepal Department of Hydrology and Meteorology 1988. overlying the prolonged seasonal melt peak. They Note: While there is an overall trend to decreasing runoff depths with increasing assessed the rainfall contribution by identifying and altitude, it can be seen that there is a distinct difference between eastern (red) and western (blue) basins. While both are curvilinear, runoff from the eastern basins is calculating the volume of these spikes in relation to greater at all altitudes below 4,000 m than in the west. the underlying snowmelt (and glacier) hydrograph. Their conclusion was that the maximum annual 4.2.7 Initial Uses of the Existing Network rainfall contribution was 20 percent, the minimum was 1 percent, and the average was 5 percent of A modern tendency in hydrological analysis is first the total inflow. Majeed (1995) similarly calculated to consider building a model or to apply a new set the contribution of direct rainfall to Mangla reservoir of data to an existing model. However, all models, on the Jhelum River as generally in the range of no matter how detailed, will always remain a highly 10 to 20 percent but, in the most extreme case, simplistic representation of real-world complexity 35 percent. (Oreskes, Shrader-Frechette, and Belitz 1994). They inevitably reflect a specific set of assumptions about Although it might be anticipated that model outputs dominant hydrological processes (Van Dijk 2011). would generally correspond to these boundaries, In the extreme, “all models are wrong, but some are they do not necessarily do so. Based on a model useful.” With respect to the HKH, still little is known output, Bookhagen and Burbank (2010) suggest of the dominant processes in runoff generation at an average rainfall contribution to the Indus of levels above 3,000 m. 26 percent and to the Jhelum of 65 percent. The magnitude of the different estimates illustrates the Before proceeding to describing a full distributed level of uncertainty in basic understanding of model of a catchment, it is profitable to apply HKH hydrology. preliminary data analysis to establish boundary conditions and also as a means of establishing Analysis by Archer and Fowler (2008) tends to the reliability of the data. Some of these support the Majeed interpretation, although it does 65 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains not specifically attempt to assess the comparative melt based on a regional ablation gradient (Haefeli runoff from rainfall and snowmelt. The authors 1962) and the area elevation curve of the glacier examined the relationships between climate from its terminus to the highest annual altitude parameters and runoff. The analysis found no reached by the mean 0ºC isotherm. Melt summed significant relationship between summer rainfall in elevation bands over this area of the glacier is and summer runoff for the main Jhelum inflow assumed to represent the net annual loss of mass to Mangla. On the other hand, there is a very to the glacier; it does not represent the loss of significant correlation (at 1 percent level) between seasonal snow. Glacier melt was then compared winter precipitation and summer runoff for the with catchment runoff computed in a similar way as Jhelum and its principal tributaries. More puzzling a summation of runoff from altitudinal belts, in turn is a very significant negative correlation between based on gauged flow data. preceding winter and spring temperature and summer runoff, and a weak but consistent negative Alford and Armstrong (2010) conclude that the correlation between summer temperature and contribution of glacier annual meltwater to annual summer runoff. streamflow in the Ganges basin from the glacierized catchments of the Nepal Himalaya represents Archer (2003) similarly used regression analysis approximately 4 percent of the total annual between climate variables and runoff to show that streamflow volume. While it is widely believed that the controlling factors in runoff were essentially a serious decline of glaciers and resulting loss of different between high-level subcatchments of the runoff from them would result in major downstream Indus, with runoff primarily derived from glacier rivers from the Himalaya becoming seasonal melt, and catchments where runoff was primarily (IPCC 2007), this assessment suggests that glacier derived from melt of seasonal snow. Summer runoff meltwater is not a major factor in determining the from the glacier-fed catchments is controlled by volume of rivers flowing from the Nepal Himalaya. concurrent temperature (and not at all by winter However, this study is by no means definitive and precipitation), while runoff from seasonal snowmelt- requires more detailed field data from glaciers and fed catchments is controlled by preceding winter and downstream river flows to clarify regional and spring precipitation (and not at all by concurrent local variations. temperature). While this analysis does not attempt to calculate the comparative contribution of glaciers Much of the concern about changes in the water and seasonal snow, it does provide an indication, resources of rivers downstream from the HKH has via the strength of the respective correlations, of arisen from studies of what might happen to glaciers the comparative contribution of glacier melt and and to resulting runoff if they were to experience seasonal snowmelt to catchment runoff. Given projected global increases in temperature. The the consistent patterns of correlation between paper by Rees and Collins (2005) is particularly independently measured climate variables and pertinent, having been quoted in a World Bank streamflow, this analysis also serves to indicate study of water problems in Pakistan by Briscoe and that both sets of data are reliable, at least at Qamar (2007). Rees and Collins applied a rate of a seasonal level. 0.06°C per year transient climatic warming to two hypothetical glaciers in the eastern and western Alford, Armstrong, and Racoviteanu (2009), Alford HKH. Their modeled results suggest a rapid decline and Armstrong (2010), and Alford (2011) attempted in glacier volume accompanied by an initial sharp to distinguish the glacier contribution to total flow increase in river flow, lasting perhaps for five in basins in Nepal and Pakistan (Figure 4.7). They decades but followed by a sharp decline in flows first used a simple methodology to compute glacier on the main stem of the Indus by 30–40 percent 66 Mountain Hydrology Figure 4.7 Estimated Glacier Melt Contribution to Total Annual Flow, HKH Mountains Comparative Glacier Melt and Total Runoff Volumes Gauged Basins - HKH Mountains 35000 30000 25000 20000 15000 10000 5000 0 or git nza r yok e ab eri ki i si sul iga tor Ko da Tam en Bh Gil Tri Hu Sh Sh As an D. Ch KG Total, mcm Glacier, mcm Source: Alford, Armstrong, and Racoviteanu 2009; Alford 2011. Note: The figure shows a transect of the estimated glacier component of runoff from selected basins of the Nepal Himalaya and the upper Indus basin. by the end of the century. They also concluded that declining proportion of glacial contribution to the impacts of declining glacier area on river flow would main stem of the Indus. On lower subcatchments, be greater in the west (Pakistan), where precipitation depending on melt of seasonal snow, annual flow is scarce, while summer snowfall in the east would has predominantly increased, with several stations suppress the rate of initial flow increase, delay peak exhibiting statistically significant positive trends. discharge, and postpone eventual disappearance of Analysis of timing of the annual hydrograph using the ice. Such a scenario prompted understandable spring onset date and center of volume date concerns, described by Briscoe and Qamar (2007) indicated no clear trends, in direct contrast to what as “terrifying.” has been observed in western North America. The results of the analysis indicate that the magnitude Despite these concerns, little attempt has been and timing of the streamflow hydrograph in made as yet to assess historical trends in streamflow, general is influenced both by the seasonally varying which, given the global increase in temperature over snowpack and by trends in temperature. the past several decades, might be expected already to show signs of increasing glacial runoff and a It is now recognized that historical climate trends change in timing of those flows. Sharif et al. (2012) in the upper Indus basin have not mirrored global have examined trends in magnitude and timing of trends (Fowler and Archer 2006). Over the period flows at 19 gauging stations in the upper Indus. 1960–2000, summer temperature trends were The response is quite opposite to that suggested by generally downward, although winter and annual Rees and Collins (2005) in that high-level glacial trends were rising. It is now also clear that trans- catchments show a falling trend in runoff and a Himalayan glaciers behave quite differently from 67 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains those in the eastern and central Himalaya, where insight into many aspects of climate change debates. significant retreat and depletion of glacier volume Coupled with basic procedures describing energy has occurred (Eriksson et al. 2009; Berthier et al. exchange, such as application of the degree-day 2007). Thickening and advance have been reported and ablation gradient concept, it may be possible to in many Karakoram glaciers in recent decades develop preliminary hypotheses concerning the water (Hewitt 2005, 2007, 2011). Analysis of ice loss budget and components of streamflow formation using satellite gravimetry during 2003–09 (Matsuo within a mountain basin, as well as the relationships and Heki 2010) seems to confirm that glacier between climate change, glacier retreat, and glacier loss is lower in the Karakoram compared to the meltwater runoff volumes. neighboring Himalaya. The goals of a hydrological monitoring program There is now an appearance of a measure of must be clearly defined. The design of a stream agreement in estimates involving relationships gauging network intended primaily to provide data among climate, glaciers, and river flow in the on the timing and volume of mountain runoff for trans-Himalayan upper Indus. However, there is a downstream uses, such as irrigation, hydroelectic tentative indication that some of these relationships energy generation, or domestic water supplies, may not have applied in the first half of the 20th will differ from one intended to supply data for a century, when summer monsoon rainfall appears water balance analysis that includes the role of to have played a bigger role in the flow of the glaciers in runoff formation within the mountain Jhelum River. It is uncertain as to how far eastward basin. If the goal is forecasting of water supplies for these relationships and trends extend, given the lowland users, then the existing monitoring network change to a monsoon-driven climate and hydrology. is probably sufficient; what will be required is better Shrivastava (2011) indicates that between 1960 and management and analysis of the data. If the goal is the mid-1990s, over a broad stretch of the Indian a better understanding of the relationship between Himalaya, the trend in summer temperatures has climate, glaciers, and streamflow, installation of also been downward, a trend clearly at variance with additional hydrometric stations in the immediate the observed glacier loss. Trends in the magnitude vicinity of selected glacier termini, at altitudes and timing of runoff response do not appear to have between 3,000 and 5,000 m, would be useful for been investigated. any studies dealing with the mesoscale hydrological regime of the HKH catchment basins. The whole region would benefit from a more comprehensive review of climate, glacier mass It would be simpler to resolve problems related to balance, and river flow, building upon extensive supply and use of runoff from the mountain basins work already carried out by Singh and colleagues if all countries of the HKH Mountains prepared (see, for example, Singh, Ramasastri, and Kumar digital databases of streamflow measurements and 1995; Singh and Kumar 1996, 1997a, 1997b; made these data generally available on request by Singh et al. 2000). Modeling, using satellite remotely legitimate data users. sensed data, would clearly benefit from a close inspection of such analyses. 4.4 Streamflow Monitoring A large, essentially unanalyzed amount of streamflow “Streamflow” is the combined result of all data is available for the Nepal and Pakistan climatological geographic factors that operate in catchment basins of the HKH Mountains. This simple a drainage basin (Herschy 1995). It is the only assessment of the data demonstrates that a deeper phase of the hydrological cycle in which the water investigation of these data might yield considerable is confined in such a way as to permit accurate 68 Mountain Hydrology measurement of the overall catchment response. network, has not yet been considered in addressing Other measurements are point measurements climate change and comparative contribution issues. for which the uncertainties on an areal basis are However, it is anticipated that the existing network difficult to estimate. The objectives of any streamflow will prove inadequate to resolve all these issues and monitoring network are to provide flow and runoff additional targeted stations will be required. information over a sufficient range of catchments with differing input sources for application to water Closely connected with network design and resources design and management, hydropower development is an appreciation of the quality of development, flood risk assessment, and river the streamflow data. Hydrometry or streamflow ecology. Additional objectives to establish such a measurement is a science in its own right, and a network in the HKH are: range of techniques have been developed to provide a continuous measurement of flow (discharge) at • To provide a sufficient duration of record to a gauging site (Herschy 1995). All methods are assess trends and periodicities, especially as these subject to error or uncertainty but these errors can be may be related to global climatic change; and magnified when river conditions are unfavorable, for • To provide sufficient information to assess year- example, as typically apply in the turbulent and mobile by-year and decadal variations and trends in rivers of the HKH, and also when human resources the comparative contribution to streamflow and skills are limited. Data users may well be unaware from glaciers, seasonal snowmelt, and liquid of the many ways they can be wrong. The application precipitation. of sophisticated modeling techniques to poor data can prove misleading or, at worst, worthless. Problems The initial design of a hydrometric (river gauging) of data quality need to be addressed both in the network is to gather a broad impression of the design and management of new stations and in the variability of water resources of a country or region. evaluation of existing flow records. Stations may be added on an ad hoc basis to meet specific needs such as hydropower development or 4.4.1 Quality of Streamflow Measurements water resources planning. With the exception of a few high-level research measurements, streamflow The measurement of discharge in a river is the gauging stations in the HKH have not been installed subject of the science of hydrometry. Discharge, individually or as a network to answer questions unlike most climate measurements, is not simply a associated with climate change or to assess the matter of recording a reading from an instrument. It comparative contributions to flow from melt of involves several stages, all of which pose difficulties glaciers, seasonal snow, and liquid precipitation. or are subject to error. Difficulties and errors are However, stations installed for one purpose are often magnified in the steep mobile channels of the HKH. found to provide useful or essential information Therefore, a cautionary approach is necessary in for another purpose. This has certainly been found using discharge data supplied by a national agency; to be the case with respect to the use of existing the records should not be assumed to be accurate HKH stations to assess questions of climate change and homogeneous. The measurement of liquid impacts or the comparative contribution to runoff flows in open channels is the subject of a series from different sources. of international standards set by the International Organization for Standardization (ISO), and of While serious climate change uncertainties still national standards set by such bodies as the remain, the existing streamflow network in the Bureau of Indian Standards.7 Selected here from HKH, in conjunction with the climate and glacier the full spectrum of topics in these standards are 7 For example, IS 1192–1981: Velocity: Area Methods for Measurement of Flow in Open Channels. 69 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains those particularly related to the HKH, which will be observers can usually be recognized, for example, covered in the remainder of this chapter. by changes that are stepped, but unless checked can lead to serious errors in discharge assessment. 4.4.2 Site Selection A long-established alternative to manual observation Establishing a new measurement site (gauging is the use of a float mechanism with the levels traced station) requires thorough consideration of channel on a chart recorder or, now more frequently, using characteristics, the range of expected flow, and a digital recorder. This is still generally the most winter access. accurate method of measurement of level, now most commonly using an optical shaft encoder and The most important requirement for the location of digital logger. However, as a float recorder needs a gauging station is that downstream features of the still water for measurement, it requires the capital river channel – which control the water level at the cost of a stilling well and housing for the recording station and determine the relationship between water instruments. This arrangement is appropriate for level and discharge – are stable and sensitive to major water resources stations, but may not be changes in discharge. Stability implies that the rating necessary on many rivers. curve for the station should not change over time, and sensitivity implies that a measureable change There are now several alternative types of in discharge should be matched by a measurable instrumentation that can be installed in or over change in level (most relevant to low flows). flowing water and therefore do not need the installation of a stilling well (although they may In the HKH, channel controls are inevitably natural need protection from erosion damage), including features of the channel, such as a rock bar or downward-looking ultrasonic devices and pressure constriction (rather than a gauging weir which is transducers, bubble sensors, and radar level sensors. impractical or uneconomic). Since upland sites often have mobile banks and beds of gravel and Over the last 20 years, the accuracy and reliability boulders, it may be difficult to find a suitable site of pressure transducers has significantly improved with stable control. Suboptimal sites are often and the cost has decreased. Many thousands of chosen, but they require more frequent discharge devices are installed around the world, and they measurements to ensure that the rating curve has are the level measurement method of choice for not drifted. Inevitably, the accuracy of measurement many applications. The term “pressure transducer” at such sites is lower than at locations with stable is applied to devices that convert changes in water control. Accessibility of sites is also a significant pressure and hence water level into electrical limitation. signals, which are then converted to a digital signal of water depth at suitably short intervals (such as 4.4.3 Water Level Measurement hourly). Data can then be stored on a digital logger from which they can be downloaded periodically The traditional approach to water level measurement on site or transmitted by modem to a base station. in the HKH has been manual observation of a A further advantage is a low power requirement: staff gauge at fixed time intervals, and often more typical dry cell battery life is at least several months. frequently during high flow when the level is more Additionally, such transducers will continue to variable. Local observers are recruited to carry out operate when partly covered in sediment. An this task. However, in the absence of supervision, it example of such an installation on a tributary of the has been frequent practice for the observer to visit upper Indus is shown in Figure 4.8. They remain the irregularly and then to fill in the intervening values cheapest and most effective solution for most remote by guesswork. Such interpolation by untrained installations. 70 Mountain Hydrology Figure 4.8 require replaceable gas cylinders but operate by Typical Arrangement for Water Level Measurement compressing gas through a measuring tube into by Pressure Transducer in the HKH the water at each selected interval (typically 15 minutes). They have a measuring range up to 30 m. Both bubble and radar sensors have low power requirements; they do not need mains power, and therefore are suitable for remote stations. It is recommended that stations where stage measurements are currently based on manual observation of a staff gauge be equipped with a pressure transducer, bubble or radar sensor, and associated logger to provide a more continuous and reliable record. They have a very low battery power requirement. An observer should be retained to manually read the staff gauge at less frequent intervals as a check on the calibration of the pressure transducer. Pressure bubble and radar sensors are also suitable for installation at remote stations where no observers are continually present. 4.4.4 Establishing a Relationship between Water Level and Discharge Since river discharge cannot be measured directly and continuously, the usual procedure is to make a continuous measurement of stage, and then convert the stage time series to a discharge time series using a relationship between water level and discharge (the rating curve). Source: D. Archer. Note: The typical arrangement for water level measurement by pressure transducer The rating curve is developed by repeated shows staff gauge, pipe to protect pressure transducer cable, and instrument housing for logger and solar panel. measurement of discharge over the full range of observed stage. Normally, the lower and medium Significant technical developments in radar level ranges present little difficulty, but since the highest sensors and bubble sensors now make these viable flood discharges are infrequent and difficult to alternatives to pressure transducers. Both can be measure, extrapolation is usually necessary to used at open water sites without stilling wells. Radar convert the highest range of stage to discharge. The sensors provide noncontact measurement from extent of extrapolation can place a severe limit on above the water surface (for example, attached to the accuracy of the discharge series. a bridge) by downward transmission and receipt of a reflected radar pulse from the water surface. If the channel and downstream control are stable, The time delay provides an accurate determination comparatively few discharge measurements may be of distance to the water surface. Unlike traditional needed once the rating curve is initially established. bubble recorders, new bubble sensors do not However, in typical HKH rivers, where the bed and 71 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains channel are unstable due to scour or aggradation, deployed internationally. In some developed frequent measurements may be needed through countries, the ADCP has replaced the current meter the life of the station to retain the reliability of as the principal method of flow measurement; its discharge estimation. use is possible at most sites where a current meter can be used. Flow measurement can be made The most frequent method of discharge from a cableway, bridge, or temporary ropeway measurement is using a current meter (cup type, across a river; it thus provides the opportunity for propeller type, or electromagnetic) to measure a wider choice of location for new stations and velocity and depth at a series of segments across for calibrating the upper range of existing stations the channel. At low depths and velocities, a (Figure 4.10). The device uses acoustic fields to measurement may be made by wading (Figure measure water velocities and depths. As the ADCP is 4.9.a), but at greater depths they must be made moved across the channel, mounted on a powered by suspension of the current meter from a fixed or remote-controlled boat or towed on a float, cableway or bridge or from a boat. Fixed cableways it collects a velocity profile and depth across the are an expensive option and are usually limited section. The instrument is connected to a rugged to the most significant stations for water resources bankside laptop with data processing software that (Figure 4.9.b). Hence, many gauging stations are gives the section discharge on completion of the located at or near bridges where bridge suspension traverse. Incorporated instruments and software can be used (even if these are not otherwise the ensure that only the downstream component of flow most suitable sites in terms of channel stability). is measured. In the past, the ADCP was mainly used Boat gauging is usually limited to wide rivers of in larger rivers such as the Amazon, but smaller limited velocity. versions are now available that can be used in river depths of less than 1 m. Velocities up to 10 The acoustic doppler current profiler (ADCP) is m per second can theoretically be accommodated being used increasingly, with many thousands but, in practice, turbulence will limit its use above Figure 4.9 Typical Discharge Measurement Devices in the HKH a. Taking a measurement by wading in a glacier- b. Fixed cableway gauging on the Astore River at fed Indus tributary, in low flow Doyien, in low flow Source: D. Archer. 