Climate Risks and A SYNTHESIS Adaptation in Asian REPORT Coastal Megacities Climate Risks and Adaptation in Asian Coastal Megacities A Synthesis Report © 2010 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, NW Washington, DC 20433, U.S.A. Telephone: 202-473-1000 Internet: www.worldbank.org/climatechange E-mail: feedback@worldbank.org All rights reserved. September 2010 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 volume 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 judgement on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applica- ble law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400; fax 978-750-4470; Internet: www.copyright.com. Cover images: Large image: Ho Chi Minh City, © Karen Kasmauski/Corbis Small images: top: Manila, © Francis R. Malasig/Corbis; middle: Bangkok, © I. Saxar/Shutterstock Images, LLC; bottom: Kolkata, © Bruce Burkhardt/ Corbis All dollars are U.S. dollars unless otherwise indicated. Table of Contents Acknowledgments...................................................................................................................................vii Abbreviations.and.Acronyms................................................................................................................. ix Executive.Summary.................................................................................................................................. xi 1.. Introduction..........................................................................................................................................1 Background and Rationale ............................................................................................................................ 1 Objective........................................................................................................................................................... 2 Process of Preparation.................................................................................................................................... 3 Overview of Methodology/Approach and Climate Parameters Selected ............................................. 3 Structure of the Report ................................................................................................................................... 4 2.. Methodologies.for.Downscaling,.Hydrological.Mapping,.and.. . Assessing.Damage.Costs...................................................................................................................5 Selection of Emissions Scenarios, Downscaling, and Uncertainties ...................................................... 5 Hydrological Modeling for Developing Scenarios of Flood Risk ........................................................... 9 Approach to Assessing Damage Costs ...................................................................................................... 12 Assessment of Damage Costs in the HCMC Study ................................................................................ 17 Assumptions about the Future of Cities in Estimating Damage Costs ............................................... 19 Conclusion: Methodological Limitations and Uncertainties in Interpreting Results of the Study ............................................................................................................................................. 20 3.. Estimating.Flood.Impacts.and.Vulnerabilities.in.Coastal.Cities.............................................23 Estimating Future Climate-related Impacts in Bangkok......................................................................... 23 Main Findings from Hydrological Analysis and GIS Mapping for Bangkok ..................................... 28 Estimating Climate-related Impacts in Manila ....................................................................................... 31 Findings from the Hydrological Analysis and GIS Mapping for Metro Manila ................................. 35 Estimating Climate-related Impacts in Ho Chi Minh City, Vietnam .................................................... 38 Main Findings from Hydrological Analysis and GIS Mapping for HCMC ......................................... 44 Conclusion ..................................................................................................................................................... 50 4.. Assessing.Damage.Costs.and.Prioritizing.Adaptation.Options..............................................51 Bangkok: Analysis of Damage Costs Related to Flooding in 2008 and 2050 ....................................... 51 Prioritization of Adaptation Options in Bangkok.................................................................................... 56 Analysis of Damage Costs Related to Flooding in Metro Manila ........................................................ 60 Prioritization of Adaptation Options in Manila ....................................................................................... 65 Analysis of Damage Costs in HCMC ........................................................................................................ 69 Analysis of Adaptation in HCMC .............................................................................................................. 72 Conclusion ..................................................................................................................................................... 