72 Mountain Hydrology Figure 4.10 Typical Examples of ADCP in Use for Discharge Measurement a. ADCP device is towed by rope across a river b. ADCP mounted on a trimaran designed to cope with fast, choppy water Source: Nick Everard. Source: D. Archer 4 m per second. Turbulence can make it difficult Dilution gauging involves the injection of a tracer to maintain the stability of the towed ADCP , and of known concentration at a constant rate into entrained air can affect the accuracy of velocity a stream and then sampling downstream where measurement. Very high sediment loads can also complete mixing of the tracer with river water has prevent penetration of the signal to the riverbed, but occurred. The discharge is a function of the ratio these can be countered by using a lower operating of the injected tracer to the downstream sampled frequency. concentration. Turbulence here is a positive advantage as it promotes mixing over a short In many situations in the HKH, the ADCP is the only river length. Fluorescent tracers have proven most viable means of discharge measurement, especially effective; preliminary analysis can be carried out in high or flood flows. It is recommended that in the field but are best verified in the laboratory. agencies responsible for streamflow gauging in the Dilution gauging requires no power source and HKH evaluate the ADCP for use as a standard flow can be used to measure discharges up to about measurement device. Capital costs are high, but the 100 m3 per second. This method is recommended to compensations are greater efficiency and flexibility of all monitoring agencies for steep turbulent streams measurement. across the HKH. In the most extreme conditions, typical of glacier 4.4.5 Transforming the Record outflows but also common elsewhere, highly of Stage to Discharge turbulent flow in rocky steams is transmitted in rapids and waterfalls. No in-river sensors are possible. In Once the rating curve has been established, it this situation, dilution gauging methods can be used, seems a trivial matter to apply it to the time series which have already been deployed in mountain river of river level (stage) to create the time series of studies in Nepal for the past 20 years (Spreafico and discharge. Simplicity is certainly the case if the time Grabs 1993). series has been collected digitally and the rating 73 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains curve can be expressed in a functional (rather than shift method is inappropriately used where the graphical) form. The usual functional relationship control is stable, as it incorporates all the errors is logarithmic. However, even here, there are of individual gaugings and does not benefit from possibilities of introducing errors, as shown in the the improvement achieved by fitting a mean line following examples: to the gauging in the rating curve. • It has been past practice, in some cases, where 4.4.6 Evaluating Historical Discharge Records stage data were collected manually to compute the daily mean flow by taking the average While the above discussion provides some guidance daily level and converting this single value to on improvement in future measurement, it is still discharge. This method will underestimate the difficult to determine how to respond to error and daily flow, derived as the average of individual- uncertainty in historical records. It is practically level measurements converted to discharge (since convenient to assume that all records are reliable the relationship is logarithmic rather than linear). as provided and to proceed directly to analysis and The effect will be greatest on small, flashy rivers modeling on that basis. However, results from such affected by intense rainfall. Digital processing of analysis may be unreliable. This is neither to the the level record to discharge eliminates the need advantage of the analyst nor agencies who depend for such simplification; upon the results. Two ways in which agencies can • It has also been past practice to compute a provide information to give a better appreciation of new rating curve on an annual basis. This has the reliability of their data are as follows: some benefits at low flow, if there are many gaugings contributing to each rating curve. • Metadata on each station are held on station However, for basins where the rating is stable, it files. The files contain records made by inspecting can be seriously detrimental to the reliability and officers of the agency on the reliability of the level homogeneity of high flows, because extrapolation measurements made both by manual observation from different ranges of low flows to the highest and by level instruments. In addition, they contain levels may yield quite different discharges for the a listing of all the discharge measurements that same stage. In such cases, it is better to retain all have been made at the station and a graphical the high flow gaugings over the period of stability display of the rating curves. Key metadata and use the same rating throughout the period; information could be displayed on a website for and each station, which can then be accessed by all • Where the riverbed is unstable, as in sand or users, who could then make their own assessment gravel channels, and the rating is constantly or of the potential use of the record. An example frequently shifting, a procedure known as Stout’s of such a website is HiFlows-UK, which provides shift method is appropriately applied to cope flood data for around 1,000 river flow gauging with the constantly changing rating. By the shift stations throughout the United Kingdom. It is method, the distance by which the level for a specifically targeted at flood estimation as a given discharge measurement differs from the part of risk assessment.8 The website requires rating curve (delta h) is determined, and then Δh individual users to make a judgment on the is interpolated between successive gauging and quality of the data on the basis of the information added (or subtracted) from the measured stage provided, and is mainly for use in flood studies; time series before converting to discharge. The and 8 http://www.environment-agency.gov.uk/hiflows/91727.aspx. 74 Mountain Hydrology • An alternative and more comprehensive system is volume are consistent between successive stations provided in Lamb et al. 2003 on the development on a river or between the sum of tributary flow and of an objective system for representing gauging downstream flow. These attributes can be checked station data quality (GSDQ). The system is now for an entire record or as a time series of annual used throughout the Environment Agency of values. England and Wales. The quality information is provided alongside actual river flow time series to 4.5 New Network Requirements users on request. As discussed, significant progress can be made by GSDQ includes statistically based estimates further application of statistical tests and modeling of uncertainty in flow measurement, derived procedures on existing data. An additional benefit from international standards where possible; could be gained from the release of recent climate quantitative attributes, such as the number and and flow data, which could be used in conjunction deviations of check gauging; and categorical with remotely sensed climate and glacier data, for attributes, such as assessment of the significance which coverage is most comprehensive during the of bypassing and the extent of missing or last decade. truncated data. Basic station information (including ratings, flow gaugings, and station The one clear location where new gauging stations dimensions) are entered and stored in the are required is sites recording the outflow from GSDQ software, a customized Excel spreadsheet glaciers. A sufficient number of sites are required application. The GSDQ spreadsheet calculates to assess differences between categories of glacier all the required attribute values from basic by region, size, orientation, and range of elevation. inputs and returns a classification score. This Hydrological understanding would benefit, even if is a number between 0.0 and 1.0, where 1.0 flow measurement is made from glaciers on which indicates best quality. The numerical score is also glacier mass balance studies are not proceeding. subdivided into three classes – caution, fair, and Such glacier outflow records must be combined good. A review of international practices at the with the main river gauging network as a basis for time of Lamb’s publication did not reveal any assessing comparative flow contributions. other established, quantitative, objective measure of data quality in widespread use. 4.6 Summary The application of such a system in HKH would • Streamflow is the only phase of the hydrological provide significant benefit to all data users, including cycle in which the water is confined in such a those in national agencies and consultants involved way as to permit measurement of the overall in water resources development and flood risk catchment response. Therefore, accurate management. measurement of streamflow is critical in linking climatic inputs of moisture and energy with In the absence of information on data quality, the storage and melt of ice and seasonal snow; analyst can usually make some judgment on the • Understanding of streamflow hydrology would quality of the data. A hydrometric network is more particularly benefit from the release of recent than the sum of the individual stations and checks climate and flow data, which could be used can be made on the consistency of a station by in conjunction with remotely sensed climate comparison with neighboring stations in association and glacier data, for which coverage is most with information on precipitation. For example, it comprehensive during the last decade; can be determined whether the runoff and flow • Manual observation of river levels should be 75 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains replaced, or complemented, by automatic Archer, D.R. 2003. “Contrasting Hydrological level measurement at all stations with the use Regimes in the Upper Indus Basin.” Journal of of pressure transducers or radar or bubble Hydrology 274 (1–4): 198–210. sensors with data stored by digital loggers. Data Archer, D.R., and H.J. Fowler. 2008. “Using storage of level data and transfer to the agency Meteorological Data to Forecast Seasonal database enables more rapid and reliable Runoff on the River Jhelum, Pakistan.” Journal of conversion to a discharge time series and Hydrology 361: 10–23. subsequent analysis; • Reliability of stage discharge relationships (rating Berthier, E., Y. Arnaud, R. Kumar, S. Ahmad, P. curves) is currently limited by the frequency and Wagnon, and P . Chevallier. 2007. “Remote difficulty of current meter gauging. Dilution Sensing Estimates of Glacier Mass Balances gauging is recommended for wider use in steep, in the Himachal Pradesh (Western Himalaya, turbulent, glacier-fed streams. The ADCP is India).” Remote Sensing of Environment 108 (3): recommended for testing and application on 327–38. doi: 10.1016/j.rse.2006.11.017. larger streams and rivers; Bookhagen, B., and D.W. Burbank. 2010. “Toward • The principal requirement for new gauging a Complete Himalayan Hydrological Budget: sites is with respect to high-elevation tributaries Spatiotemporal Distribution of Snowmelt and fed primarily by glaciers or permanent snow. A Rainfall and Their Impact on River Discharge.” sufficient number of sites are required to assess Journal of Geophysical Research 115: F03019. differences between categories of glacier by doi: 10.1029/2009JF001426. region, size, orientation, and range of elevation; and Briscoe, J., and U. Qamar. 2007. Pakistan’s Water • Release of streamflow metadata would enable Economy Running Dry. Karachi: World Bank/ users to judge the quality of the data for Oxford University Press. particular applications. Eriksson, M., J. Xu, A.B. Shrestha, R.A. Vaidya, S. Nepal, and K. Sandström. 2009. The Changing References Himalayas: Impact of Climate Change on Water Resources and Livelihoods in the Greater Alford, D. 2011. Hydrology and Glaciers in the Himalayas. Kathmandu: International Center for Upper Indus Basin. World Bank Technical Report, Integrated Mountain Development. South Asia Agriculture and Rural Development Fowler, H.J., and D.R. Archer. 2006. “Conflicting Unit. Washington, DC: World Bank. Signals of Climate Change in the Upper Indus Alford, D., and R. Armstrong. 2010. “The Role Basin.” Journal of Climatology 19: 4276–92. of Glaciers in Streamflow from the Nepal Goudie, A., D. Brunsden, D. Collins, E. Derbyshire, Himalaya.” The Cryosphere Discussions R. Ferguson, Z. Hashmet, D. Jones, F. Perrott, 4: 469–94. www.the-cryosphere-discuss. M. Said, R. Waters, and W. Whalley. 1984. “The net/4/469/2010/. Geomorphology of the Hunza Valley, Karakoram Alford, D., R. Armstrong, and A. Racoviteanu. Mountains, Pakistan.” In The International 2009. Glacier Retreat in the Nepal Himalaya: Karakoram Project, ed. K. Miller. London: An Assessment of the Role of Glaciers in the Cambridge University Press. Hydrologic Regime of the Nepal Himalaya. Haefeli, R. 1962. “The Ablation Gradient and Project Completion Report to World Bank. the Retreat of a Glacier Tongue.” International Association of Hydrological Sciences Publications 58: 49–59. 76 Mountain Hydrology Herschy, R.W. 1995. Streamflow Measurement, 2nd Mool, P.K., D. Wangda, S.R. Bajracharya, K. ed. London: E. & F.N. Spon. Kunzang, D.R. Gurung, and S.P . Joshi. 2001. Inventory of Glaciers, Glacial Lakes and Glacial Hewitt, K. 2005. “The Karakoram Anomaly? Glacier Lake Outburst Floods. Kathmandu, Nepal: Expansion and the ‘Elevation Effect,’ Karakoram ICIMOD. Himalaya.” Mountain Research and Development 25 (4): 332–40. Nepal DHM (Nepal Department of Hydrology and Meteorology). 1988. Hydrological Records of Hewitt, K. 2007. “Tributary Glacial Surges: An Nepal: Streamflow Summary, Updated Version. Exceptional Concentration at Panmah Glacier, Kathmandu: Government of Nepal, Ministry of Karakoram Himalaya.” Journal of Glaciology 53: Water Resources, DHM. 181–88. Oreskes, N., K. Shrader-Frechette, and K. Belitz. Hewitt, K. 2011. “Glacier Change, Concentration, 1994. “Verification, Validation, and Confirmation and Elevation Effects in the Karakoram Himalaya, of Numerical Models in the Earth Sciences.” Upper Indus Basin.” Mountain Research and Science 263: (5147): 641–46. Development 31 (3): 188–200. http://dx.doi. org/10.1659/MRD-JOURNAL-D-11-00020.1. Rees, H.G., and D.N. Collins. 2006. “Regional Differences in Response of Flow in Glacier- Immerzeel, W.W., L.P.H. van Beek, and M.F.P. Fed Himalayan Rivers to Climatic Warming.” Bierkens. 2010. “Climate Change Will Affect Hydrological Processes 20 (10): 2157–69. the Asian Water Towers.” Science 328 (5984): 1382–85. Sharif, M., D.R. Archer, H.J. Fowler, and N. Forsythe. 2012. “Trends in Timing and Magnitude of Flow IPCC (Intergovernmental Panel on Climate in the Upper Indus Basin.” Hydrology and Earth Change). 2007. Contribution of Working Systems Sciences Discussions 9 (9): 9931–66. Group II to the Fourth Assessment Report of the doi: 10.5194/hessd-9-9931-2012. www.hydrol- Intergovernmental Panel on Climate Change, earth-syst-sci-discuss.net/9/9931/2012. 2007 – Working Group II: Impacts, Adaptation and Vulnerability. Geneva: IPCC. Sharma, C. 1983. Water and Energy Resources of the Himalayan Block. Kathmandu: Navana Lamb, R., M.D. Zaidman, D.R. Archer, T.J. Marsh, Printing Works. and M.L. Lees. 2003. River Gauging Station Data Quality Classification (GSDQ). R&D Technical Shrivastava, S.K. 2011. Evaluation of Temporal And Report W6-058/TR. Swindon, United Kingdom: Spatial Climatic Variability over Indian Himalaya. UK Environment Agency. Doctoral thesis. Roorkee: Indian Institute of Technology. Majeed, A. 1995. Forecasting of Seasonal and 10-Day Inflows into Mangla Reservoir. Report for Singh, P., and N. Kumar. 1996. “Determination the Pakistan Ministry of Science and Technology, of Snowmelt Factor in the Himalayan Region.” Karachi. Hydrological Sciences Journal 41: 301–10. Majeed, A., T. Hussain, and M.R. Asghar. 2010 Singh, P., and N. Kumar. 1997a. “Effect of “Long and Medium Range Forecast for Inflow at Orography on Precipitation in the Western Tarbela.” Pakistan Journal of Meteorology 3 (6): Himalayan Region.” Journal of Hydrology 199: 37–52. 183–206. Matsuo, K., and K. Heki. 2010. “Time-Variable Singh, P., and N. Kumar. 1997b. “Impact of Climate Ice Loss in Asian High Mountains from Satellite Change on the Hydrological Response of a Snow Gravimetry.” Earth and Planetary Science Letters and Glacier Melt Runoff Dominated Himalayan 290: 30–36. doi: 10.1016/j.epsl.2009.11.053. River.” Journal of Hydrology 193: 316–50. 77 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Singh, P., K.S. Ramasastri, and N. Kumar. 1995. Van Dijk, A.I.J.M. 2011. Model-Data Fusion: “Topographical Influence on Precipitation Using Observations to Understand and Reduce Distribution in Different Ranges of Western Uncertainty in Hydrological Models. Paper Himalayas.” Nordic Hydrology 26: 259–84. presented at 19th International Congress on Modelling and Simulation, Perth, Australia, Singh, P., K.S. Ramasastri, N. Kumar, and M. December 12–16, 2011. http://mssanz.org.au/ Arora. 2000. “Correlations between Discharge modsim2011. and Meteorological Parameters and Runoff Forecasting from a Highly Glacierized Himalayan Young, G.J., and K. Hewitt. 1993. Basin.” Hydrological Sciences Journal 45: “Glaciohydrological Features of the Karakoram 637–52. Himalaya: Measurement Possibilities and Constraints.” In Snow and Glacier Hydrology: Spreafico, M., and W.E. Grabs. 1993. Proceedings of the International Symposium, “Determination of Discharge with Fluorescence Kathmandu, Nepal, November 16–21, 1992, Tracers in the Nepal Himalayas.” In Snow ed. G.J. Young, 273–83. IAHS Publication 218. and Glacier Hydrology: Proceedings of the Wallingford, United Kingdom: International International Symposium, Kathmandu, Nepal, Association of Hydrological Sciences. November 16–21, 1992, ed. G.J. Young, 17–27. IAHS Publication 218. Wallingford, United Kingdom: International Association of Hydrological Sciences. 78 Indigenous Glacier Monitoring 5. Indigenous Glacier Monitoring When the early Western mountain climbers are talking about what is happening in their approached Mount Everest, the Sherpa living at its area instead of reading out a weather report base could not understand why the foreigners would from Kathmandu that might have no relevance want to climb the mountain. Today, with training to them. and economic incentives, the Sherpa are among the leading high-altitude climbers internationally, with The remote sensing and the satellites give dozens of ascents of Mount Everest and other high us the eagle-eye view, which is essential Himalaya peaks made possible each year by their but not enough. In a country as diverse expertise. Drilling in an ablation stake or maintaining geographically and socially as Nepal – there a climatological station on or near a Himalaya are more than 90 languages and 103 caste glacier is no more technical than climbing an 8,000 m and ethnic groups – the eagle-eye view needs Himalayan peak. The ability to move safely and to be complemented by the view from the function efficiently in the often-hostile environment at ground, what I call “toad’s-eye” science … the high altitudes of potential monitoring sites in the High science should … meet up with civic HKH Mountains is presumed to be a characteristic science and traditional knowledge, in order of the indigenous mountain peoples throughout to understand what is happening, so that the region. To successfully monitor the climate and national governments can also plan. … glaciers of the HKH, the mountain people must be involved. For example, Dipak Gyawali, Former The solutions have to come out of the Minister of Hydrology and Research Director of the watershed and out of the problem-shed. You Nepal Water Conservation Foundation, illustrates a can talk about big solutions – building high possible role for local people. He explains that an dams – which can take 40 years. We don’t area as diverse as the Himalaya needs localized, know in Nepal if a government will last 40 “toad’s-eye” science if it is to learn how to adapt to days. The solutions have to be what these climate change (Thompson et al. 2007): millions of households can take. Can they be helped? How can they be helped? We just We have some suggestions for how to do it. haven’t done the science for that. We need For instance, you put a weather monitoring civic science; ground-level truth. station in every school in Nepal, and get the children to do the readings and get Over the past decade, increasingly with international the schoolmaster to fax the readings back, news about climate change impacts on indigenous your data points increase from around 450 communities, the movement has broadened to to around 4,000. You are suddenly rich in include special considerations for social science data, and the local people are involved in data and an indigenous research paradigm (Pulsifer understanding the dimensions of the problem. et al. 2011). This chapter explores the potential for engaging mountain communities in the HKH It will be a long, drawn out process, but it is in a research agenda that would involve them starting with rain gauges in the schools, linked in monitoring their environmental conditions, up with the local FM radio stations. Suddenly specifically glaciers and glacier-related conditions. It the FM stations are very excited because they demonstrates how community-based and collectively 79 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains held knowledge can provide valuable insights, ever, is there a reciprocal aspect in which indigenous complement scientific efforts, and support adaptive knowledge is formally structured at the outset into the strategies. Recently, concerns about adaptation content of scientific research questions and methods. to climate change have moved to center stage. A Effective, positive working relationships between “climate change adaptation has been enshrined in mountain communities and research institutions the policy debate through its appearance in Article take time and commitment to develop and require 2 of the United Nations Framework Convention a substantial level of institutional investment and on Climate Change (UNFCCC), where the support. These factors need to be considered as much ultimate objective of the Convention concedes as possible when it comes to weaving traditional that adaptation to climate change in relation to ecological knowledge or local participation into the food production, ecosystem health and economic monitoring and data collection structures that support development can and will occur.” Adger et al. the aims of glacier hazard mitigation and support (2009b, p. 336): resilience under conditions of climate change. Residents of the HKH region as well as elsewhere Traditional ecological knowledge and indigenous in the Global South are predominantly poor and knowledge are, for the most part, synonymous underrepresented in the political and scientific and refer to the evolving knowledge acquired by arena, and thus lack the capacity to influence indigenous peoples over hundreds, in some cases, hazard planning and adaptation efforts taking place thousands of years through direct contact and at the national and international levels (CCCD interaction with the natural world. This knowledge is 2009; Tschakert and Dietrich 2010). Mountain place specific, reflecting long-standing relationships communities, with limited assets, information, and between natural phenomena, biota, landscapes, support, have little choice but to attempt to adapt aquatic systems, and the timing of events. Cultural to changing environmental conditions, which can adaptations to the risks and hazards posed by create a troubling dilemma for the scientific and life in the extreme mountains of the Hindu Kush, humanitarian community. A report by the Global Karakoram, and western Himalaya are well known Leadership for Climate Action (GLCA 2009a, p. – for example, building practices and architectural 6) states: “The world’s poor, who have contributed styles in the region reflect indigenous knowledge the least to greenhouse gas emissions, will suffer about the geophysical environment and approaches the worst impacts of climate change and have to coping with a range of mountain hazards. The the least capacity to adapt.” These vulnerabilities traditional timber-laced construction pattern proved underscore the need for proactive engagement highly resistant to the earth movements caused by with the populations and social groups that will the Kashmir earthquake. likely continue to face acute social stress and environmental risks. As a number of researchers have pointed out, the linkages between traditional ecological knowledge 5.1 Indigenous Monitoring: and understanding climate change is essential for Overview and Purpose verifying and evaluating climate change scenarios developed by scientists. For others, the traditional Historically, models of cooperation between the ecological knowledge of indigenous peoples is now scientific community and indigenous communities seen as a repository for solutions to the vagaries were structured for scientists to assist communities and uncertainties associated with climate change. with research projects that were shaped by the As Raygorodetsky (2011) emphasizes, community- scientists’ research plans. The models for interaction based, collectively held knowledge – along with have largely been unequal and one-sided. Rarely, if local processes of observing and interpreting 80 Indigenous Glacier Monitoring changes in land, sky, and sea – is essential to educational, and technical instruments for dealing complement scientific data with chronological and with changing environmental circumstances. landscape-specific precision. Data construction and management should In light of the “tipping points” being approached be designed to serve the expressed needs of in several agro-ecological zones, United Nations communities. It is recommended that an exploratory agencies, including the United Nations Educational, study be conducted on the potential of an interactive Scientific and Cultural Organization (UNESCO) global positioning system (GPS) to be of local and the Food and Agriculture Organization (FAO) interest and of practical use in mountain conditions. of the United Nations, have been exploring the Another strategy would be to assess whether collective knowledge of indigenous peoples. A the technical hardware and software is robust tipping point is described by Lenton et al. (2008) or adaptable enough to support potential uses as the critical threshold at which a tiny perturbation without an unacceptably high probability of failure. can qualitatively and quantitatively alter the state of Data gathered could be hydrometeorological and development of a system, often with unanticipated, glaciological impact assessments of water access, rapid, and abrupt changes. Interest in traditional agricultural production, and food security, enabling ecological knowledge has grown in the past the generation of enhanced profiles of vulnerability decade in part because it has become increasingly and adaptation across the region. Potential recruits recognized that such knowledge can help the global for data gathering and monitoring could include community avert tipping points, and can also inform high-elevation hunters and pastoralists, given their the conservation of biodiversity, protection of rare knowledge of local geography and their direct and endangered species (Colding and Folke 2001), experience with the physical dimensions of mountain management of protected areas, development environments. Indigenous monitoring should of ecological processes (Alcorn 1993), and enhance communication about the process of proliferation of sustainable resource use in general. land use planning and classifying areas in terms of hazard, risk, and vulnerability. Involving indigenous people in observing and monitoring glaciers and the glacial environment has been done for the following reasons: (a) to 5.2 Vulnerability of Mountain build collaboration between indigenous knowledge, Communities: Some Considerations science, and technology; (b) to develop tools for active engagement of mountain communities in The HKH region is similar to other ecologically glacial science and community-based monitoring; and socially vulnerable areas that are already (c) to develop approaches to engage indigenous experiencing the impacts of climate change (Barnett, communities in mapping glacier hazards and Adam, and Lettenmaier 2005; Battisti and Naylor glacier hazard management; (d) to expand the 2009; GERES 2009; Orlove, Wiegandt, and use and understanding of glacier science and Luckman 2008). Temperature and precipitation technologies in everyday life in ways that resonate are exhibiting increased variability and more with indigenous systems and technologies; and (e) to pronounced and frequent extremes than historical preserve data created from indigenous knowledge norms (Bahadur 2004; Mool et al. 2001). It is as a lasting legacy and a benefit for current and well established that mountain environments are future generations. Ultimately, a major goal should susceptible to climate-induced environmental be to serve not only the scientific community but change (Salick, Zhendong, and Byg 2009). These also the vast range of glacier-dependent mountain changes have affected the productivity and reliability communities. In this regard, data gathering and of local agricultural practices, threatening existing documentation could serve as new cultural, household livelihood strategies and increasing 81 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains vulnerability to water-related hazards among disappear. Other documented local issues include already vulnerable, resource-dependent populations declining water availability, changing patterns in (Erikkson, Fang, and Dekens 2008; Halvorson precipitation, demise of springs, increasing extreme and Hamilton 2010; Ives 2004; Jodha 2005; weather events, and shifts in the length of the Marston 2008; Norberg-Hodge 1992; Siebert and growing season, adding uncertainty to an already Wangchuk 2011; Xu et al. 2007; Xu et al. 2008; Xu challenging situation where new economic growth et al. 2009). in various sectors, including mining, hydropower development, industrial agriculture, and tourism, has Along with these changes, glacier-related risks led to increasing and competing demands for water. and hazards in the HKH are threatening mountain communities. Human vulnerability to glacier A major shortcoming in the work to date is that the recession, particularly in this region, has recently potential of communities and households to adapt come into the focus of investigation and analysis to glacier-related climate change has not been among research institutions and the scientific effectively addressed, nor have there been studies community. The Intergovernmental Panel on Climate of how households are already shifting resources Change (IPCC 2007, p. 15) defines “vulnerability” and labor in response to climate changes. Future as the “degree to which a system is susceptible scientific efforts need this information so they might to, and unable to cope with, adverse effects of support the community’s particular adaptation climate change.” Recent work on vulnerability to strategies and capacity-building activities, for climate change focuses on “exposure, sensitivity, example, by forming indigenous monitoring networks and resilience of human systems to various forms in the region. of climate-related risk,” and IPCC argues that the strengthening of adaptive capacity and livelihood The high-mountain regions of the HKH lie within security must be a central concern of development an active glacier hazard zone. The impacts of efforts. Nevertheless, in the HKH region, consistent glacial lake outburst floods (GLOFs) in this region empirical datasets drawn from grounded and have been significant and deadly. The catastrophic detailed analyses of the consequences of glacial impacts of the 1994 Thortomi GLOF in Bhutan, recession and climatic variability for mountain where the moraine-dammed lake failed upon communities are lacking. As a result, there is no heavy rainfall, underscored the country’s need coherent picture of the complex socioecological for preparedness for such potential disasters. The processes that influence vulnerability to glacial growing number of dammed lakes in the upper recession at the local level across the region, and reaches of basins, often behind unstable moraines, the ways in which particular social groups – the makes it a critical time for populations increasingly poor, ethnic minorities, women, children, and the exposed to GLOFs to initiate glacier and climate elderly – that are at risk are regularly overlooked. monitoring programs. However, the glacier monitoring program in Bhutan is in its infancy. 5.2.1 Glacial Recession Recently, attention in Bhutan has focused on a GLOF prevention program that involves a United Nations- What has been reported is that glacial recession sponsored engineering project to partially drain has increased rapidly and is currently affecting the one of the glacial lakes in the upper reaches of the rates of seasonal glacial discharge and the array Lunana valley. of water resources upon which communities rely. Moreover, local perceptions of glacier recession The changes and social upheaval associated in the region confirm the scientific observations with glacier hazards complicate the mountain and underscore the worrisome trends as glaciers development controversies that countries in the HKH 82 Indigenous Glacier Monitoring already face. The development trajectory of this 21.7 years in Pakistan; 25.1 years in India; mountainous region has deepened its involvement 21.4 years in Nepal; and 24.6 in Bhutan.9 The in geopolitics and the global economy. The relative youth of the population has influenced the socioeconomic and political conditions associated level of preparedness, planning, response, and with these trends, as well as local and regional recovery capacity of these countries, because a large struggles for access to and control over natural portion of the population lacks the experience and resources, are exacerbated in the aftermath of skills to deal with such disasters. Those who have natural disasters. Furthermore, the highly gendered been through disasters can help quell fears and social structure and social differentiation drawn encourage hope among victims of glacier-related along class, religious, and ethnic lines greatly hazards, as they see that people have survived influence disaster experiences, resource access, these events, and they can provide guidance on employment opportunities, and coping abilities. how to stay alive or, at the very least, what strategies worked in the past to help mitigate In theory, the population’s vulnerability must be environmental risks and damage. Such experiences considered in relation to the physical environment, and memories are fundamental to the development, social structures, axes of social difference, and the maintenance, and transmission of disaster recovery local–global interactions of pressure-applying forces. knowledge. Typically, this type of information has Both internal and external factors would influence been held and shared by elders. However, the the community members chosen to participate in push for Western-oriented modes of education and the monitoring of glaciers and glacial environments population shifts suggests a potential loss of interest in this region. Internal factors such as the status of in indigenous knowledge. family members, class, gender, ideology, ethnicity, poverty, education, access to communication 5.2.3 Gaps in Knowledge and Awareness of networks, and the seasonal or permanent out- Mountain Hazards migration of community members shape the social fabric and resiliency in coping with glacier hazards. Residents of the HKH region tend to lack scientific Other contributing factors include externally driven knowledge about the geophysical processes that changes such as large infrastructure projects (for cause mountain hazards and awareness of the steps example, dams and roads); expansion of industrial individuals can take to protect themselves from agriculture; forest clearing and devegetation their impacts. Because they lack sound information, through commercial or illegal logging, mining, indigenous people may blame these types of events and mineral extraction; and acts of aggression, on metaphysical phenomena or social groups war, and displacement. Commerce, construction, rather than focusing attention on preparedness and urbanization, and economic development in planning. Outside specialists may be perceived the HKH have continued with little regard for as having undue power and influence, thus their environmental risks and hazards. Several of these information dissemination techniques are generally factors are discussed in the next sections. not well received in the HKH countries, whose populations have strongly hierarchical social and 5.2.2 Demographics educational structures. The power relationships and informational disconnect between government The demographic picture of the region, especially officials and local residents, and between scientists the young age of the population, has contributed to and the public, result in a large gap between risk in several ways. The HKH countries in this study community members’ knowledge and vital life- have low median ages: 16.6 years in Afghanistan; saving information. 9 United Nations Statistics Division: Social Indicators. http://unstats.un.org/unsd/demographic/products/socind/default.htm. 83 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 5.2.4 Male Out-migration research in the areas of hydroclimatic measurements and glacier monitoring could be enriched with The increase in male out-migration from mountain spatially distributed social analyses. Coupling communities is another cause of the information the physical science with the social science is void. Men who have left rural areas indefinitely to critical to policy planning and the development of pursue opportunities in urban centers may learn new methodological tools for glacier change research. critical information but fail to pass it on to family back home. This leaves some of the most vulnerable Practical implications of this glacier hazard social groups – rural women and children in remote management discussion continue to emerge. First, mountain communities – without information about advancing glacier science education and mountain disaster preparedness, medical care, emergency hazard education – with a focus on indigenous services, assistance, and so forth. Women and girls knowledge and developing a culture of hazard risk find themselves burdened with greater responsibilities reduction – is critical to reducing loss of life and in caring for family members and doing farm property due to glacier-related disasters in the future. work, thereby constraining the time and resources Mitigation and preparedness methods drawing from available for school and attending critical disaster traditional knowledge have been suggested for preparedness classes. use in disaster preparedness, the development of early warning systems, and postdisaster rebuilding. 5.3 Glacier Hazard Management Second, these educational initiatives will create Issues knowledge that will help communities prepare and plan adaptation measures in response to Against this complicated regional background, the environmental and climate change challenges. aim of the current study is to examine potential And third, disaster risk reduction plans should bridges between indigenous monitoring and maintain a combination of community involvement scientific glacier monitoring in the HKH. The need and expert input by teaming elders, educators, is great for applying new frameworks and scientific and leaders representing both men and women. tools for describing and analyzing glacier hazard Education about the physical process and existing vulnerability (IPCC 2007) and glacier hazard geophysical hazards, and training in evacuation and management more generally in the HKH context. first aid, should be fundamental to all glacier hazard Glacier hazard management includes a constellation management efforts in the region. of measures that reduce the risk of glacier-related disasters, including prediction of glacier hazards, 5.4 Solutions for Indigenous Glacier drainage or containment of glacial lakes to remove Monitoring in the HKH Region the possibility of a GLOF, minimization of the portion of the population exposed to glacier hazards, and This review of scientific and social science reduction of human vulnerability. information about indigenous monitoring of glaciers in the Himalaya revealed, first and foremost, In view of recent changes in HKH drainage systems, a lack of data and information on indigenous special attention should be given to the social monitoring. This lack hampers attempts to project conditions that influence vulnerability to assist in the likelihood of success in taking action to involve preparing for and mitigating impacts. No adequate mountain communities proactively in monitoring documentation exists of the magnitude of extreme and glacier hazard management. A glacier weather events and measurements of the impact monitoring station was established in 2011 by the on livelihoods, which is a major shortcoming for Pakistan Meteorological Department in the upper modeling the impacts of climate change. Scientific Hunza valley at Pasu Glacier (elevation 4,500 m). 84 Indigenous Glacier Monitoring Additional monitoring stations will be extended to audio, maps, or even video, to capture individual other glaciated valleys in the Hunza basin. or collective knowledge. The terms of data-sharing agreements would need to be carefully worked As with glacier monitoring, there is a need to begin through to adequately recognize intellectual property quantifying and qualifying community interactions and privacy. with glaciers and impacts of glacial recession on water security, hazards, food security, and well-being. Developments that could have dramatic implications Current knowledge comprises mainly anecdotal for hydrological systems are fraught with critical observations (often conveyed in news media) and unknowns. A major unknown in many river basins is pieced-together stories of societal change in relation the downstream–upstream relationships associated to changing environmental circumstances. There are with hydropower development trajectories. Also very only a small number of individual research projects much unknown are the cumulative environmental focused on climate change impacts. and socioeconomic impacts of the hydropower development agenda on freshwater systems. Efforts From a monitoring perspective, these observations to promote renewable energy sources that involve point to two distinct concerns. The first relates to major investments in hydropower can come at large how the information and data generated can be costs and risks to local populations. shared with mountain communities and the country government to support adaptation to climate 5.5 Observations and change; the second involves questions about what the data on glacial hydrology actually mean to the Recommendations local community water user groups or stakeholders based in the drainage basins. Each mountain In order to reduce the vulnerability to glacier-related community is dealing with different sets of problems disasters in the HKH, the need for integrating efforts to and geographies of glacier hazards. Solutions include indigenous communities seems apparent. The must be site specific, if they are to be effective. As findings from this review underscore several challenges suggested by Gearheard et al. (2011) in the context for integrating mountain communities in an indigenous of Arctic environmental monitoring, the development monitoring effort. Scientific research centers, political of locally appropriate technology for monitoring institutions, and disaster management agencies need in high-mountain settings will require an iterative to be convinced that local participation in monitoring design process and engineering refinements. These environmental processes is worthwhile. Governments authors suggest that design of the system should and international agencies, by recognizing these needs consider the inclusion of community representatives and abilities of local citizen scientists, stand to gain to use GPS and field computers to document key allies in their efforts to prepare for, mitigate, and environmental observations, track their travel respond to glacier hazards. Teaming community routes, log their observations and experiences, and elders with disaster preparedness specialists would log the weather they encounter. The system would further create a holistic and trusting venue for be easily updatable with timely observations. As conveying data and information. Gearheard and colleagues noted, it is possible from an engineering and technology design perspective 5.5.1 Observations to produce an indigenous software interface as a result of interaction and exchange between local The following are observations regarding indigenous participants and Western scientists. monitoring concluded from the current study: Data collections from indigenous knowledge projects • The countries in the HKH region exhibit weak generally involve an array of media, including text, outreach and communication linkages with 85 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains mountain communities and have a high risk of Development of indigenous monitoring glacier-related hazards and disasters; teams, as in “citizen scientist” programs • The picture of emergency preparedness and disaster response planning at the local and • Educate community leaders in core concepts and drainage basin scale is weak, with attention to methods to raise awareness and community support; glacier-related concerns being, in most cases, • Train teams in basic methods and methodological limited or nonexistent; and approaches, record keeping, core skills, and • Countries in the region have weak scientific assessment of household and community capacity at the local level. perceptions of problems; and • Project goals should include glacial hazard 5.5.2 Recommendations investigations, glacial monitoring, weather monitoring, informing engineering projects The authors of the present study recommend that that are scientifically sound and innovative in a set of best practices in indigenous monitoring approach, and disaster risk reduction programs. of glaciers and glacial hazards for major basins be developed in collaboration with the Support and enhance hazard preparedness scientific community, with minimum standards and disaster risk reduction at local level and indicators that include the social aspects of indigenous monitoring. The intent of these • The preparedness plan should include a disaster recommendations is to ensure that the specific management plan that lays out the line of needs and contributions of mountain communities communication and coordination among the are institutionalized into glacier-related monitoring scientific networks; and risk reduction plans: • The emergency preparedness plan should include an assessment format and prioritization of existing General community interventions resources and capacities; • The mechanism for early warning and reporting • Establish regional glacier science outreach and a potential disaster should be strengthened training centers; in coordination with government agencies, • Organize curricular initiatives across grades levels nongovernmental organizations, and civil society; and in higher education that provide science- and based climate- and glacier-related education; • Indigenous knowledge and an intergenerational • Reduce information vulnerability among mountain perspective should be considered in designing people. A starting point would be promoting monitoring systems. Elderly people may have literacy, developing appropriate curriculum memories of past disasters or environmental materials, and training teachers. Training stresses that could pass life-saving information on should address life-saving earthquake drills, to younger generations. basic weather and climate change science, risk awareness, and information about hydroclimatic One important strategy would be to build capacity systems and glacier studies; within regional higher education institutions. • Respect mountain communities as able Colleges and universities are key players in stakeholders and disaster management understanding glacier science and monitoring professionals; and activities. These institutions can both help build • Conduct glacial recession-related vulnerability capacity at the local level and provide leadership and needs assessments. and technical support to mountain communities. 86 Indigenous Glacier Monitoring References the Western Himalayan Region.” In Proceedings of the GERES International Seminar on Energy Adger, W.N., S. Huq, K. Brown, D. Conway, and M. and Climate Change in Cold Regions, Hulme. 