73 iii 5.. Conclusions.and.Policy.Implications............................................................................................75 Key Findings and Lessons for Policy Makers........................................................................................... 75 Lessons on Methodology for Follow-up Studies .................................................................................... 78 Bibliography..............................................................................................................................................81 Annexes A. Vulnerability of Kolkata Metropolitan Area to increased Precipitation in a Changing Climate ................................................................................................................................ 85 B. Scenarios Applied in the Hydrodynamic Modeling in the HCMC study...................................... 91 C. Adaptation to Increased Flooding: Brief Overview........................................................................... 93 D. Comparison of Costs across Cities ....................................................................................................... 97 Figures Figure 1.1 Asian Megacity Hotspots ................................................................................................................. 2 Figure 2.1 Hydrometeorological Model Schematic for Chao Phraya Watershed ......................................11 Figure 2.2 Manila Rainfall-Runoff Calibration Hydrographs ..................................................................... 12 Figure 2.3 Estimation of Damage to Buildings, Assets, and Inventories in the Bangkok and Manila Cases............................................................................................................................. 14 Figure 2.3 Estimating Impacts--A Flow Chart .............................................................................................. 13 Figure 2.5 Possible Relationships between Flood Duration and Land Value Loss .................................. 18 Figure 3.1 Location of Bangkok in the Chao Praya River Basin.................................................................. 23 Figure 3.2 Land Elevations, 2002 versus 2050 Land Subsidence................................................................. 28 Figure 3.3 Maximum Water Depth for 1-in-30-year event, 2008 and 2050, A1FI...................................... 29 Figure 3.4 Bangkok Flood Hazard Relationship............................................................................................ 29 Figure 3.5a Affected Condensed (Poor) Community of Case C2008-T30 ................................................... 31 Figure 3.5b Affected Condensed (Poor) Community of Case C2050-LS-SR-SS-A1FI-T30 ........................ 31 Figure 3.6 Metro Manila and its Watershed ................................................................................................... 32 Figure 3.7 Different Climatic Regimes in the Philippines ............................................................................ 33 Figure 3.8 Major Watershed and Drainage Areas of Manila........................................................................ 34 Figure 3.9 Comparison of Population Affected by Flooding under Different Scenarios ........................ 36 Figure 3.10 Areas of High Population Density and with High Risk of Inundation under A1FI Scenario ................................................................................................................................... 37 Figure 3.11 Areas at High Risk from Flooding under Different Scenarios .................................................. 37 Figure 3.12 HCMC: Frequently Flooded Areas under Current Conditions ................................................ 40 Figure 3.13a HCMCCity Case Study: Comparison of 1-in-30-year Flood for 2008 ...................................... 45 Figure 3.13b HCMCCity Case Study: Comparison of 1-in-30-year Flood for 2050 A2 Scenario ................ 45 Figure 3.14a HCMC Poverty Rates by District .................................................................................................. 48 Figure 3.14b Districts Vulnerable to Flooding.................................................................................................... 48 Figure 3.15 Impact on Waste Management Sector .......................................................................................... 49 Figure 3.16 HCMC 2050 A2 1-in-30-year Flood Inundation Overlaid on Projected Land Use Patterns ...................................................................................................................................... 49 Figure 3.17 HCMC Droughts and Salinity Intrusion in 2050 ........................................................................ 50 Figure 4.1 Damage Cost Associated with a 1-in-30-year Flood (C2050-LS-SR-SS-A1FI-T30)................. 52 Figure 4.2 Loss Exceedance Curves, Bangkok .............................................................................................. 52 Figure 4.3 Maximum Inundation Area Without and With the Proposed Adaptation ............................. 58 Figure 4.4 Flood Costs under Three Return Periods and Two Climate Scenarios (PHP) ........................ 