2003. “Adaptation to Climate Change in April 21–24, 2009. the Developing World.” Progress in Development GLCA (Global Leadership for Climate Action). –95. Studies 3 (3): 179­ 2009a. Facilitating an International Agreement on Alcorn, J.B. 1993. “Indigenous Peoples and Climate Change: Adaptation to Climate Change. Conservation.” Conservation Biology 7: 424–26. Washington, DC: United Nations Foundation, GLCA Bahadur, J. 2004. Himalayan Snow and Glaciers: Associated Environmental Problems, Progress, and GLCA (Global Leadership for Climate Action). Prospects. New Delhi: Concept Publishing. 2009b. Toward a Post-2012 Agreement on Climate Change: Recommendations of Global Barnett, T.P., J.C. Adam, and D.P. Lettenmaier. 2005. “Potential Impacts of a Warming Climate on Leadership for Climate Action. http://www. Water Availability in Snow-Dominated Regions.” globalactionnow.org/images/pdf/glca_recomm_ Nature 438: 303–9. post2012_agreement_climatechange.pdf. Battisti, D., and R. Naylor. 2009. “Historical Warnings Halvorson, S.J., and J.P. Hamilton. 2010. “In the of Future Food Insecurity with Unprecedented Aftermath of the 2005 Qa’yamat: The Kashmir Seasonal Heat.” Science 323: 240–44. Earthquake Disaster in Northern Pakistan.” Disasters: The Journal of Disaster Studies, Policy, CCCD (Commission on Climate Change and and Management 34 (1): 184–204. Development). 2009. Closing the Gaps: Disaster Risk Reduction and Adaptation to Climate Change IPCC (Intergovernmental Panel on Climate in Developing Countries. Stockholm, Sweden: Change). 2007. Climate Change 2007: Impacts, CCCD. Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Colding, J., and C. Folke. 2001. “Social Taboos: ‘Invisible’ Systems of Local Resource Management Report of the Intergovernmental Panel on Climate and Biological Conservation.” Ecological Change (M.L. Parry, O.F. Canziani, J.P. Palutikof, Applications 11: 584-600. P.J. van der Linden, and C.E. Hanson, eds.) Cambridge: Cambridge University Press. Erikkson, M., J. Fang, and J. Dekens. 2008. “How Does Climate Change Affect Human Health in Ives, J.D. 2004. Himalayan Perceptions: the Hindu Kush-Himalaya Region?” Regional Environmental Change and the Well-Being of Health Forum, International Centre for Integrated Mountain Peoples. London: Routledge. Mountain Development (ICIMOD) 12 (1): 11–15. Jodha, N. 2005. “Adaptation Strategies against Gearheard, S., C. Aporta, G. Aipellee, and K. Growing Environmental and Social Vulnerabilities O’Keefe. 2011. “The Igliniit Project: Inuit Hunters in Mountain Areas.” Himalayan Journal of Document Life on the Trail to Map and Monitor Sciences 3 (5): 33–42. Arctic Change.” The Canadian Geographer 55 Lenton, T., H. Held, E. Kreigler, J.W. Hall, W. Lucht, (1): 42–55. S. Rahmstorf, and H.J. Schellnhuber. 2008. GERES (Groupe Energies Renouvelables, “Tipping Elements in the Earth’s Climate System.” Environnement et Solidarités). 2009. “Impacts of Proceedings of the National Academy of Sciences Climate Change in Ladakh, Lahaul and Spiti of 105 (6): 1786–93. 87 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Marston, R. 2008. “Land, Life, and Environmental Siebert, S., and S. Wangchuk. 2011. Government Change in Mountains.” Annals of the Association Policies, Market Opportunities and Climate of American Geographers 98 (3): 507–20. Change: Agricultural Transformation in Bumthang, Mool, P., D. Wangda, S. Bajracharya, K. Kunzang, Bhutan. Paper presented at joint AFHVS, ASFS, & D.R. Gurung, and S.P . Joshi. 2001. Inventory of SAFN Conference, June 9–12, 201l, Missoula, Glaciers, Glacial Lakes and Glacial Lake Outburst MT. Floods, Bhutan. Kathmandu, Nepal: International Thompson, M., M. Warburon, T. Hatley, and D. Centre for Integrated Mountain Development. Gyawali. 2007, Uncertainty on a Himalayan Norberg-Hodge, H. 1992. Ancient Futures: Learning Scale: An Institutional Theory of Environmental from Ladakh. San Francisco, CA: Sierra Club Perception and a Strategic Framework for the Books. Sustainable Development of the Himalaya. Lalitpur, Nepal: Himal Books. Orlove, B., E. Wiegandt, and B. Luckman, eds. 2008. Darkening Peaks: Glacier Retreat, Science, Tschakert, P., and K. Dietrich. 2010. “Anticipatory and Society. Berkeley, California: University of Learning for Climate Change Adaptation and California Press. Resilience.” Ecology and Society 15 (2): 11. Pulsifer, P., G.L. Liadler, D.R. Fraser Taylor, and Xu, J.C., R. Grumbine, A. Shrestha, M. Eriksson, A. Hayes. 2011. “Towards an Indigenist Data X. Yang, Y. Wang, and A. Wilkes. 2009. “The Management Program: Reflections on Experiences Melting Himalayas: Cascading Effects of Climate Developing an Atlas of Sea Ice Knowledge and Change on Water, Biodiversity, and Livelihoods.” Use.” The Canadian Geographer/Le Geographe Conservation Biology 23 (3): 520–30. Canadien 55 (1): 108–24. Xu, J.C., R. Sharma, J. Fang, and Y. Xu. 2008. Raygorodetsky, G. 2011. Why Traditional Knowledge “Critical Linkages between Land-Use Transition Holds the Key to Climate Change. United Nations and Human Health in the Himalayan Region.” University. http://unu.edu/articles/global- Environment International 34 (2): 239–47. change-sustainable-development/why-traditional- Xu, J.C., A.B. Shrestha, R.A. Vaidya, M. Eriksson, knowledge-holds-the-key-to-climate-change. and K. Hewitt. 2007. The Melting Himalayas: Salick, J., F. Zhendong, and A. Byg. 2009. “Eastern Regional Challenges and Local Impacts of Climate Himalayan Alpine Plant Ecology, Tibetan Change on Mountain Ecosystems and Livelihoods. Ethnobotany, and Climate Change.” Global Kathmandu, Nepal: International Centre for Environmental Change 19: 147–55. Integrated Mountain Development (ICIMOD). 88 Satellite Imagery and Digital Elevation Models 6. Satellite Imagery and Digital Elevation Models Glaciers have been monitored on a large scale since • Availability of semiautomated mapping 1894, when the International Glacier Commission approaches that are cost-effective, because often was established at the sixth International Geological numerous satellite scenes can be analyzed within Congress in Zurich. Today, several global, regional, a relatively short time or even simultaneously. and national glacier inventorying and monitoring initiatives exist, two examples of which are described However, as Raup, Kääb et al. (2007) have noted, in this chapter. While earlier inventories used mere availability of satellite imagery and DEMs topographic maps and field observations for glacier does not necessarily equate to accurate thematic monitoring, today indirect measurements from information production and glacier parameter satellite imaging and DEMs are increasingly being estimation. Remote sensing has some disadvantages: used, although direct field measurements are still the necessity for adequately trained operators; important. Ideally, direct and indirect measurements analytical problems in glacier identification and go hand in hand, though often circumstances such mapping; comparability of results from various as remoteness, inaccessibility due to political and methodological approaches; interference from security issues, or time and funding limitations for atmospheric conditions (for example clouds, smoke); field monitoring campaigns only allow for indirect and sensor malfunctions. measurements. However, in too many glacier mapping and monitoring studies, insufficient 6.1 Literature Review fieldwork has been done to collect ground-truth data to determine error calculations for results from The most recent and comprehensive review of the remote sensing analyses. use of satellite imagery and DEMs in glacier mapping and monitoring is Kargel et al. 2013. This book Remote sensing technologies are used to identify and summarizes results from over 10 years of research delineate glaciers, characterize glaciogeomorphic on glacier monitoring within the international project parameters, describe glacier fluctuations and Global Land Ice Measurements from Space (GLIMS); velocities, and document (through multitemporal the study described in this chapter relies heavily on analysis) changes in glacial extent and volume, this source, particularly on the following chapters: among other tasks. Remote sensing technologies and data collection offer the following advantages: • Bishop, Bush et al. 2013 on remote sensing science and technology for glacier assessment; • Spatial coverage of data of larger regions; • Ramachandran et al. 2013 on satellite image • Availability of data for remote areas with limited acquisition, preprocessing, and special products; field-based glaciological measurements; • Kääb et al. 2013 on glacier mapping and • Availability of data of the same area from time monitoring based on spectral data; slices that often reach back some decades, • Quincey et al. 2013 on digital terrain modeling allowing multitemporal analyses; and glacier topographic characterization; • Availability of data at various spatial and spectral • Racoviteanu et al. 2013 on Himalayan glaciers in resolutions; India, Bhutan, and Nepal; and • Availability of data at low cost or even no cost; • Bishop, Shroder et al. 2013 on Afghanistan and and Pakistan. 89 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Other important reviews of the use and usefulness of via a web-based portal that is easy to use; remote sensing in glacier monitoring are Raup, Kääb • Reach out to and educate the general community et al. 2007, with a focus on the GLIMS project; and about glacier monitoring activities and results; Racoviteanu, Williams, and Barry 2008, with a focus • Address the needs of local communities (for on the Himalaya. The USGS as part of its Satellite example, the implementation of a GLOF warning Image Atlas of Glaciers of the World (Williams system); and and Ferrigno 2010) includes chapters on Pakistan • Identify questions for the applied research (Shroder and Bishop 2010), India (Vohra 2010), and community. Nepal (Higuchi et al. 2010). 6.3 Mesoscale Satellite Imagery 6.2 Requirements for Glacier Monitoring Program Until the early 1970s, aerial photography was the primary remote sensing technology in glacier With regard to the specific environmental, cultural, mapping and monitoring. Although it offers logistic, and security circumstances in the HKH many advantages, this technology also has many region, a glacier monitoring program that makes use restrictions, for example, in ground coverage, its of remote sensing data should meeting the following limited availability for study areas such as the HKH, conditions: and the high cost of aircraft and regular flights. • Cover the entire region; Satellite imagery analysis was then introduced in • Respect existing security issues and be safe for all studies of the cryosphere. Since the early 1970s, participants; medium-resolution (10–90 m) optical satellite data • Be carried out by regional and local institutions, have become available, particularly with the launch agencies or others, and be advised by an of sensors such as the Landsat Multispectral Scanner international expert board; (MSS), Landsat Thematic Mapper (TM), Système Pour • Train local and regional geospatial information l’Observation de la Terre (SPOT), Indian Remote technology and field specialists in glacier Sensing (IRS) including Cartosat and Resourcesat, monitoring; Landsat 7 Enhanced Thematic Mapper Plus (ETM+), • Employ low-cost or open-source, widely used Advanced Spaceborne Thermal Emission and data types and software packages; Reflection Radiometer (ASTER), and Advanced Land • Develop an inventory of all glaciated areas using Observing Satellite (ALOS) (Figure 6.1). Today, one single approach (same imagery and DEM large-scale (less than 10 m) imagery suitable for types, same analytical methods); detailed glacier studies at basin scale is available, • Monitor as many glaciers as possible and for example, from IKONOS, Quickbird, and as often as possible (in a near real-time GeoEye-1. However, the narrow swath, long revisit environment) to describe the general status of cycles, and high cost limit their use for systematic glaciers and glacial changes throughout the glacier monitoring of larger regions such as the HKH; HKH. CORONA data from 1960 to 1972 were • Monitor a number of benchmark glaciers in more declassified in 1995 but are only available for some detail and intensity; glacierized areas within the HKH region. • Implement one regional glacier monitoring entity, or at least guarantee the exchange of data and The Landsat program is the longest-running results between participating countries; enterprise for acquisition of satellite imagery of • Make status, data, results, reports, and so on, Earth. Since its beginning in 1972, seven individual available to the general community (open access) satellites have been launched, of which Landsat 90 Satellite Imagery and Digital Elevation Models 5 (1984) and Landsat 7 (1999) are still in orbit. Williams, and Barry 2008). ASTER’s features include Landsat images cover an area of 185 km2 and the following: come at 15–60 m spatial resolution, depending on the wavelength range. Both Landsat 5 and 7 carry • Spatial resolution of 15 m in visible (VIS) and the TM and MSS; the MSS on board Landsat 5 was near infrared (VNIR) is adequate for regional- powered down in 1995. On May 31, 2003, the scale glacier studies; Scan Line Corrector in the Landsat ETM+ instrument • High spectral resolution with three VNIR bands, failed, causing the loss of approximately 22 percent six mid-infrared (MIR) bands, and five thermal of the data in a scene. However, data products are infrared (TIR) bands allows for multispectral available with the missing data optionally filled in image classification; using other Landsat 7 data selected by the user. • Off-nadir viewing band in the near infrared (NIR) Scenes can be downloaded free of charge from the enables high-resolution along-track stereoscopic USGS imagery archive – and requests placed for vision; processing of scenes not downloadable – or from • Adjustable sensor gain settings provide increased the Global Land Cover Facility. Landsat imagery has contrast over bright areas (snow and glaciers); been extensively and successfully used in glacier • The revisit interval is only 16 days; monitoring studies, particularly for band ratio • Data are provided at no cost for noncommercial analysis and land cover classification. As of April users participating in the GLIMS project; 2012, approximately 3,400 Landsat 5 TM scenes • High-priority data acquisition requests submitted and 9,700 Landsat 7 ETM+ scenes displaying the by the researcher to the ASTER Science Team HKH study area were available for downloading ensure adequate quality of the acquired data for from the USGS Earth Explorer website.10 glacier monitoring; for example, data acquisition requests include specifications on instrument gain The many advantages of remote sensing by ASTER, settings for each ASTER band, the acquisition launched in 1999, have increased its use in window (start and end time for the acquisition), numerous glacier monitoring studies (Racoviteanu, and specific glaciers to be targeted in the field; and Figure 6.1 • Various products are available from the Land Landsat ETM+ Index Map for the HKH Region Processes Distributed Active Archive Center,11 for example, surface kinetic temperature, surface emissivity, surface reflectance, and the orthorectified product package that includes the relative DEM constructed on demand using Silcast software. As of April 2012, about 8,000 ASTER scenes were available for download from NASA’s Earth Observing System Data and Information System Reverb|ECHO website, providing coverage over the HKH region.12 Figure 6.2 is an example of this imagery, showing glaciers of Bhutan and China Source: Bajracharya and Shrestha 2011. obtained with ASTER imagery. 10 USGS Earth Explorer: http://earthexplorer.usgs.gov/. 11 Land Processes Distributed Active Archive Center website: https://lpdaac.usgs.gov/. 12 Reverb|ECHO website: http://reverb.echo.nasa.gov/. 91 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains ALOS, launched in 2006, combines visible remote Figure 6.2 sensing with active microwave techniques using ASTER Image of Glaciers in the Himalaya of Bhutan three sensors: the Panchromatic Remote-Sensing and China Instrument for Stereo Mapping, suitable for detailed digital elevation mapping at 5 m accuracy; the Advanced Visible and Near Infrared Radiometer type 2 (AVNIR-2), suitable for glacier mapping and measuring of albedo of the glacier surface; and the Phased Array type L-band Synthetic Aperture Radar, suitable for day-and-night and all-weather land observation, and probably for estimating snow cover depth, which may assist in determining accumulation rates on glaciers. An example of ALOS imagery showing the glaciers of Sagarmatha Source: J. Kargel. National Park in eastern Nepal is shown in Figure 6.3. Data can be requested through ALOS data Figure 6.3 nodes for noncommercial use at costs incurred by ALOS AVNIR-2 Scene Covering Sagarmatha the participating ALOS data node organizations National Park, Nepal by region. 6.4 Glacier Monitoring Using Satellite Imagery and DEMs Although today semiautomated glacier mapping from optical satellite imagery is a standard tool in glacier monitoring, problems still exist in the methodology that cause different results in different glacier mapping studies. For example, Kamp, Krumwiede et al. (2013) compared studies of glaciers in the Mongolian Altai Mountains and found that the reported glacier number and glacial extent varied widely: while one study put the number of “glaciers” in the Mongolian Altai at 120, another one identified around 731 “glaciers and glacierized areas;” as a result, the exact number and spatial coverage of Mongolia’s glaciers are still unknown. Kamp, Krumwiede et al. (2013) identified two main reasons for the varying results: the definition of a Source: PASCO and JAXA. glacier and methods for data collection. First, for many studies, it was unclear what (2009) mapped only glaciers larger than definition of “glacier” was used and what exactly 0.05 km2. Bolch et al. (2010) mapped glaciers had been mapped. When inventorying glaciers in larger than 0.01 km2, but then compared only western Canada, Bolch, Menounos, and Wheate glaciers larger than 0.1 km2. Also, Paul et al. 92 Satellite Imagery and Digital Elevation Models (2009) set the lower limit of what constitutes a electromagnetic spectrum, which makes it glacier at 0.01 km2, because a glacier any smaller relatively easy to map debris-free glaciers would be too difficult to accurately identify through (Figure 6.4). Snow and ice are characterized by a platform with a spatial resolution of 15–30 m. high reflectivity (albedo) in the VIS wavelengths GLIMS defines a glacier, identified by a single (0.4–0.7 micrometers); medium reflectivity in the GLIMS glacier ID, as “a body of ice and snow that NIR (0.8–2.5 micrometers); low reflectivity and high is observed at the end of the melt season” (Raup, emissivity in the TIR (2.5–14 micrometers); and low Kääb et al. 2007). absorption and high scattering in the microwave. While in clear weather the high albedo of snow and The second reason results vary is that many studies ice make them easily distinguished from do not – or do so only in a sketchy way – explain surrounding terrain using visible infrared (VIR), employed methods; inaccuracies and errors; source optically thick clouds are also highly reflective in data type (topographic maps, aerial photographs, VIR, hence they confound the classification. or satellite imagery); or source data acquisition However, they are reflective in the NIR, and are date. This lack of information calls into question a thus discriminated from snow and ice. study’s reliability. Kamp, Krumwiede et al. (2013), therefore, concluded that it is important to realize Single band ratios (VIS/NIR) and NDSI are that reported numbers of glaciers, glacierized area, commonly used to separate the bright snow and glacial changes, and estimated climate changes ice from darker landscape features (Figure 6.5). from existing studies have to be viewed with When applying band ratios, a threshold of 2 was extreme caution. found to be most suitable (Bishop et al. 2008; Bolch et al. 2010); when using NDSI, a threshold The two indicators used most frequently for a of 0.4 was found to differentiate snow from glacier’s response to climate forcing are changes nonsnow, and thresholds of 0.5–0.6 proved in glacial area and terminus position, which are successful in delineating glacier ice in the Andes of relatively easy to extract from multispectral satellite Peru (Racoviteanu, Williams, and Barry 2008). Both images (Racoviteanu, Williams, and Barry 2008). band ratios and NDSI methods produced Glacial area is calculated from glacier outlines and satisfactory mapping results for shaded glacier used as input for volume–area scaling techniques. parts, and have the advantage of being fast and Glacier outlines combined with a DEM help to robust, and thus relatively easy to automate over derive glacier parameters such as hypsometry and extensive areas (Bolch and Kamp 2006; Paul, the ELA. Kääb, and Haeberli 2007; Bishop et al. 2008; Racoviteanu, Williams, and Barry 2008). However, Raup, Kääb et al. (2007) identified the five most using band ratios has some problems in mapping commonly used methods in glacier identification glaciers due to the presence of fresh snow on the from satellite imagery: manual digitization, spectral glacier surface, supraglacial debris, and proglacial band ratio and threshold, normalized difference and supraglacial lakes (Racoviteanu, Williams, and snow index (NDSI), geomorphometric based, and Barry 2008). thermal band methods. Despite present difficulties, Krumwiede et al. 6.5 Monitoring Debris-free Glaciers (2013) and Kamp, McManigal et al. (2013) argued for a simple threshold ratio mapping approach, Clean glacial ice has a distinct spectral signature, allowing for faster processing time by using an with uniqueness in the VNIR part of the unsupervised classification scheme. This scheme 93 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Figure 6.4 Delineation Results for Glaciers in the Northern Tien Shan a. Delineated glaciated areas (red lines) using b. Comparison with glacier outlines (blue lines) in Landsat TM4/TM5 ratio images the topographic map 1:10,000 Source: Bolch and Kamp 2006. Figure 6.5 Simple Threshold Ratio Mapping Approach Using Landsat 7 Bands 4 and 7 for Parts of the Himalaya in India (33°N 77°E) a. Extraction of snow and ice b. Extracted snow and ice data draped over natural color composite Source: Krumwiede et al. 2013. 94 Satellite Imagery and Digital Elevation Models uses bands 4 and 7 from the Landsat 5 TM and also showed, for example, that the number was Landsat 7 ETM+ sensors and performs simple 59 glaciers using band ratio TM5/TM3, 61 raster mathematical calculations. This approach glaciers using NDSI, and 62 glaciers using band can be incorporated directly into satellite imagery ratio TM5/TM4. The “overprediction” of glacier processing methods and can generate output numbers for the latter approaches occurred datasets, including raster overlays and glacier in particular in the glacier class of less than outline shapefiles with area calculations. These 0.05 km2 and was due to erroneous identification output datasets can then be incorporated with of shadows as glaciers. Pan et al. (2013) also DEMs to determine glacier hypsometric areas compared glacier mapping results using thresholds and glacier area with respect to aspect. Using this of 2, 3, and 4, and concluded that a threshold of simple method makes it is easier to analyze several 2 produced the most accurate results. However, images in a shorter amount of time and allow for the mapping using band ratio TM4/TM7 also still multitemporal change detection and analysis. has problems when differentiating between clean ice and debris-covered ice. Note that TM4/TM7 These findings of Krumwiede et al. (2013) and is the only band ratio that did not have errors Kamp, McManigal et al. (2013) were supported by and produced the best results when mapping Pan et al. (2012, who found that band ratio disconnected glaciers. TM4/TM7 produced the most accurate results when compared to results from visual interpretation 6.6 Monitoring Debris-covered and from other frequently used band ratios Glaciers (TM3/TM5, TM5/TM3, TM5/TM4) and NDSI, because it is the only approach that did not Debris-covered glaciers are more difficult to map, erroneously map shadowed terrain and proglacial because the supraglacial material might have lakes (Figure 6.6). The Pan et al. (2013) study very similar VIS/NIR spectral signatures to the in the Ikh Turgen Range in the northeastern surrounding terrain (Figures 6.7 and 6.8), although Mongolian Altai at the Russian border, using band spectral information has been successfully used to ratio TM4/TM7, identified 52 glaciers. The results detect characteristics of supraglacial debris load and rock types (Figure 6.9) (Bishop, Shroder, and Figure 6.6 Ward 1995). This complicates not only the (semi) Glacier Mapping Results Using Different Band Ratios Applied to a Landsat Image of Ikh Turgen Figure 6.7 Range Characteristics of Supraglacial Debris of Glaciers in Northern Pakistan Derived from SPOT Imagery Multispectral Analysis a. Debris load at Raikot Glacier; b. Surface lithology at Batura Glacier Source: Pan et al. 2012 Source: Bishop, Shroder, and Ward 1995. 95 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Figure 6.8 Figure 6.9 Results from Different Glacier Mapping Steps for Results from Morphometric Glacier Mapping the Nun Kun Mountains in Zanskar (MGM) of Glaciers in Himalaya Range of Zanskar, India, Using ASTER Satellite Imagery and ASTER DEMs 1. Parkachik Glacier; 2. Drang Drung Glacier Source: Kamp, Byrne, and Bolch 2011. Note: (a) profile curvature; (b) slope angle; (c) cluster; (d) thermal information; Source: Kamp, Byrne, and Bolch 2011. (e) supervised land cover classification using the maximum likelihood method; (f) Note: In a. and b. figure 2. MGM-based delineation (yellow line) picks up clean morphometric glacier mapping. PG = Parkachik Glacier; UG = unnamed glacier; ice as well as debris-covered parts of the glacier. SV = Suru valley; UV = unnamed valley. automated identification, but also, often, the manual Another promising approach is mapping from identification of the glacier margins, because they thermal imagery, because supraglacial debris appear contiguous with the proglacial zone. As the that is thinner than 2 cm generally has a lower spectral information alone is insufficient in mapping temperature than the surrounding moraines debris-covered glaciers, DEMs are used to support and terrain. Above this threshold, the thicker the mapping process. For example, small elevation debris actually insulates the underlying ice and differences between the proglacial zone and glacier the supraglacial debris cover might now show surface or the exact location of a catchment divide a temperature similar to the surrounding of the can be more easily identified in a DEM than with glacier, which prevents any glacier monitoring. spectral information alone (Quincey et al. 2013). As in most other cases, information about the Bolch et al. (2010) found that SRTM3 data did thickness of the supraglacial debris is not available; provide sufficient results to calculate ice divides, thermal information analysis is applicable only to which are important for defining individual glaciers. glacier parts under thin debris cover, while any This finding will help automate the process of other parts of the glacier that are covered with delineating glaciers. The combination of band ratios thicker debris might not be identified correctly. and topographic information has successfully been Therefore, the use of thermal information in glacier used to map debris-covered glaciers. monitoring can only have a supportive role and 96 Satellite Imagery and Digital Elevation Models is usually performed in combination with other project, in which it was able to guide the ASTER mapping techniques. instrument to acquire imagery of Earth’s glaciers that was optimal (best season and instrument gain More sophisticated glacier mapping approaches settings) for glacier monitoring.13 Regional centers “use both first- and second-order topographic and affiliated stewards are responsible for a specific derivatives to segment landscape units accordingly, region; for the HKH region these are the ICIMOD, and make use of statistical, artificial intelligence as the regional center for the Himalaya (Bhutan, and hierarchical structural methods” (Quincey et India, Nepal); and the University of Nebraska, al. 2013). Examples include pattern recognition, Omaha, as the regional center for Southwest Asia artificial intelligence techniques, and object-oriented (Afghanistan and Pakistan). mapping. Object-oriented mapping “employs various DEM derived terrain-object properties Of importance for the glacier monitoring approach such as slope angle, slope azimuth, curvature, in the HKH region proposed here is that for and relief to identify locally contiguous portions registered GLIMS-related researchers, ASTER data of the landscape, which are iteratively aggregated are free of charge. Details about GLIMS can be to form higher-order landform objects at larger found in Bishop et al. 2004; Kargel et al. 2005; and larger scales” (Quincey et al. 2013). Even Kargel et al. 2013; Rau et al. 2005; Raup, Kääb et more complex, integrative mapping approaches al. 2007; and Raup, Racoviteanu, et al. 2007. combine sophisticated techniques, for example, a hybrid (anthropogenic–computational) method GLIMS analysis results include digital glacier that includes object-oriented mapping and artificial outlines and related metadata, and they can also neural networks (Raup, Kääb et al. 2007); or include snowlines, center flow lines, hypsometry the MGM method, which combines band ratios, data, surface velocity fields, and literature topographic analyses, cluster analysis, supervised references. The program also develops tools land cover classification, and thermal information to aid in glacier mapping and for transfer of (Paul, Huggel, and Kääb 2004; Bolch and Kamp analysis results for archiving to the National Snow 2006; Bolch et al. 2007; Kamp, Byrne, and Bolch and Ice Data Center. These include GLIMSView, 2011). The latter has been found to be useful documented procedures for GLIMS analysis, and in distinguishing debris cover on glaciers, and web-based tools for data formatting and quality both Landsat 7 ETM+ (at 60 m resolution) and control. More than 60 institutions across the globe ASTER (at 90 m resolution) include thermal bands are involved in GLIMS. (Racoviteanu, Williams, and Barry 2008). The GLIMS webpage14 has guides and tutorials, 6.7 Global Land Ice Measurements for example, for glacier classification, compilation from Space of glacier inventory data from digital sources, analysis algorithms, and specific guides from and The international project GLIMS, established for regional centers. It offers the following ways of in the late 1990s, was designed to monitor the viewing the database: world’s glaciers using data from optical satellite instruments such as ASTER. Compared to other • GLIMS Glacier Viewer. The GLIMS Glacier glacier inventories, GLIMS is the first attempt to Viewer is an interactive map of the data in build a globally complete, high-resolution map of the GLIMS glacier database. Different layers glacier extents. It began as an ASTER Science Team in this interface can be viewed and spatially 13 GLIMS: Global Land Ice Measurements from Space – Monitoring the World’s Changing Glaciers. http://glims.org. 14 Ibid. 97 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains queried, including GLIMS glacier outlines, ASTER However, while some regions are covered completely footprints, regional center locations, the World (for example, British Columbia in Canada, the Glacier Inventory, and the fluctuations of glaciers. Caucasus, China, and – of importance for this report The GLIMS glacier outlines layer contains the – Nepal, although based on somewhat old maps), results of glacier mapping within the GLIMS the datasets for India and Pakistan are incomplete, initiative. Each polygon in this layer represents the with many glaciers still missing. For example, for extent of a particular glacier at a specific time, Southwest Asia (Afghanistan and Pakistan), the as well as other possible features of the glacier, number was 411 glaciers covering around 710 km2, such as the extent of debris cover or the location and for the Himalaya (Bhutan, India, Nepal) the of supraglacial and proglacial lakes. The GLIMS number was 3,707 glaciers covering around glacier outlines can also be downloaded as ESRI 6,860 km2 (GLIMS database). shapefiles, MapInfo tables, Geographic Markup Language files, Keyhole Mark-up Language 6.8 Macroscale Satellite Imagery (Google Earth), and the Generic Mapping Tools multisegment format (Figure 6.10); Ground-based measurements, whether from • GLIMS text search interface. This interface manually operated or automatic weather stations, provides access to the GLIMS glacier database sample the climate at a single point. Assessment through a text-based search form. The of regional or catchment climate depends on parameters the user can search on, as well interpolation and (often dubious) extrapolation. as the result fields that can be returned, are Satellite remote sensing provides the potential for customizable. This allows the user to search on obtaining complete areal coverage of aspects of and return only the criteria that are relevant to the climate of a region, covering locations that are their needs. Query results in this interface can be not measured or cannot be measured at ground downloaded individually or as part of the larger level. Spatial data derived by remote sensing has a result set. Downloaded data are available in the same formats as from the map server interface; and Figure 6.10 • Open Geospatial Consortium server. Viewing GLIMS ASTER Browse Data within Another way of accessing data within the GLIMS Google Earth glacier database is through the Web Map Service and Web Feature Service protocols of the Open Geospatial Consortium. These services allow access to GLIMS glacier data directly from desktop GIS software products such as ArcGIS, GRASS, or Google Earth, as well as other applications such as MapServer. The data are divided into annual subgroups for the years 2000 to the present. Subsequently, each annual group is divided into three layers representing image center points, image bounding boxes (polygons), and browse image overlays, as shown in Figure 6.10. As of March 21, 2012, the GLIMS glacier database Source: GLIMS website http://glims.org. Note: View the GLIMS ASTER browse data within Google Earth by downloading the included 115,051 glacier analyses (snapshots). file GLIMS_ASTER.kmz and opening it in Google Earth. 98 Satellite Imagery and Digital Elevation Models theoretical basis in identified relationships between free. Given the cloud climatology of the upper Indus, the intensity of emission of electromagnetic radiation most satellite scenes are partially cloud covered. In at specific wavelengths (spectral bands) and specific general, pixels over valley bottoms will be cloud-free surface or atmospheric properties. The following more often than pixels covering mountaintops. Thus, sections describe the most important spatial data the count of observations contributing to eight-day sources for application to the climate of the HKH, aggregates will most often be greater for low- what they are designed to sense, and what are their elevation zones than higher ones. uses and limitations. The principal use of MODIS-derived land surface 6.8.1 Moderate Resolution Imaging temperature has been to derive lapse rates of Spectrometer temperature over the full altitudinal range of a catchment, rather than over the limited range NASA’s MODIS instruments capture data in 36 available to surface weather stations. Derived spectral bands ranging in wavelength from 0.4 to temperatures are then used in conjunction with a 14.4 micrometers and at varying spatial resolutions. melt model, distributed by altitude. They are designed to provide measurements of large-scale global dynamics, including changes in MODIS, like other radiometrically derived data Earth’s cloud cover, radiation budget, and processes products, has limitations in its capacity to accurately occurring in the oceans, on land, and in the lower assess parameters of interest. It detects the presence atmosphere. NASA software extracts time series or absence of snow but cannot measure either datasets with given resolution and time averaging falling precipitation or snow water equivalent on from specific sensed wavelengths. Records are the ground. The theoretical basis for assessing available from early 2000 to the present. snow-covered area and land surface temperature depends, among other things, on relationships With respect to the climate of the HKH, MODIS identified between the intensity of emission of provides the potential to develop a spatial electromagnetic radiation at specific wavelengths characterization of the snow-covered area and (spectral bands) and specific surface or atmospheric land surface temperature, which can serve as properties. Potential exists for misidentification analogs, respectively, for precipitation and air of radiometrically similar climate features. This temperature. Land surface temperature and snow- is illustrated by the challenge of accurately covered area datasets are available as daily or differentiating snow versus cloud (both are cold and eight-day aggregates. The eight-day snow-covered “bright”). Crucial algorithms for determination of area datasets provide 500 m horizontal resolution snow cover and land surface temperature routinely maximum snow cover extent (product MOD10A2) employ cloud-masking approaches. (Hall et al. 2001). For land surface temperature, the eight-day datasets provide 1 km horizontal resolution In other mountainous study areas, the systematic mean surface temperature for daytime (near misidentification of snow as cloud in the transition local noon) and nighttime (near local midnight) zone between snow-covered and snow-free overpasses (product MOD11A2) (Wan 2008). areas has been found to occur in earlier versions of the MODIS snow algorithm. The ground- The algorithms for both snow-covered area and based data available in the HKH generally do land surface temperature depend upon “clear sky not allow assessment of whether such issues conditions.” This means that, for a given pixel, in have been resolved in the current version of the order to determine the snow cover state or calculate algorithm. Independent of cloud-masking issues, its land surface temperature, the sky must be cloud- the topography of the study area for mountainous 99 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains regions may also affect accuracy of snow cover frequency observation, although with substantial detection. In winter, large shadows resulting from gaps in the spatiotemporal coverage of the higher- low sun angles can also result in underdetection of resolution data needed to calibrate the lower- snow in steep terrain. resolution continuous imagery from the geostationary platforms. The TRMM 3B43v006 data product 6.8.2 Tropical Rainfall Monitoring Mission has provided a continuous time series of monthly estimated precipitation totals at 0.25 decimal degree The Tropical Rainfall Monitoring Mission (TRMM) horizontal resolution from January 1998. satellite carries multiple sensors that serve the following functions: The TRMM 3B42 data, while having a high temporal resolution, contain several artifacts and generally • Precipitation rainfall radar, designed to provide result in unreliable measurements in mountain three-dimensional maps of storm structure ranges. However, attempts have been made to giving information on intensity and distribution correct TRMM 3B42. Forsythe et al. ( 2011) had of rainfall, rain type, storm depth, and height at much stronger reservations about the accuracy of which the snow melts into rain. It has a horizontal TRMM 3B42 rainfall for modeling flow in the upper resolution at the ground of 5 km and a swath Indus basin. They found, for example, that the TRMM width of about 250 km; catchment average accumulated precipitation over • TRMM microwave imager (TMI), which is a a period of years is only a fraction (40–60 percent) passive microwave sensor designed to provide of the observed river discharge. However, this quantitative rainfall information over a wide applied to seasonal snowmelt driven subcatchments swath by quantifying water vapor, cloud (with little or no glacial contribution), where an water, and rainfall intensity in the atmosphere. explanation based on drawdown of glacial volume TMI supplements the similar special sensor cannot be applied, and, therefore, the reliability of microwave/imager (SSM/I), which has been quantitative assessment of precipitation by TRMM operating since 1987; 3B42 is questioned. They also found that TRMM • Visible and infrared scanner, which mainly uses precipitation estimates did not correspond well the association between cloud top temperature with ground-based data in terms of seasonality or and height and the occurrence of precipitation. orographic gradient. Higher cloud tops are positively correlated with precipitation for convective clouds; and TRMM way well provide reliable quantitative • Lightning imaging sensor detects and locates estimates of summer monsoon convective rainfall, lightning over the tropical region of the globe. but application to orographically enhanced winter snowfall from westerly systems may prove more The TRMM high-resolution observations are very problematic (Dinku et al. 2008). Comparisons limited in observational frequency, with direct repeat to available local long-record observations of observations only about twice per month. The TRMM precipitation and river discharge data in the upper specific observations, however, are merged with Indus suggest that the TRMM estimates provide a additional passive microwave observations from quantitative index of monthly precipitation rather several other satellite-borne instruments (SSM/I, than a measure of absolute magnitude. In this they DMSP , AMSR-E, and AMSU-B, among others), as are similar to the local long-record meteorological well as the near-continuous low-resolution infrared observations, which also do not directly represent and thermal imagery from geostationary weather catchmentwide precipitation but do correlate well satellites. This multisensor composite can thus provide as indicators of mass inputs for seasonal snowmelt- a balance between good spatial resolution and high driven catchments. 100 Satellite Imagery and Digital Elevation Models 6.8.3 Advanced Very High Resolution (overall planimetric and vertical accuracy), data Radiometer management, interpretation, and application (Quincey et al. 2013). Accurate glacier assessment The Advanced Very High Resolution Radiometer using topographic information is frequently an (AVHRR) is a radiation detection imager that can be issue of DEM quality and the methodology used in used for remotely determining cloud cover, surface DEM generation. For example, a major difficulty temperature, and snow cover extent. The first AVHRR in using digital photogrammetry to generate was a four-channel radiometer that was launched elevation data over glacier surfaces is that accurate in 1978; the second (AVHRR/2), a five-channel DEMs can only be derived if the glacier surface instrument, was launched in June 1981. The latest shows sufficient texture to correlate image pairs; instrument – version AVHRR/3, with six channels – otherwise, grossly inaccurate elevation estimates was launched in May 1998. Although with more develop over large expanses of bare ice or snow limited spectral resolution, this long record offers (Quincey et al. 2013). Without knowledge of the the potential to greatly increase the overlap of the algorithms used to generate the DEM, an operator spatial data products with local observations, thus may accept such results as being reliable. Great refining quantification of relationships between them, care is required to ensure that the data selected and to extend the range of observations by MODIS for an application are appropriate, processing is to better capture present spatiotemporal climate carried out with a high level of expertise, and errors variability. in any derived data are accurately reported, so that real geophysical patterns and features can be 6.9 DEMs and Geomorphometry differentiated from image and processing artifacts (Quincey et al. 2013). DEMs are digital representations of the Earth’s surface. In glacier monitoring, they are required for 6.9.1 Source Data image orthorectification and radiometric calibration, debris-covered glacier mapping, surface energy Usually DEMs are generated from stereo airborne balance studies, glacier ice volume loss and mass photographs or satellite images. While the former balance estimates, glacier hypsometry, and ELA are not available for the entire HKH, the latter estimation. DEMs are generated from digitized do cover the entire region. Such satellite imagery topographic maps, satellite stereo-imagery (for is collected by different sensors (for example, example, ASTER, IRS, SPOT), and data derived from ASTER, Cartosat, SPOT) and, hence, is of different radar interferometry (for example, SRTM, TerraSAR-X) parameterization and quality. Most of the DEMs used and laser altimetry (for example, light detection and for glacier monitoring represent elevation using a ranging (LiDAR)). Unfortunately, no perfect DEM regular grid of constant spatial resolution. Gridded type for glacier mapping exists, because the different DEMs derived from satellite imagery are available DEMs all have their advantages and disadvantages. at various scales, which affects their suitability for Hence, the user must carefully select the DEM type glacier monitoring. For example, medium-resolution with regard to the application. (10–30 m) satellite sensors such as ASTER and European Remote Sensing (ERS) -1/2 cover a larger DEM generation and analysis require a high area but provide less topographic detail, while fine- degree of operator expertise. Digital terrain resolution (less than 10 m) satellites such as SPOT-5, modeling is a complex process involving acquisition Quickbird, or GeoEye offer much more detail of of source data, interpolation techniques, and a smaller area. Quincey et al. (2013) describe the surface modeling. In addition, quality control advantages and disadvantages of square-gridded is necessary, including accuracy assessment DEMs as follows: 101 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains They are directly comparable with remote selection or mosaicking. Its relatively high accuracy sensing imagery of equal spatial resolution, of 7–14 m is a result of extensive postprocessing simple to analyze statistically, computationally that includes cloud masking to remove cloudy pixels, easy to represent and can be saved in a stacking all cloud-screened DEMs, removing residual range of formats (for example, GeoTIFF, bad values and outliers, averaging selected data to HDF). Conversely, they are poor at create final pixel values, and then correcting residual representing abrupt changes in elevation and anomalies. While version 1, released in 2009, was heterogeneous topography, particularly at an experimental stage and had residual anomalies coarse resolutions, and can generate large and artifacts that affected the accuracy of the file sizes at fine resolution, as well as having a product and hindered effective utilization for certain large amount of data redundancy across flat applications, these errors have been eliminated in areas. version 2. ASTER DEMs of medium (30 m) spatial resolution DEMs are also generated from synthetic aperture can be relatively easily generated using various radar (SAR) satellite images from, for example, software packages or purchased at low costs from ERS-1/2, Radarsat-1/2, SRTM, TanDEM-X, and Land Processes Distributed Active Archive Center. TerraSAR-X, and from LiDAR datasets from airborne Numerous studies have shown that ASTER DEMs sensors. However, most of these data have are suitable in glacier studies, and the GLIMS significant disadvantages that limit their suitability for project employs ASTER DEMs as the main elevation glacier monitoring; for example, ERS satellites have source. However, the production of “absolute” DEMs large footprints in the range of kilometers, which referenced to mean sea level requires ground control is of no use for mapping relatively small Alpine points (GCPs) that have been collected in the field or glaciers. from other sources, such as orthorectified images or topographic maps. SRTM provides near-global elevation data at 90 m spatial resolution and 10 m accuracy In addition, other DEM sources come with some (Figure 6.11). However, despite some advantages, disadvantages. For example, while SPOT-derived the slope-induced errors characteristic of InSAR DEMs produced good results in glacier studies, the data make SRTM unsuitable for glacier change high cost of the imagery limits its use over large detection at small timescales and over small areas, and the generation of CORONA-derived glaciers (Racoviteanu, Williams, and Barry 2008). DEMs is relatively difficult due to complicated image Furthermore, for the mid-latitudes and the outer geometry and flight parameters, especially in rugged tropics, SRTM’s acquisition month overlapped with terrain (Racoviteanu, Williams, and Barry 2008). the accumulation season, resulting in possible overestimations of SRTM-derived elevations over A good solution is the ASTER GDEM, the only DEM glaciers (Racoviteanu, Williams, and Barry 2008). that covers the entire land surface of the Earth at a Bishop et al. (2008) noticed that SRTM data high resolution of 30 m. It was produced using the were greatly affected by shadows caused by high entire 1.5-million-scene ASTER archive acquired topographic relief. Despite these disadvantages, from the start of observation in 2000 in cooperation SRTM data-derived elevations are often used to with the Japan-United States ASTER Science Team fill gaps in DEMs generated from other sources, and it was released free of charge. ASTER GDEM for example, ASTER. In such cases, postprocessing was developed based on a grid of 1x1 degree is necessary to smooth the transition between the of latitude and longitude and requires no scene original DEM and SRTM elevations. Furthermore, 102 Satellite Imagery and Digital Elevation Models these merged DEMs must be used with caution, Figure 6.11 because the scale of topographic detail varies across SRTM Index Map for the HKH Region the DEM. TanDEM-X provides a global DEM to HRTI-3 specifications, that is, 12 m spatial resolution, less than 2 m relative vertical accuracy, and less than 10 m absolute vertical accuracy; studies on the use of TanDEM-X data in glacier monitoring have yet to be carried out. LiDAR DEMs are of high quality, allowing for detailed glacier analyses; however, procuring these DEMs can be expensive. ICESat’s Geoscience Laser Altimeter System, a space-based LiDAR, has a smaller footprint of only 60 m but has Source: Bajracharya and Shrestha 2011. been found useful only for evaluating other DEMs or to estimate glacier elevation changes when compared with other multitemporal elevation data the Khumbu Himalaya and found a mean difference (Quincey et al. 2013). of 0.43±16.7 m. Fujita et al. (2008) compared SRTM DEM elevations and ASTER DEM elevations Comparative studies of DEMs from different sources with ground survey data in the Bhutan Himalaya and demonstrate that some uncertainty still exists about reported a mean elevation difference of 11.3 m for the reliability of extracted elevations. For the Tien the SRTM DEMs and 11 m for the ASTER DEMs. Shan at the border between Kazakhstan and Kyrgyzstan, Bolch, Kamp, and Olsenholler (2005) The GTOP30 DEM was developed from several compared elevations from (a) the SRTM DEM raster and vector sources by the USGS and and a DEM derived from contour lines (reference completed in 1996. While it offers global coverage, DEM); and (b) the SRTM DEM with an ASTER the horizontal resolution of approximately 1 km DEM, concluding that, for the former, the average does not allow for any detailed glacier mapping, difference between elevations was only about 6 m although it provides a useful overview of the study on average, while for the latter the difference was region and has been used to fill data gaps in other up to 100 m, particularly at southeast and north- DEM datasets. to-northwest exposed steeper slopes. For Cerro Sillajhuay in the Andes at the border between Bolivia To overcome the problems regarding data gaps due and Chile, an ASTER DEM was compared with a to cloud cover or removed artifacts in some types DEM from contour lines (reference DEM) and it was of DEMs, data from different DEM types are usually found that the ASTER DEM showed increasingly merged. For example, in a first step, existing data lower elevations with increasing elevation. For the gaps in a high-Z-resolution ASTER DEM are filled Bernina Group in the Swiss Alps, a comparison of with data from a low-Z-resolution ASTER DEM; in an ASTER DEM and an SRTM DEM with a DEM a second step, remaining data gaps are filled with derived from contour lines (reference DEM) found SRTM data; in a third step, still remaining data that ASTER elevations were generally too high gaps are filled with data from the GTOPO30 DEM; (8.3 m on average), and SRTM elevations were the final step is a filtering for smoothing purposes. generally too low (–9.8 m on average). Berthier et Although the final DEM includes different horizontal al. (2007) compared SRTM DEM elevations with and vertical scales across the study area, it is useful SPOT-5 DEM elevations for nonglaciated terrain in for some applications. 