60 Figure 4.5 Loss Exceedance Curves for Manila (PHP) ................................................................................. 63 iv | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Figure 4.6 Damage Costs Associated with Different Scenarios (PHP) ...................................................... 63 Figure 4.7 Damages to Buildings from a 1-in-30-year Flood (2008 PHP) .................................................. 63 Figure 4.8 Flood Costs as a Percent of 2008 GDP .......................................................................................... 65 Figure 4.9 Annual Benefits from Adaptation Investments in Metro Manila ............................................. 68 Boxes Box 2.1 Strengths and Limitations of Different Downscaling Techniques Selected for this Study ......... 7 Box 2.2 Downscaling from 16 GCMs ................................................................................................................ 8 Box 2.3 Some Basic Principles for Hydrological Mapping ...........................................................................11 Box 3.1 The Bangkok Metropolitan Region (BMR): Some Assumptions about the Future .................... 26 Box 3.2 What does Metro Manila Look Like in the Future? ....................................................................... 35 Box 3.3 HCMC in 2050 ..................................................................................................................................... 41 Box 3.4 Overview of Downscaling and Hydrological Analysis Carried out for HCMC Study ............ 42 Box 4.1 Examining Building Damages, Income Losses, and Health Costs in Bangkok ......................... 55 Box 4.2 Expected Annual Benefits from Adaptation in Bangkok ............................................................... 58 Box 4.3 Increased Time Costs and Health Risk from Flooding in Manila ................................................ 65 Box 4.4 Rough Estimate of Viability of Proposed Flood Control Measures ............................................. 72 Tables Table 2.1 Climate Change Forecasts for 2050 ..................................................................................................... 8 Table 2.2 Summary of City Case Study Hydrologic Modeling ......................................................................11 Table 2.3 Direct and Indirect Costs from Flooding ......................................................................................... 13 Table 2.4 Flood Damage Rate by Type of Building in Manila ....................................................................... 15 Table 3.1 Poverty Line and the Poor in the BMR1........................................................................................... 24 Table 3.2 Bangkok Monthly Average Temperature and Precipitation.......................................................... 25 Table 3.3 Climate Change and Land Subsidence Parameter Summary for Bangkok ................................ 27 Table 3.4 Bangkok Inundated Area under Current Conditions and Future Scenarios .............................. 28 Table 3.5 Exposure of Bangkok Population to Flooding ................................................................................ 30 Table 3.6 Manila: Monthly Average Temperature and Precipitation ............................................................ 33 Table 3.7 Manila Climate Change Parameters ................................................................................................. 35 Table 3.8 Manila: Comparison of Inundated Area (km2) with 1-in-100-year flood for 2008 and 2050 Climate Change Scenarios with only Existing Infrastructure and with Completion of 1990 Master Plan .............................................................................................. 36 Table 3.9 Affected Length of Road by Inundation Depth .............................................................................. 38 Table 3.10 HCMC District Poverty Rates. 2003.................................................................................................. 39 Table 3.11 Ho Chi Minh City: Monthly Average Temperature and Precipitation ........................................ 40 Table 3.12 Climate Change Parameter Summary for HCMC ......................................................................... 44 Table 3.13 Summary of Flooding at Present and in 2050 with Climate Change ........................................... 44 Table 3.14 District Population Affected by an Extreme Event in 2050 ........................................................... 46 Table 3.15 Districts Affected by Flooding in Base Year and in 2050 ............................................................... 47 Table 3.16 Effects of Flooding on Future Land Use under 2050 A2 Extreme Event .................................... 49 Table 4.1 Summary of Damages Assessed in the Bangkok Study ................................................................ 51 Table 4.2 Summary of Flood and Storm Damages, Bangkok (million 2008 THB) ...................................... 53 Table 4.3 Changes in Income Losses to Wage Earners, Commerce, and Industry ..................................... 55 Table 4.4 Damage Costs in Bangkok and Regional GRDP ............................................................................. 56 Table 4.5 Investment Costs for Adaptation Projects in Bangkok (million THB)......................................... 