103 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains For a proposed glacier monitoring program in the After generation of ortho-images, the root mean HKH region, gaining access to high-quality DEMs square error (RMSE), which summarizes the residual can be problematic. It is crucial that one single type values, is calculated. If the RMSE is within an of DEM is available across the whole area and that acceptable limit, elevation information can then be the data are of sufficient quality and of low cost or, extracted from the DEM. Quincey et al. (2013) put even better, free of charge. These requirements limit the maximum RMSE at two times the pixel size of the the pool of potentially useful DEM sources for glacier imagery. For ASTER DEMs, Lang and Welch (1999) monitoring of extensive areas. For example, for the suggested the RMSE should be ±7 to ±30 m. The politically sensitive HKH region, many border areas RMSE has become the standard measure of DEM are not accessible to researchers for GCP collection accuracy, although it does not very well describe the in the field, and the quality of existing topographic statistical distribution of the vertical error. map series is often not known. Therefore, at this point, ASTER GDEM and SRTM3 data provide the 6.9.3 Ground Control Points best (or only) available and least expensive sources. Although errors in ASTER-derived DEMs of steep, GCPs can be derived from traditional surveying high-mountain relief sometimes produce artifacts in or from global navigation satellite systems in classification results, they still provide a quality of combination with field receivers (differential and output that exceeds classification methods that do postprocessed), or they can come from identifying not employ a DEM of any sort (Quincey et al. 2013). landscape features on topographic maps. GCPs are important for DEM quality assessment, as they 6.9.2 Error Calculation give relatively accurate information on location and elevation of a landscape feature. Quincey et al. In the DEM generation process, vertical (altimetric) (2013) put the minimum accuracy of any ground or horizontal (planimetric) errors might be control data at twice that of the expected DEM. introduced. Planimetric errors result from horizontal Toutin (2008) showed that accuracy, number, and shifts of elevation values to erroneous geographic distribution of GCPs across the imagery can have coordinates and can be corrected by quantifying a significant impact on DEM quality. It is difficult to the orthophoto parallax followed by modeling, give a minimum number of required GCPs; while while altimetric errors are often systematic biases more than the theoretically necessary four GCPs and might increase or decrease elevations in the usually improves DEM quality, low-accuracy GCPs DEM (Quincey et al. 2013). Knowledge about the might actually produce low-quality DEMs (Quincey planimetric and altimetric errors is of importance, et al. 2013). If no GCPs are available at all, only because smaller changes in glacier margin position relative DEMs may be produced. or glacier surface altitude by only some meters might be within the error and thus might not be detected 6.9.4 Postprocessing or be only partly detected. This erroneous mapping could lead to incorrect calculations of, for example, DEMs derived from contour maps and satellite data glacier volume, mass balance, and meltwater require the interpolation of (calculated) elevation discharge. values between (measured) elevation sample points. The interpolators most commonly used are inverse To derive elevation from the stereo images, distance weighted (IDW), spline, triangular irregular photogrammetric techniques are employed using network (TIN), and kriging. All of these come with suitable imaging processing software. This process advantages and disadvantages: IDW that strongly requires information about the geometry between the weights only a few sample points produces a rough sensor and terrain at the time of image acquisition. surface; TINs often produce a typical “terracing” 104 Satellite Imagery and Digital Elevation Models effect in valleys and on ridges; spline interpolators analyses involving DEM and satellite imagery. often produce overshoots (“holes” and “spikes”); The most common software packages in use and and kriging requires experience when applying a currently available are as follows: semivariogram model (Quincey et al. 2013). For all interpolators, it seems to be the case in general that Satellite Imagery and DEMs the smoother the terrain and wider the contour lines on related topographic maps, the lower the quality • Geomatica OrthoEngine of PCI Geomatics of the resulting DEM. Although it has been proposed (www.pcigeomatics.com). Probably the most to introduce breaklines in the generation of DEMs frequently used software for DEM generation. for relatively flat terrain, it seems that they do not Horizontal and vertical accuracy of ±15 me significantly improve overall DEM quality (Quincey and ±20 m (1 σ), respectively. Postprocessing et al. 2013). A solution for relatively flat terrain tool for manual editing. Tools can also be could be to manually include additional contour incorporated into ESRI ArcGIS. Windows and lines derived from topographic maps or stereoscopic Linux operating systems. License costs: analysis of aerial photographs, but this editing is Contact PCI Geomatics customer service for time consuming. current pricing;. • LPS Core of ERDAS (www.erdas.com). Provides DEMs can include different types of errors, such as the most comprehensive control over the DEM those that result from pixel matching, correlation extraction and editing process. Postprocessing failures, cloud cover, shadowing, and low image tool for manual editing. Tools can also be contrast over clean snow and ice, any of which can incorporated into ESRI ArcGIS. Windows erroneously represent real position or elevation operating system. License costs: US$2,800; (Quincey et al. 2013). If this is the case, in a • ENVI DEM Extraction Module postprocessing step, manual editing or fusion (http://www.exelisvis.com). Accuracy of better of data from the originally generated DEM and than ±20 m for relative DEMs, and for absolute data from other sources helps reduce the error. DEMs ±30 horizontal and ±15 m vertical. For example, artifacts in ASTER GDEM have been Postprocessing tool for manual editing. Tools can successfully filled with SRTM or GTOPO30 data. also be incorporated into ESRI ArcGIS. Windows, It is important to understand that this DEM fusion Macintosh, and Unix operating systems. License generates a DEM that includes varying spatial costs: about US$3,500; resolution or resampling, topographic detail, and • Desktop Mapping System Softcopy, Version accuracy within its grid. Such “corrected” DEMs 5.0. Horizontal and vertical accuracy of ±15 are specifically problematic in multitemporal DEM m to ±25 m (1 σ). Windows operating system. analyses, for example, of changes in glacier surface License costs: about US$6,000; and elevation. Quincey et al. (2013) summarized • Silcast of Sensor Information Laboratory Corp., systematic errors in SRTM data reported by a number Japan (www.silc.co.jp). Exclusively developed of authors who found a nearly linear elevation bias for ASTER DEM generation. Written in IDLR6.1; for parts of the Alps, Patagonia, the Himalaya, and can be executed with IDL VM without IDL license. British Columbia, and reported regional patterns of Does not accept GCPs. Automatic postprocessing the bias. of errors without any operator control. Was used for generation of global (up to 83° latitude) 6.9.5 Software Packages GDEM with 30 m x 30 m ground resolution with ±7 m accuracy (1 σ). Windows, Macintosh, and Several institutional, private, and commercial Linux operating systems. License costs: about software packages are available for geospatial US$4,500/license noncommercial mode. 105 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Satellite Imagery, DEMs, and GIS Macintosh, and Linux operating systems. License costs: Open source; • ESRI ArcGIS (www.esri.com/software/arcgis/ • SAGA GIS (www.saga-gis.org). Designed index.html). ESRI has become the leader with for easy implementation of spatial algorithms regards to GIS software. Its latest version and geospatial methods to analyze vector and (ArcGIS 10) includes spatial analysis, data raster datasets through an easily approachable management, mapping, visualization, and interface. Developed by the Department of advanced imagery processing capabilities. Field Physical Geography, Göttingen University. capabilities are available through ArcGIS Mobile. Windows and Linux operating systems. License Windows operating system. License costs: about costs: Open source; and US$1,500–7,000 depending on license level; • gvSIG (www.gvsig.org). Desktop GIS application • IDRISI (www.clarklabs.org). An integrated raster- used for capturing, storing, and analyzing any based GIS software with nearly 300 modules kind of geographic information. Several features for the analysis and display of digital spatial available for the analysis of vector and raster information. Windows operating system. License remote sensing data. Licensed under the GNU costs: US$1,750 floating license (single user at General Public License and available in several one time), US$1,250 new single license, US$425 different languages. Windows, Macintosh, and upgrade license; Linux operating systems. License costs: • GRASS (grass.osgeo.org). Originally developed Open source. by the United States Army Construction Engineering Research Laboratory in the 6.10 DEM Analysis 1980s, this GIS software is used for geospatial data management and analysis, including 6.10.1 Geomorphometry image processing, geospatial modeling, and visualization. The software is supported by the Geomorphometry – the science of quantitative Open Source Geospatial Foundation. Windows, land surface analysis – draws from mathematics, Macintosh, and Linux operating systems. License computer science, and geosciences (Pike 1995, costs: Open source; 2000). The field of geomorphometry has two • OSSIM (www.ossim.org). Provides advanced modes of study: the study of individual or specific geospatial image processing and GIS landforms and features (such as glaciers), and the capabilities. Has been under development study of the general land surface or region, such since 1996 and has continued support from the as the Himalaya (Evans 1972). Glacier mapping open source software development community. usually includes the geomorphometric analysis of the Windows, Macintosh, and Linux operating glacier surface, and most software packages have systems. License costs: Open source; tools for such geomorphometric analysis. However, • QuantumGIS (www.qgis.org). User-friendly GIS as much as geomorphometric parameters help in software licensed under the GNU General Public identifying, describing, and classifying glaciers, their License and official project of the Open Source quality depends on the accuracy of the input DEM, Geospatial Foundation. Supports numerous and studies have shown that, for many applications, vector, raster, and database formats. Capable increased DEM resolution only introduces disruptive of performing terrain analysis, hydrological noise (Quincey et al. 2013). In general, it is essential modeling, and additional analysis through that systematic biases are removed prior to the an extensible plug-in architecture. Windows, geomorphometric analysis. 106 Satellite Imagery and Digital Elevation Models 6.10.2 Land Surface Parameters characterize individual geomorphic features of the glacial system, such as lateral moraines, crevasses, or Most glacier monitoring approaches include the icefalls, based on their shape and textural properties analysis of primary (first-order derivates of an (Quincey et al. 2013). Two of the available methods elevation field) and secondary (second-order are variogram and fractal dimension analyses. derivates of an elevation field) land surface The former has been used to characterize terrain parameters. The most common primary land surface morphological features and to calibrate secondary parameters are slope angle, aspect, and hypsometry; morphological indices, while the latter helps to the most common secondary parameters are improve landform classification accuracy and to planimetric curvature and vertical curvature, which explain operational-scale and process–structure both describe shape (Kamp, Bolch, and Olsenholler relationships (Quincey et al. 2013). 2003). By clustering surfaces with similar curvature characteristics, it is possible to differentiate between 6.10.3 Topographic Radiation Modeling glacier surfaces and valley bottoms (low convexity), ridges and lateral moraines (high convexity), medial Glacier monitoring increasingly includes the moraines (moderate convexity), and the transitions modeling of solar radiation variation across the between glacier margins and lateral moraines (high study area. Surface orientation and shadowing concavity) (Quincey et al. 2013). Since the glacier are important parameters that can easily be surface topography changes with time because of calculated using a DEM employing standard glacier movement, mass balance changes, or other software packages; surface irradiance and ablation factors, so do the geomorphometric parameters. gradient calculations are input parameters for Hence, the description of the latter in multitemporal energy balance melt models that are fundamental glacier analysis can provide information about to understanding the relationship between glacier glacial changes. For example, an increasing behavior and climate (Quincey et al. 2013). Fine- steepness at the glacier margin over years might resolution modeling has been used to calculate melt reflect a glacier advance. A simple but effective rates of snow and ice; however, accurate field and method to describe glacier surface elevation meteorological data are required. changes is to compare the topographic profiles along the flow line of the glacier from multitemporal 6.10.4 Altitudinal Functions DEMs (Quincey et al. 2013). When DEMs are combined with complementary GIS data such as Elevation is one of the most important factors land cover classifications, and neighborhood and responsible for variations in glacier characteristics, change detection analyses, slope angle data can and the most fundamental altitude is the ELA that be used as the primary morphological characteristic separates ablation and accumulation areas of the to delineate glacial terrain. However, specific slope glacier. In glacier monitoring, it is usually assumed thresholds vary from region to region, with ice that the ELA is equal to the snowline position at the masses in the European Alps generally characterized end of the melting season (Quincey et al. 2013). by short, steep tongues relative to the long (greater One simple method to extract the ELA using DEM than 10 km), flat tongues of some glaciers in the elevation values is the toe-to-summit method (Benn Himalaya (Quincey et al. 2013). and Lehmkuhl 2000), which can be performed efficiently, especially with well-defined valley glaciers. Increasingly, automated and semiautomated glacier ELA calculations are fundamental for estimating mapping approaches also include the analysis of the the mass balance of a glacier and, thus, serve as hierarchical organization of topography and aim to indicators for climate variations. 107 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains The area–altitude profile (hypsometric curve) [T]he severe geometric and radiometric and the altitude–slope function (clinometry) help distortions and speckle (“noise”) require in describing erosion rates and mass balance complicated processing and accurate digital fluctuations within a glacial system (Quincey et al. elevation models (DEMs), which are not 2013). Differential erosion occurs with altitude, always readily available. Other techniques and studies have shown that glaciation produces such as passive microwave systems, radar, and the greatest mesoscale relief at high altitudes, while laser altimetry show promise for increasing our warm-based glaciation reduces relief at intermediate understanding of glacier characteristics in altitudes (Bishop et al. 2001). Altitude–velocity the Himalayas. profiles are used in mass balance calculations and glacier flow regimes. For a proposed institutionalized glacier monitoring program that covers the entire HKH region, it is 6.11 Summary crucial to keep the costs of source data as low as possible. Thus, it is easier to make use of 6.11.1 Satellite Imagery existing and free available datasets for regionwide monitoring, and then purchase satellite imaging and Despite the great variety of available satellite DEM data for monitoring the glaciers of specific imagery – including at large-scale resolution – in interest, for example, identified benchmark glaciers their review of the use of optical remote sensing in or glaciers that represent a potential hazard. At glacier characterization, Racoviteanu, Williams, this time, imagery from Landsat and ASTER has and Barry (2008) concluded that medium-scale some important advantages: it covers the entire (10–90 m) resolution optical sensors in multispectral HKH region; it comes at low cost or even no cost; mode, with relatively large swath widths and short it has sufficient spatial resolution for most relevant revisit cycles, are useful for regular glacier mapping glacier analyses; it has extensive experience in its over extensive areas, and that ASTER may still be application in glacier studies, with existing results; the most suitable sensor for glacier monitoring that and ASTER data are also used for the generation of includes mass balance applications. Authors of this the ASTER GDEM. study support that view and suggest that Landsat and ASTER are probably the most useful sensors for a 6.11.2 DEMs proposed glacier monitoring program in the HKH region. After reviewing existing studies, particularly the review by Quincey et al. (2013), authors of this However, sensors operating in the visible and near IR study conclude that, today, DEMs are fundamental (VNIR) ranges, such as Landsat, ASTER, and ALOS, and a standard tool in any glacier monitoring are limited to daylight and cloud-free conditions, project. Square-gridded DEMs are most widely used, because they provide more realistic terrain which are difficult to obtain in the HKH region representations than DEMs derived from other data (Racoviteanu, Williams, and Barry 2008). Although sources. In particular, ASTER remains the most widely synthetic aperture radar is efficient at night and in used data source for DEM generation because of cloudy conditions, the authors caution that: its stereoscopic capability, wide spectral range, 108 Satellite Imagery and Digital Elevation Models medium-to-fine spatial resolution, and, importantly, represented in DEM data. Therefore, while the its low cost (Quincey et al. 2013). For many regions resolution of the generated DEM is limited by the with glaciers, ASTER DEMs come with an accuracy resolution of the input source data, those making of ±15–30 m after some significant postprocessing. scientific interpretations should also be mindful to However, in areas with steep rock headwalls restrict analyses to the natural scale of the terrain- and large, low-contrast accumulation areas, the dependent application; accuracy is often only ±60 m. • As a guide, the resolution of the derived DEM can provide a practical indication of the scale At the same time, DEMs are far from perfect. of information content; analyses of processes Examples of the many drawbacks that still exist are or features that occur on a finer scale than this found in the literature (Evans 1972; Kamp, Bolch, should be made with caution; and Olsenholler 2003; Pike, Evans, and Hengl • Currently, scientists employ a range of analytical 2009; Quincey et al. 2013), and the following tools, algorithms, processing approaches, and are typical: software for the generation, manipulation, and interpretation of topographic data. Methods are • It is not easy to detect abrupt changes in highly empirical, thus the type and quality of the topography; derived data are dependent, to a large extent, • Producing fine-resolution DEMs with low data on the analyst. Consequently, the replication of storage requirements is difficult; existing results can be difficult; even more so the • It is also difficult to generate reliable elevation application of one published technique to a new data at any resolution finer than that offered by area; and SRTM (90 m) for many regions; • Finally, challenges remain in identifying and • Persistent cloud cover affects optical sensors, quantifying altimetric errors, particularly when particularly in mountain regions; comparing DEMs from different sources, and in • Lack of ground control data, particularly in areas simply gaining access to reliable data for some of of political sensitivity and rugged terrain, makes it the most politically sensitive regions of the world. difficult to verify results; • Inconsistencies occur between studies in the Several methods have been prescribed by Quincey manner in which errors are reported. A limited et al. (2013) to correct for the limitations on number do not even quantify the expected achieving optimum results that are present with all error (or uncertainty) in the presented data, current DEMs: and, of those that do, the accuracy data can sometimes be highly dependent on the number • Longer observational periods to detect and positioning of chosen checkpoints. This magnitudes of change in multitemporal studies; makes it difficult for an independent researcher • Adhering to the limitations of resolution in the to replicate results, or even establish the exact scale of the information and terrain-dependent error quantification that has taken place and how applications; and reliable it is; • Developing standards and protocols for • The interaction between topography and most information extraction and integration in order geophysical processes occurs over a range of to maintain data quality and accuracy across spatial scales, which cannot currently be truly different study areas and research efforts. 109 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains References Satellite Multispectral Imaging of Glaciers, ed. J.S. Kargel, G.J. Leonard, M.P . Bishop, A. Kääb, and Bajracharya, S.R., and B. Shrestha, eds. 2011. B. Raup. Berlin: Praxis-Springer. The Status of Glaciers in the Hindu Bishop, M.P ., J.F. Shroder, A.B.G. Bush, U. Kush-Himalayan Region. Technical Paper. Haritashya, H.H.H. Bulley, and J. Olsenholler. Kathmandu: International Centre for Integrated 2013. “Afghanistan and Pakistan.” In Global Mountain Development (ICIMOD). Land Ice Measurements from Space: Satellite Benn, D.I., and F. Lehmkuhl. 2000. “Mass Balance Multispectral Imaging of Glaciers, ed. J.S. Kargel, and Equilibrium Line Altitudes of Glaciers in G.J. Leonard, M.P . Bishop, A. Kääb, and B. Raup. High Mountain Environments.” Quaternary Berlin: Praxis-Springer. International 65/66: 15–29. Bishop M.P ., J.F. Shroder, and J.L. Ward. 1995. “SPOT Multispectral Analysis for Producing Berthier, E., Y. Arnaud, K. Rajesh, A. Sarfaraz, Supraglacial Debris-Load Estimates for Batura P. Wagnon, and P . Chevallier. 2007. “Remote Glacier, Pakistan.” Geocarto International 10 (4): Sensing Estimates of Glacier Mass Balances 81–90. in the Himachal Pradesh (Western Himalaya, India).” Remote Sensing of Environment 108 (3): Bolch, T., M. Buchroithner, A. Kunert, and U. 327–38. Kamp. 2007. “Automated Delineation of Debris- Covered Glaciers Based on ASTER Data.” In Bishop, M.P ., R.G. Barry, A.B.G. Bush, L. Copeland, GeoInformation in Europe: Proceedings of the J.L. Dwyer, A.G. Fountain, W. Haeberli, D.K. 27th Symposium of the European Association of Hall, A. Kääb, J.S. Kargel, B.F. Molina, J.A. Remote Sensing Laboratories (EARSeL), June 4–7, Olsenholler, F. Paul, B.H. Raup, J.F. Shroder, D.C. 2007, Bozen, Italy, ed. M.A. Gomarasca, Trabant, and R. Wessels. 2004. “Global Land 403–10. Rotterdam: Millpress. Ice Measurements from Space (GLIMS): Remote Sensing and GIS Investigations of the Earth’s Bolch, T., and U. Kamp. 2006. “Glacier Mapping in Cryosphere.” Geocarto International 19 (2): High Mountains Using DEMs, ASTER and Landsat 57–85. Data.” Proceedings of the eighth International Symposium on High Mountain Remote Sensing Bishop, M.P., R. Bonk, U. Kamp, and J.F. Shroder. Cartography, March 20–27, 2005, La Paz, 2001. “Terrain Analysis and Data Modeling for Bolivia. Grazer Schriften der Geographie und Alpine Glacier Mapping.” Polar Geography 25 Raumforschung 41: 13–24. (3): 182–201. Bolch, T., U. Kamp, and J. Olsenholler. 2005. Bishop, M.P., A.B. Bush, E. Collier, L. Copland, U.K. “Using ASTER and SRTM DEMs for Studying Haritashya, S.F. John, S.C. Swenson, and J. Wahr. Geomorphology and Glaciation in High 2008. Advancing Glaciers and Positive Mass Mountain Areas.” In New Strategies for European Anomaly in the Karakoram Himalaya, Pakistan. Remote Sensing: Proceedings of the 24th Meeting Paper presented at the American Geophysical of the European Association of Remote Sensing Union Fall Meeting, December 15–19, 2008, Laboratories (EARSeL), 25–27 May 2004, San Francisco, United States. Dubrovnik, Croatia, ed. M. Oluic, 119–27. Bishop, M.P., A.B. Bush, R. Furfaro, A.R. Gillespie, Rotterdam: Millpress. D.K. Hall, U. Haratashya, B. Raup, S.J.S. Khalsa, Bolch, T., B. Menounos, and R. Wheate. 2009. and R. Armstrong. 2013. “Remote Sensing “Landsat-Based Inventory of Glaciers in Western Science and Technology for Glacier Assessment.” Canada, 1985–2005.” Remote Sensing of In Global Land Ice Measurements from Space: Environment 114: 127–37. 110 Satellite Imagery and Digital Elevation Models Bolch, T., T. Yao, S. Kang, M.F. Buchroithner, Kääb, A., F. Paul, M.P. Bishop, J. Kargel, R. Furfaro, D. Scherer, F. Maussion, E. Huintjes, and C. B. Raup, A. Gillespie, and A. Nolin. 2013. Schneider. 2010. “A Glacier Inventory for the “Glacier Mapping and Monitoring Based on Western Nyainqentanglha Range and Nam Co Spectral Data.” In Global Land Ice Basin, Tibet, and Glacier Changes 1976–2009.” Measurements from Space: Satellite Multispectral The Cryosphere 4: 419–33. Imaging of Glaciers, ed. J.S. Kargel, G.J. Leonard, M.P . Bishop, A. Kääb, and B. Raup. Dinku, T., S. Chidzambwa, P . Ceccato, S.J. Connor, Berlin: Praxis-Springer. and C.F. Ropelewski. 2008. “Validation of High Resolution Satellite Rainfall Products Kamp, U., T. Bolch, and J. Olsenholler. 2003. over Complex Terrain.” International Journal “DEM Generation from ASTER Satellite Data for of Remote Sensing 29 (14): 4097–110. Geomorphometric Analysis of Cerro Sillajhuay, doi:10.1080/01431160701772526. Chile/Bolivia.” In ASPRS 2003 Annual Proceedings, 5–9 May 2003, Anchorage, Alaska, USA. Evans, I.S. 1972. “General Geomorphometry, Derivatives of Altitude, and Descriptive Statistics.” Kamp, U., M. Byrne, and T. Bolch. 2011. “Glacier In Spatial Analysis in Geomorphology, ed. R.J. Fluctuations between 1975 and 2008 in the Chorley, 17–90. Harper and Row. Greater Himalaya Range of Zanskar, Southern Ladakh.” Journal of Mountain Science 8 (3): Forsythe, N., C. Kilsby, H. Fowler, and D. Archer. 374–389. 2011. Assessing Climate Pressures on Glacier- Melt and Snowmelt-Derived Runoff in the Hindu Kamp, U., B. Krumwiede, K.G. McManigal, M. Kush-Karakoram Sector of the Upper Indus Walther, and A. Dashtseren. 2013. “Glaciers of Basin. Newcastle University, School of Civil Mongolia.” In Satellite Image Atlas of Glaciers Engineering and Geosciences, Water Resources of the World: Glaciers of Asia, ed. R.S. Williams Research Group. Jr. and J.G. Ferrigno (online addendum). Washington, DC: United States Geological Fujita, K., R. Suzuki, T. Nuimura, and A. Sakai. Survey, U.S. Government Printing Office. 2008. “Performance of ASTER and SRTM DEMs, and Their Potential for Assessing Glacial Lakes in Kamp, U., K.G. McManigal, A. Dashtseren, and M. the Lunana Region, Bhutan Himalaya.” Journal of Walther. 2013. “Documenting Glacial Changes Glaciology 54 (185): 220–28. between 1910, 1970, 1992 and 2010 in the Turgen Mountains, Mongolian Altai, Using Hall, D.K., G.A. Riggs, V.V. Salomonson, J.S. Repeated Photographs, Topographic Maps and Barton, K. Casey, J.Y.L. Chien, N.E. DiGirolamo, Satellite Imagery.” Geographical Journal 179 (3): A.G. Klein, H.W. Powell, and A.B. Tait. 2001. 248–63. Algorithm Theoretical Basis Document (ATBD) for the MODIS Snow and Sea Ice-mapping Kargel, J.S., M.J. Abrams, M.P . Bishop, A. Bush, Algorithms. http://modis.gsfc.nasa.gov/data/ G. Hamilton, H. Jiskoot, A. Kääb, H.H. Kieffer, atbd/atbd_mod10.pdf. E.M. Lee, F. Paul, F. Rau, B. Raup, J.F. Shroder, D.L. Soltesz, L. Stearns, and R. Wessels. 2005. Higuchi, K., O. Watanabe, H. Fushimi, S. Takenaka, “Multispectral Imaging Contributions to Global and A. Nagoshi. 2010. “Glaciers of Nepal: Land Ice Measurements from Space.” Remote Glacier Distribution in the Nepal Himalaya with Sensing of Environment 99 (1–2): 187–219. Comparisons to the Karakoram Range.” In Satellite Image Atlas of Glaciers of the World: Glaciers of Kargel, J.S., G.J. Leonard, M.P. Bishop, A. Kääb, Asia, ed. R.S. Williams Jr. and J.G. Ferrigno, 293– and B. Raup, eds. 2013. Global Land Ice 320. Washington, DC: United States Geological Measurements from Space: Satellite Multispectral Survey, U.S. Government Printing Office. Imaging of Glaciers. Berlin: Praxis-Springer. 111 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Krumwiede, B.S., U. Kamp, G.J. Leonard, A. Pike, R.J., I.S. Evans, and T. Hengl. 2009. Dashtseren, and M. Walther. 2013. “Recent “Geomorphometry: A Brief Guide.” In Glacier Changes in the Altai Mountains, Western Geomorphometry: Concepts, Software, Mongolia: Case Studies from Tavan Bogd Applications, ed. T. Hengl and H.I. Reuter, 1–30. and Munkh Khairkhan.” In Global Land Ice Developments in Soil Science Series, Elsevier. Measurements from Space: Satellite Multispectral Quincey, D., M.P . Bishop, A. Kääb, E. Berthier, Imaging of Glaciers, ed. J.S. Kargel, B. Flach, T. Bolch, M. Buchroithner, U. Kamp, G.J. Leonard, M.P . Bishop, A. Kääb, and S.J.S. Khalsa, T. Toutin, U. Haritashya, A.E. B. Raup. Berlin: Praxis-Springer. Racoviteanu, J.F. Shroder, and B. Raup. Lang, H.R., and R. Welch. 1999. Algorithm 2013. “Digital Terrain Modeling and Glacier Theoretical Basis Document for ASTER Digital Topographic Characterization.” In Global Elevation Models, Version 3.0. Pasadena: Jet Land Ice Measurements from Space: Satellite Propulsion Laboratory. Multispectral Imaging of Glaciers, ed. J.S. Kargel, G.J. Leonard, M.P . Bishop, A. Kääb, and B. Raup. Pan, C., U. Kamp, K. McManigal, and M. Walther. Berlin: Praxis-Springer. 2012. Inventory of Mongolian Glaciers for the Global Land Ice Measurements from Space Racoviteanu, A.E., T. Bolch, J. Kargel, G. Leonard, (GLIMS) Program. Presentation at the University of A. Kääb, U. Kamp, E. Berthier, Y. Arnaud, A.V. Montana Graduate Student and Faculty Research Kulkarni, M.P. Bishop, J.F. Shroder, I.M. Baghuna, Conference, April 14, 2012, Missoula, MT, R. Bhambri, R. Furfaro, S. Bajracharya, and P . United States. Mool. 2013. “Himalayan Glaciers (India, Bhutan, Nepal).” In Global Land Ice Measurements from Paul, F., R.G. Barry, J.G. Cogley, H. Frey, W. Space: Satellite Multispectral Imaging of Glaciers, Haeberli, A. Ohmura, C.S.L. Ommanney, ed. J.S. Kargel, G.J. Leonard, M.P . Bishop, A. B. Raup, A. Rivera, and M. Zemp. 2009. Kääb, and B. Raup. Berlin: Praxis-Springer. “Recommendations for the Compilation of Glacier Inventory Data from Digital Sources.” Racoviteanu, A.E., M.W. Williams, and R.G. Barry. Annals of Glaciology 50 (53): 119–26. 2008. “Optical Remote Sensing of Glacier Characteristics: A Review with Focus on the Paul, F., C. Huggel, and A. Kääb. 2004. Himalaya.” Sensors 8 (5): 3355–83. “Combining Satellite Multispectral Image Data and a Digital Elevation Model for Mapping Ramachandran, B., J. Dwyer, B. Raup, and J.S. Debris-Covered Glaciers.” Remote Sensing of Kargel. 2013. “Satellite Image Acquisition, Environment 89 (4): 510–18. Preprocessing, and Special Products.” In Global Land Ice Measurements from Space: Satellite Paul, F., A. Kääb, and W. Haeberli. 2007. “Recent Multispectral Imaging of Glaciers, ed. J.S. Kargel, Glacier Changes in the Alps Observed from G.J. Leonard, M.P . Bishop, A. Kääb, and B. Raup. Satellite: Consequences for Future Monitoring Berlin: Praxis-Springer. Strategies.” Global and Planetary Change 56 (1–2): 111–22. Rau, F., F. Mauz, S. Vogt, S.J.S. Khalsa, and B. Raup. 2005. Illustrated GLIMS Glacier Classification Pike, R.J. 1995. “Geomorphometry: Progress, Manual: Glacier Classification Guidance for the Practice, and Prospect.” Zeitschrift für GLIMS Glacier Inventory, Version 1.0, 2005-02- Geomorphologie, Supplementband 10. http://www.glims.org/MapsAndDocs/assets/ 101: 221–38. GLIMS_Glacier-Classification-Manual_V1_2005- Pike, R.J. 2000. “Geomorphometry: Diversity in 02-10.pdf. Quantitative Surface Analysis.” Progress in Physical Geography 24 (1): 1–20. 112 Satellite Imagery and Digital Elevation Models Raup, B.H., A. Kääb, J.S. Kargel, M.P . Bishop, G. Toutin, T. 2008. “ASTER DEMs for Geomatic Hamilton, E. Lee, F. Paul, F. Rau, D. Soltesz, S.J.S. and Geoscientific Applications: A Review.” Khalsa, M. Beedle, and C. Helm. 2007. “Remote International Journal of Remote Sensing 29 (7): Sensing and GIS Technology in the Global Land 1855–75. Ice Measurements from Space (GLIMS) Project.” Vohra, C.P. 2010. “Glaciers of India: A Brief Computers and Geosciences 33: 104–25. Overview of the State of Glaciers in the Indian Raup, B., A. Racoviteanu, S.J.S. Khalsa, C. Helm, Himalaya in the 1970s and at the End of the R. Armstrong, and Y. Arnaud. 2007. “The GLIMS 20th Century.” In Satellite Image Atlas of Glaciers Geospatial Glacier Database: A New Tool for of the World: Glaciers of Asia, ed. R.S. Williams Studying Glacier Change.” Global and Planetary Jr. and J.G. Ferrigno, 259–91. Washington, Change 56 (1–2): 101–10. DC: United States Geological Survey, U.S. Government Printing Office. Shroder, J.F., and M.P. Bishop. 2010. “Glaciers of Pakistan.” In Satellite Image Atlas of Glaciers of Williams, R.S. Jr., and J.G. Ferrigno, eds. 2010. the World: Glaciers of Asia, ed. R.S. Williams Satellite Image Atlas of Glaciers of the World: Jr. and J.G. Ferrigno, 201–57. Washington, Glaciers of Asia. Washington, DC: United States DC: United States Geological Survey, U.S. Geological Survey, U.S. Government Printing Government Printing Office. Office. 113 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains 7. Monitoring of the HKH Cryosphere If it is to be successful, cryosphere monitoring in considered here make no special provisions for extra the HKH Mountains must be developed as a broad allowances that would make it attractive to staff to program, involving institutional collaboration, work under harsh conditions. staff training, instrument network development, coordination of monitoring methodologies and Institutional approaches for HKH monitoring are as procedures, and competent management and follows: oversight. The Integrated Global Observing Strategy (IGOS) cryosphere theme serves as a • Monitoring through institutionalized government useful guide in meeting these requirements. IGOS authorities and procedures with a long-term unites the major satellite and ground-based perspective and through dedicated units systems for global environmental observations of anchored in the responsible institutions; the atmosphere, oceans and land in a framework • Monitoring through indigenous and foreign that delivers maximum benefit and effectiveness in research programs on a longer-term basis at their final use. It is a strategic planning process, well-established sites; involving many partners, that links research and • Monitoring through nongovernmental operational programs as well as data producers organizations, or even local companies, that and users. IGOS has compiled the experiences execute monitoring tasks under well-defined of high-mountain monitoring programs and has framework conditions. If governments were to produced lessons learned and recommendations outsource monitoring to private entities, the that are now integrated in the follow-up Global advantage would be that payments are then Cryosphere Watch program. This chapter will made independently of government rules that summarize these observations on snow and glacier apply to civil servants; and issues. The recommendations made here will build • Periodic research expeditions similar to the on the experience of past programs, look at future Japan Glaciological Expedition to Nepal, which requirements, and work on the basis of well-known, collected over 15 years of data and observations, robust, and reliable technologies and procedures could offer an opportunity complementary to institutionalized monitoring activities. These would 7.1 Considerations and Technical be complementary because local governments, Procedures for HKH Monitoring thus far, have not been prepared to take up monitoring at the end of a project, either in terms In addition to the physical challenges of remoteness, of cost or expertise. altitude, terrain, and temperature, the challenges of carrying out monitoring of glaciers climate and Depending on local conditions, logistic arrangements runoff in very high mountains, especially in the HKH are made by the executing entity or with the assistance region, include insufficient funding of continued, of a well-established trekking agency as a partner. long-term observation programs, weak institutions, Finding human resources (for example, porters) for and the difficulty of regularly sending government logistic support has recently become difficult and officials to remote, high-mountain regions on more expensive, at least for the conditions in Nepal, missions that often last several weeks. Also, civil as younger people migrate out of the country in large service rules in the countries of the HKH Mountains numbers for better job opportunities. 114 Monitoring of the HKH Cryosphere 7.2 Selection of Monitoring Sites and 7.3 Practical Procedures for Monitoring Logistical Considerations the HKH Cryosphere Monitoring sites should meet the following technical 7.3.1 Guiding Principles and organizational criteria: Recommended guiding principles for establishing a • Sites should be representative of the phenomena cryosphere monitoring network in the HKH are as to be monitored; follows: • Existing sites should be refurbished or upgraded to a working condition to allow continuity of • Adopt Global Climate Observing System (GCOS) earlier data collection programs; monitoring principles for all operational satellites • Sites should be accessible for as long as possible and in situ sites; and throughout the year; • Observations should follow, to the extent • En route support from local villages should be possible, the well-established Global Hierarchical available; and Observing Strategy that has been developed by • Establishment of telecommunication links should the Global Terrestrial Observing System, and be possible. which is a standard applied by the World Glacier Monitoring Service. Local observers should be available close to the site to keep equipment in repair and guard against 7.3.2 Essential Variables vandalism, and have access to station houses set up with essential supplies and spare parts. Essential (climate-sensitive) variables, as defined by GCOS, that should be observed in a monitoring In general, field visits and station maintenance program include the following hydrometeorological need to be undertaken using local facilities. For variables: cost-effectiveness and sustainability of the installed infrastructure, it is not advisable to leave all • Solid precipitation: In situ climate and synoptic observations, maintenance, and station surveillance (manual, auto), remote sensing; to office staff back in the city; rather, this effort • Snow: Snow water equivalent, depth, extent, should be delegated as much as possible to density, snowfall, albedo, in situ climate and locally available staff, who may take great pride synoptic (manual, auto), remote sensing; in performing these works if their services are • Glaciers: Mass balance (accumulation/ adequately recognized and rewarded. ablation), thickness, area, length, (geometry), firn temperature; For example, after a project ends, using helicopters • Snowline/equilibrium line: Snow on ice, ground for supplying stations is not usually affordable for based (in situ), remote sensing; a government organization or any other locally • Frozen ground/permafrost: Soil temperature/ operating entity. Thus monitoring programs need thermal state, active layer thickness, borehole local, well-trained technical personnel to reduce travel temperature, extent; and and mission costs and time lags in reaching a station • Snow cover: In situ (manual, auto), remote after a problem has occurred. This is technically sensing. feasible through adequate capacity-building programs that enable local personnel to perform essential In addition, the following factors need to be technical functions based on well-defined, station- measured on a regular basis: specific, standing operating procedures. 115 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains • Temperature; 7.3.4 Components of a Cryosphere • Wind; Monitoring Network • Radiation; • Humidity; Network components could include dedicated snow • Stream flow; and glacier hydrology stations at high altitudes, • Observation of glacier ablation; with preference to the rehabilitation or refurbishing • Monitoring glacier lake changes; of older existing stations with station records, and • Monitoring of glacier tongue changes; establishment of potentially new stations at high • Mass balance studies (as snapshots perhaps and altitudes (above 3,500 m) in glacierized basins. In not continuously); addition, multi-tiered observation and monitoring • Glacier varve observations and sedimentation networks could include the following elements: observations in glacier lakes and moraines; and • Snow cover and seasonal changes, snow depth • Terrestrial observations, including specialized and water equivalent. networks, existing synoptic and hydrological stations, and stations of the GCOS (terrestrial Note that observations of hydrological factors, the and upper air observations), and stations foremost of which is streamflow, are essential in the affiliated to the WMO Global Telecommunication monitoring and evaluation of snow and glacier melt System; runoff characteristics and trends. More than any • Space-based observations; other variable, streamflow observations provide an • Observations from core network stations in insight into hydrological processes in glacierized or discrete time intervals (near real time, for largely (seasonal) snow-covered basins. example); • Observations from complementary networks, 7.3.3 Requirements Document including research stations; and • Snapshot-type observing campaigns, regularly Development of core information: scheduled (every six months, once a year), such as those in place for glacier mass balance • Description of core variables to be observed; observations. • Minimum density of networks for terrestrial observations; 7.3.5 Historical Data Records • Minimum requirements for space-based observations, including repetition frequencies, A necessary activity is recovery and processing ground resolution, and observation paths; of long time series archived data relevant to the • Preferred methods of observations (for terrestrial development and construction of cryospheric and space-based observations); fundamental climate data records. This includes the • Error bandwidth of observations, achievable documentation of glacier inventories in the countries levels of accuracy of observations; and in relevant centers. • Spatial and time resolution; • Reporting frequency; Likewise, it would be useful to detect old records and • Quality assurance procedures; undertake data rescue activities as an important part • Minimum technical qualifications of staff of a monitoring program, as these records constitute performing the observations; the “climate memory” of hydrometeorological and • Required capacity-building programs; and cryospheric processes. • For field visits and on-site explorations, required field equipment and gear, and minimum required logistic preparations. 116 Monitoring of the HKH Cryosphere 7.3.6 Telecommunications institutions. The WMO Information System provides detailed guidance and information in this regard. Priority should be given to the utilization of existing telecommunication facilities, such as code division 7.4.1 Access to Data and Information multiple access (CDMA) protocols, general packet radio service (GPRS) (a mobile phone standard As has been learned in past projects and programs, communication protocol), and high frequency (HF) an agreed data policy needs to be developed, radios in Nepal; HF in Bhutan, with GPRS expanding covering the different data streams the program rapidly; and GPRS, HF radios, and, to some extent, will monitor from a multitude of sources, though meteor burst communication in Pakistan. Conditions mostly from national sources (such as national in India vary, with HF radios in place and transition hydrometeorological networks). occurring to GPRS communication and satellite communication facilities. While these technologies The guiding principles for data collection and have been established, are robust, and are management are as follows: reasonably reliable, and are supplied by commercial providers in all countries, the costs of these services • There should be equal, nonhierarchical access to (that is, prepaid mobile services) are an issue, all project or program data by all partners; especially when dues are not paid in time or SIM • Data providers are the custodians of the data cards are not electronically recharged, or budget they generate and continue to be the owners of lines have not been established to ensure continued these data, even if they are pooled or aggregated payments and thus reliable service and maintenance in program-related databases and data (for example, if a server breaks down). The use of management systems; dedicated satellite connections, such as through • Selected data are published for the general public Inmarsat, needs to be carefully considered because in a fashion agreed to by program partners along of the high operation costs. Access to and use of the lines of WMO Resolution 40 (on access to data from the WMO Global Telecommunication meteorological data) and WMO Resolution 25 System are encouraged and can be achieved (on access to hydrological data); through the national meteorological services. • Data from research projects that are partners Global data streams from satellite operations can be in the program can be utilized by partners as acquired, but based on the availability of broadband they become available and, in the interest of connections, may have limitations in use. researchers, are made public at the latest two years after generation of the data; and 7.4 Data Management • All data have to undergo a rigid data quality control procedure that has to be designed and Only main guiding principles are outlined here, implemented. as the field of data management is very wide. However, these principles must be specific under 7.4.2 Metadata well-defined conditions in the operational planning and implementation phase of the monitoring A web-based metadata catalog must be established program. In general, data lose their value if they are to ensure accessibility to data generated, including not managed in a transparent, replicable manner; critical information concerning the physical it must be ensured that data management systems properties of observations and related data, data are built on established principles of data quality sources and dates, an indication of the quality of the control and the interoperability of different data data, information on the source and format of the management systems in different countries and data, and conditions for acquiring the data. 117 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains All metadata must conform to the ISO 19115 technologies provide an integrated approach to standard for geographic metadata, and to the WMO generating spatial and temporal consistent datasets Core Metadata Profile, allowing metadata to be of snow, glaciers, hydrology, frozen soil, and other easily updated from all data centers or data services cryospheric (and related land surface) variables. of partners in the context of the program. Consequently, there is the need to develop 7.4.3 Database Management Systems integrated, operational analysis products based on cryospheric data assimilation, models, satellite- The development and establishment of a uniform generated data, and in situ data, and to develop data management architecture does not seem requirement-oriented, operational cryospheric realistic, as national entities or organizations forecasting capability. For example, the World already have well-established systems in place. Data Center for Glaciology and Geocryology However, links have to be developed that allow the in Lanzhou, China, has developed a land data interoperability of different data sources and centers, assimilation system that can assimilate remote including the management of (near) real-time data sensing observations into land surface models and with a central program database that preferably then produce reanalyzed cryospheric datasets with should be established in a dedicated regional center high spatial (0.25 degrees) and temporal (one hour) or regional centers of excellence for the purpose resolutions. of the program. Components of the overall data management system must include the following: To make the products interoperable under different institutional and country settings, standard data • Data archives for physical observations in (near) formats and protocols must be established for real time; distributed (Web-based) data visualization services. • Data archives for physical observations, quality controlled; A wide range of relevant approaches to the use of • Data archives for physical time series models to generate advanced data products as well observations; as publications are available (for example, Schaefli • Data archives for related observations, including and Huss 2011) that use specific hydrological image archives; modeling approaches coupled with glacier mass • Libraries for observation campaigns, such as for balances. From a strategic viewpoint, it would be glacier mass balance studies; advantageous to assimilate cryospheric products in • Libraries for literature, meeting reports, contact next-generation global circulation models, medium- partners; range, seasonal, and interannual forecasting, and • Product archives, including for visualized products climate models. From an applications viewpoint, it and model results; will be necessary to develop interannual forecasting • Management of real-time data; and capabilities for snow and glacier dynamics, including • Dedicated data quality assurance control mass balance changes. However, modeling aspects protocols and quality control reports. must be considered in a separate study when discussing the generation of user-oriented products 7.4.4 Data Integration and Management from the observation networks. With data from a multitude of networks, a data 7.4.5 Data Management and Reanalysis management scheme needs to be defined. A major outcome of the program should be the establishment Reanalysis of past cryosphere data presents a and operation of a cryosphere integrated data and clear picture of past conditions, independent of information service. Recent land data assimilation the many varieties of instruments used to take 118 Monitoring of the HKH Cryosphere measurements over the years. Through a variety 7.4.6 Development of Analysis of methods, observations from various instruments and Forecast Procedures are added together onto a regularly spaced grid of data. Placing all instrument observations onto a Development of complex, model-based analyses regularly spaced grid makes comparing the actual and forecasts will certainly take time and staff observations with other gridded datasets easier. with sufficient scientific and practical operations In addition to putting observations onto a grid, background. Following a pragmatic approach, reanalysis also holds the gridding model constant— it is recommended to first develop a suite of it doesn’t change the programming, keeping initial procedures that build on quality-checked the historical record uninfluenced by artificial observations. The following provides a first list of factors. Reanalysis gives a level playing field for all recommended procedures for a variety of uses and instruments throughout the historical record. It also: general information in the educated public domain: • Promotes detailed validation of reanalysis • Time series and analysis of observations with projects for cold climates and cryosphere-related recurrence periods and associated probabilities elements; of exceedance and threshold values, for example, • Promotes the use of reanalysis as a monitoring for established warning levels; tool; • Analysis of current observations and projections • Evaluates the maturity of new data products that (perhaps on a seasonal basis) in the context of can be assimilated by models or used for model historical seasons; and verification; • Use of proxy observations at lower altitudes to • Promotes the further development of data deduce hydrometeorological and glaciological assimilation schemes and objective analyses for processes at higher elevations based on simple cryospheric variables, together with a thorough regression models where this is appropriate. treatment of error covariances; • Establishes appropriate dynamical downscaling 7.5 Institutional Setup and techniques of reanalysis of data to facilitate their Organization use in cryospheric impact models that operate in high-mountain terrain at about 10–100 m spatial The underlying assumption for the institutional setup resolution or better; and governance of any envisaged program activities • Facilitates the development of a climate in a monitoring program is that its implementation system reanalysis with inclusion of cryospheric and day-to-day management would be facilitated components; through a regional institution such as ICIMOD or • Improves the utilization of satellite data in another dedicated center of excellence in one of automated analyses and incorporate fractional the participating countries. Preferably, this would ice cover and ice dynamics in global circulation be a national hydrometeorological service that is models. also mandated to undertake, on a routine basis, • Investigates indirect methods of combining snow and glacier observations. Because university multiple remote sensing products and physically staff have a much higher level of fluctuation, it based models to infer ice thickness; and is advisable that higher education institutions be • Improves algorithms for estimating global sea ice included as support partners in underpinning the concentrations from passive microwave sensors science of program activities rather than as day- by using data assimilation techniques, and to-day implementing partners. Universities would compare results with those from sensors with a also be useful in helping to conduct glaciological higher spatial resolution. measuring activities in the context of the program 119 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains and in the configuration of data and information Financial oversight, as well as a technical and products generated from observational data. scientific monitoring and evaluation, is also provided within this group, essentially a program steering To ensure broad-based governmental support, it is committee. Meeting frequency of this group could be also essential, right from the start, during country twice a year. Day-to-day execution of the program is consultations to obtain backup from relevant ensured through a management unit housed within ministries and their line agencies, particularly the principal regional organization, which would as products will be generated on the basis of a include the following personnel: sustained demand for specific products from these line agencies. • Scientific officer; • Technical officer; The program consortium should comprise the • Asset management officer; following participants: • Financial controller; and • Administrative support (secretary). • Core consortium, including those national services that contribute the majority of the data 7.6 Cryosphere Monitoring Program required for the program; Components • Partners that contribute to the program with complementary data and information; With a view to ensuring sustained program services, • Research partners and those organizations a long-duration program is envisaged. A reasonable and individuals that execute field expeditions, cost estimate cannot be made at this time and needs including periodic observation campaigns; to be discussed in light of more advanced program • Partners that contribute through the generation of planning and approaches for the implementation products and services; of the project. It is envisaged that the program will • Donors and representatives of the hosting involve the following components: organizations; and • Expert services; • Invited experts and observers on an ad hoc basis. • Financial support to national executing institutions; The full group of partners should meet initially • Financial support of implementing regional center for the program commencement and, thereafter, with its project management unit; periodically, especially when major program • Capital investments (instruments); milestones have been achieved and consensus • Capital investments (spare parts); needs to be reached about follow-up program • Capital investments (civil works at stations); phases (as an indication: every 20 months). • Communication (satellite observations, However, the group would be too large to effectively telecommunication, data streams in general); manage the program. This requires a dedicated • Routine field observations (visits to stations); group that provides the governance of the program. • Dedicated expedition-like observation campaigns Thus members of this group should be as follows: (such as glacier mass balance studies); • Capacity-building activities at national and • Identified national focal points; regional levels; • International organizations; • Study tours; • Donors and representatives of the hosting • Project coordination (for example, through regional organization; and steering committee meetings); • Invited experts on an ad hoc basis. • Project planning, country consultations, overall 120 Monitoring of the HKH Cryosphere project coordination at donor level with main engineers at the outset of system design and implementing organizations; and implementation; • Contingencies and holdback for unrealized • The conversion of research observing systems currency exchange losses and gains. to long-term operations in a carefully-planned manner should be promoted; and An in-kind contribution from national governments • Data management systems that facilitate access, and partners in the order of an additional 30–40 use and interpretation of data and products percent of the estimated cost of the project should be included as essential elements of seems reasonable. A key indicator for project climate monitoring systems. sustainability would be the demonstration of by- the-year increasing budgets for national partners to Furthermore, operators of satellite systems for counterbalance project implementation costs, which monitoring climate need to: ideally should reach 100 percent by the time the project is completed within the funding cycle. • Take steps to make radiance calibration, calibration-monitoring and satellite-to-satellite 7.7 IGOS Monitoring Principles cross-calibration of the full operational constellation a part of the operational satellite • The impact of new systems or changes to system; and existing systems should be assessed prior to • Take steps to sample the Earth system in such a implementation; way that climate-relevant (diurnal, seasonal, and • A suitable period of overlap for new and old long-term inter-annual) changes can be resolved. observing systems is required; • The details and history of local conditions, Thus satellite systems for climate monitoring should instruments, operating procedures, data adhere to the following specific principles: processing algorithms, and other factors pertinent to interpreting data (that is, metadata) should be • Constant sampling within the diurnal cycle documented and treated with the same care as (minimizing the effects of orbital decay and orbit the data themselves; drift) should be maintained; • The quality and homogeneity of data should be • A suitable period of overlap for new and old regularly assessed as a part of routine operations; satellite systems should be ensured for a period • Consideration of the needs for environmental and adequate to determine inter-satellite biases and climate-monitoring products and assessments, maintain the homogeneity and consistency of such as IPCC assessments, should be integrated time-series observations; into national, regional and global observing • Continuity of satellite measurements (for example, priorities; elimination of gaps in the long-term record) • Operation of historically-uninterrupted stations through appropriate launch and orbital strategies and observing systems should be maintained; should be ensured; • High priority for additional observations should • Rigorous pre-launch instrument characterization be focused on data-poor regions, poorly and calibration, including radiance confirmation observed parameters, regions sensitive to change, against an international radiance scale provided and key measurements with inadequate temporal by a national metrology institute, should be resolution; ensured; • Long-term requirements, including appropriate • On-board calibration adequate for climate sampling frequencies, should be specified to system observations should be ensured and network designers, operators and instrument associated instrument characteristics monitored; 121 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains • Operational production of priority climate year, pragmatically. Typical maintenance jobs products should be sustained and peer- are cleaning, recalibration (on site), painting, reviewed new products should be introduced as dehumidifying, changing batteries, replacing appropriate; consumables, checking functionality, and • Data systems needed to facilitate user access retrieving backup data from loggers; to climate products, metadata and raw data, – Maintenance costs: Generally, 25–35 including key data for delayed-mode analysis, percent of the capital investment in should be established and maintained; instruments over the standing time of the • Use of functioning baseline instruments that instruments, which is in the order of 8–10 meet the calibration and stability requirements years before replacement is necessary. Some stated above should be maintained for as instruments last longer, depending on the long as possible, even when these exist on manufacturer and operating conditions; and decommissioned satellites; – Emergency contingency to cover loss of • Complementary in situ baseline observations for sensors, destruction of stations (for example satellite measurements should be maintained due to flood or avalanche): In the order of through appropriate activities and cooperation; and 10 percent of the capital investment costs. • Random errors and time-dependent biases in satellite observations and derived products should 7.8.2 Selection of Location be identified. As the network will continue to be sparse in such high-altitude environments, the key criteria for the 7.8 General Considerations selection of locations are: 7.8.1 Costs of Field Trips • Stations are situated in glacierized basins; • Altitudes are above 3,500 m; These are highly variable from country to country • Station location is representative of a larger and by location, and need to be assessed during area or comparable to other stations in similar country consultations. The duration of the visit altitudes; no stations are placed in locations with includes access by car or plane to the nearest a specific microclimate that bears no similarity to accessible point, then walking with porters. other station locations; • Locations are preferably in the glacierized • Expedition-style campaigns, for example, for headwaters of streams that have importance for the purpose of process studies or full-scale water management (for example, hydropower, assessment of local meteorological, hydrological, irrigation, water supply); and glaciological conditions, need to be costed • Local observers can be found, which may require separately. Also, the costs of establishing a the establishment of solid station houses where new station (including civil works) need to be the observer can stay. They may then be allowed estimated, including the following items: to run the station house as a small lodge for – Duration of establishment: Four weeks for tourists in order to earn some extra income to station house and station; maintain the house and to act as an incentive for – Construction costs: Building materials, the observers to stay there; porters, labor, food on site; • Stations have year-round accessibility, with the – Maintenance requirements: As per exception of glacier stations that work as satellites manufacturer’s specifications, mostly to the main stations; and performed during regular visits to the • Telecommunication (GPRS, GSM, HF radio) is stations, which should occur three times a possible. 122 Monitoring of the HKH Cryosphere Reference Haeberli, W., J. Cihlar, and R.G. Barry. 2000. “Glacier Monitoring within the Global Climate Schaefli, B., and M. Huss. 2011. “Integrating Observing System.” Annals of Glaciology 31: Point Glacier Mass Balance Observations into 241–46. Hydrologic Model Identification.” Hydrology and Haeberli, W., M. Hoelzle, F. Paul, and M. Zemp. Earth System Sciences 15: 1227–41. 2007. “Integrated Monitoring of Mountain Glaciers as Key Indicators of Global Climate Recommended Reading on Monitoring Change: The European Alps.” Annals of Glaciology 46: 150–60. General Integrated Global Observing Strategy (IGOS). 2007. Tartari, G. 2009. High Altitude IGOS Cryosphere Theme for the Monitoring of Environmental Monitoring: The SHARE Project Our Environment from Space and from Earth. and CEOP-HE. Paper presented at European Geosciences Union General Assembly, Vienna, Guidelines and Standards Relating to Austria, April 19–24, 2009. Measurement of Glacier Fluctuations Glacier Monitoring Anonymous. 1969. “Mass-Balance Terms.” Journal of Glaciology 8 (52): 3–7. Zemp, M. 2011. The Monitoring of Glaciers at Local, Mountain, and Global Scale. Habilitationsschrift Forel, F.A. 1895. “Instructions pour l’Observation zur Erlangung der Venia Legendi der des Variations des Glaciers: Discours Mathematisch-naturwissenschaftlichen Fakultät Préliminaire.” Archives des Sciences Physiques et der Universität, Mai 2011. Naturelles XXXIV: 209–29. Guidelines and Standards Relating to the Kaser, G., A. Fountain, and P. Jansson. 2003. A International Glacier Monitoring Strategy Manual for Monitoring the Mass Balance of Global Climate Observing System (GCOS). 2010. Mountain Glaciers with Particular Attention to Systematic Observation Requirements for Satellite- Low Latitude Characteristics: A Contribution from Based Products for Climate: Supplemental the International Commission on Snow and Ice Details to the Satellite-Based Component of the (ICSI) to the UNESCO HKH-Friend Program. IHP- Implementation Plan for the Global Observing VI Technical Documents in Hydrology 59. Paris: System for Climate in Support of the UNFCCC. UNESCO. Report GCOS-107, WMO/TD 1338, updated in Kasser, P., ed. 1967. Fluctuations of Glaciers 1959– 2010 (GCOS-138). 1965, Volume I. Zurich, Switzerland: Permanent Haeberli, W. 1998. “Historical Evolution and Service on Fluctuations of Glaciers. Operational Aspects of Worldwide Glacier Østrem, G., and A. Stanley. 1969. Glacier Mass Monitoring.” In Into the Second Century of World Balance Measurements: A Manual for Field and Glacier Monitoring: Prospects and Strategies, ed. Office Work. Canadian Department of Energy, W. Haeberli, M. Hoelzle, and S. Suter, 35–51. Mines and Resources and Norwegian Water Paris: UNESCO. Resources and Electricity Board. Haeberli, W. 2006. “Glaciers and Ice Caps: Patterson, W.S.B. 1969. The Physics of Glaciers. Historical Background and Strategies of Oxford: Pergamon Press. Worldwide Monitoring.” In Mass Balance of the Cryosphere:
Observations and Modelling United Nations Educational, Scientific and of Contemporary and Future Changes, ed. J.L. Cultural Organization (UNESCO). 1970/1973. Bamber and A.J. Payne, chapter 15. Combined Heat, Ice and Water Balances at 123 Monitoring of Glaciers, Climate, and Runoff in the Hindu Kush-Himalaya Mountains Selected Glacier Basins. Part I: A Guide for and Ice Resources, 314–28. Washington, DC: Compilation and Assemblage of Data for National Academy of Sciences. Glacier Mass Balance Measurements. Part II: United States Army Corps of Engineers. 1955. Specifications, Standards and Data Exchange. Snow Hydrology: Summary of Report of Snow UNESCO/IAHS Technical Papers in Hydrology 5. Investigations. World Glacier Monitoring Service (WGMS). 2009, Climate Monitoring and earlier versions. Submission of Glacier Fluctuation Data to the World Glacier Monitoring Hauer, F., J. Baron, D. Campbell, K. Fausch, S. Service: General Guidelines and Attribute Hostetler, G. Leavesley, P. Leavitt, D. Descriptions. Zurich, Switzerland: WGMS. McKnight, and J. Stanford. 1997. “Assessment of Climate Change and Freshwater Ecosystems Snow Monitoring of the Rocky Mountains, USA and Canada.” Crook, A.G. 1985. “SNOTEL Data Acquisition Hydrological Processes 11 (8): 817–1067. System: A Tool in Runoff Forecasting.” In A Karl, T.R., V.E. Derr, D.R. Easterling, C.K. Folland, Critical Assessment of Forecasting in Water D.J. Hofmann, S. Levitus, N. Nicholls, D.E. Parker, Quality Goals in Western Water Resources and G.W. Withee. 1995. “Critical Issues for Long- Management: Proceedings of a Symposium Held Term Climate Monitoring.” Climate Change 31 in Seattle, Washington, June 11­–13. (2–4) 85–221. Elder, K., J. Dozier, and J. Michelsen. 1991. “Snow Malby, A., J. Whyatt, R. Timmis, R. Wilby, and Accumulation and Distribution in an Alpine H. Orr. 2007. “Long-Term Variations in Watershed.” Water Resources Research 27 (7): Orographic Rainfall: Analysis and Implications 1541–52. for Upland Catchments.” Hydrological Sciences Elder, K., W. Rosenthal, and R. Davis. 1998. Journal 52 (2): 276–91. “Estimating the Spatial Distribution of Snow Wood, F. 1990. Water Equivalence in a Montane Watershed.” “Monitoring Global Climate Change: The Hydrological Processes 12: 1791–1808. Case of Greenhouse Warming.” Bulletin of the Farnes, P. 1971, “Mountain Precipitation and American Meteorological Society 71 (1): 42–144. Hydrology from Snow Surveys.” In Proceedings Hydrological Monitoring of 39th Annual Western Snow Conference, April Dunne, T., and L. Leopold. 1978. Water in 20–22, Billings, Montana, 44–49.  Environmental Planning. New York: W.H. Freeman Gurung, D.R., A. Giriraj, K.S. Aung, B. Shrestha, and Co. and A.V. Kulkarni. 2011. Snow-Cover Mapping Klemes, V. 1983. “Conceptualization and Scale in and Monitoring in the Hindu Kush-Himalayas. Hydrology.” Journal of Hydrology 65: 1–23. Kathmandu: ICIMOD. Klemes, V. 1990. “The Modeling of Mountain National Weather Service: National Operational Hydrology: The Ultimate Challenge.” In Hydrologic Remote Sensing Center. 2013. Hydrology of Mountainous Areas, IAHS NOHRSC Science and Technology. http://www. Publication 190. nohrsc.noaa.gov/technology/. Loucks, D.P., E. van Beek, J.R. Stedinger, J.P.M. Steppuhn, H., and G.E. Dyck. 1974. “Estimating Dijkman, and M.T. Villars. 2005. Water True Basin Snow Cover.” In Proceedings of Resources Systems Planning and Management: Interdisciplinary Symposium on Advanced An Introduction to Methods, Models and Concepts and Techniques in the Study of Snow Applications. Paris: UNESCO. 124 Notes Notes