57 Table 4.6 Flood Damage Costs With and Without a 30-year Return Period Flood Protection Project (million THB) .......................................................................................................................... 59 TableofContents | v Table 4.7 Net Present Value of Adaptation Measures to Provide Protection Against a 1-in-30 and 1-in-10-year Flood (million THB) .............................................................................. 59 Table 4.8 Flood Damage Costs in Manila (2008 PHP) .................................................................................... 61 Table 4.9 Income and Revenue Losses to Individuals and Firms Associated with Floods (2008 PHP) ............................................................................................................................................ 64 Table 4.10 Damage Costs from 1-in-10, 1-in-30, and 1-in-100-year Floods in Different Scenarios (2008 PHP) ............................................................................................................................................ 65 Table 4.11 Adaptation Investments Considered for Different Return Periods and Climate Scenarios ................................................................................................................................ 67 Table 4.12 Investment Costs and Net Present Value of Benefits Associated with Different Flood Control Projects in Manila (PHP) using a 15 percent discount rate.................................. 67 Table 4.13 Expected Cost of Flooding based on Quadratic Relationship between Duration of Flooding and Land Values in HCMC........................................................................................... 70 Table 4.14 Present Value of the Cost of Floods up to 2050 using the GDP Estimation Method ................. 71 Table 4.15 Summary of Present Value of Climate Change Costs in HCMC (USD) ...................................... 72 Table 4.16 Proposed Implementation Arrangements for HCMC.................................................................... 74 vi | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Acknowledgments T his synthesis report is a product of a joint pro- the Metro Manila Development Authority, National gram on Climate Adaptation in Asian Coastal Statistics Office, Professor Emma Porio and staff Megacities undertaken by the World Bank in at Ateneo de Manila University, CTI Engineering collaboration with the Asian Development Bank International, and ALMEC Corporation, as well and the Japan International Cooperation Agency. It as the local government officials and community is based on extensive collaboration among the three leaders who provided valuable inputs to the report. agencies, who jointly agreed to undertake several We also gratefully acknowledge the contributions city-level studies and prepare a synthesis report. of Professor Akimasa Sumi (University of Tokyo), The core team preparing this synthesis report Professor Nobuo Mimura (Ibaraki Universty), consisted of Poonam Pillai (Sr. Environmental Spe- and Dr. Masahiro Sugiyama (Central Research cialist and task team leader), Bradford Ryan Philips Institute of Electric Power Industry) for providing (Sr. Civil Engineer, consultant), Priya Shyamsundar the analytic framework for downscaling the IPCC (Sr. Environmental Economist, consultant), Kazi climate models. We thank the Kolkata team and Ahmed (consultant) and Limin Wang (Sr. Envi- in particular Subhendu Roy and the INRM team ronmental Economist, consultant) and included for their inputs and collaboration, and to Adriana extensive collaboration with the different city-level Damianova for initially leading the Kolkata study. teams. In particular, we would like to thank Jan Suggestions from Ian Noble also helped strengthen Bojo (World Bank), who led the Bangkok study; the analysis. We are especially grateful to Daniel Megumi Muto (JICA), who led the Manila study Hoornweg, Anthony Bigio, and Tapas Paul for peer and was the main focal point from JICA; Jay Roop reviewing this report. (ADB), who led the Ho Chi Minh City study and was A special thanks to James Warren Evans (Direc- the main focal point from ADB, and Maria Sarraf tor, Environment Department, World Bank), Magda and Susmita Dasgupta (World Bank), who led the Lovei, (Sector Manager, EASER, World Bank), Neeraj Kolkata study. For the Bangkok report, we are also Prasad (Lead Carbon Finance Specialist, ENVCF), grateful to the team at Panya consultants; to Bang- Megumi Muto (Research Fellow, JICA) and Jay Roop kok Metropolitan Administration professionals; (Environmental Specialist, ADB) for initiating this and to Manuel Cocco, Pongtip Puvacharoen, and collaborative activity and to Kseniya Lvovsky (Pro- Yabei Zhang. For the HCMC study, we thank the gram Manager, Climate Change team, Environment consulting team at the International Centre for En- Department), Michele De Nevers (Senior Manager, vironmental Management, including Jeremy Carew- Environment Department, World Bank), Gajanand Reid, Anond Snidvongs, Peter-John Meynell, John Pathmanathan (Manager, SASDO and Acting Sector Edmund Sawdon, Nigel Peter Hayball, Tran Thi Ut, Manager, SASDI), Nessim Ahmad (Director, Environ- Tranh Thanh Cong, Nguyen Thi Nga, Nguyen Le ment and Safeguards, Asian Development Bank), Dr Ninh, Nguyen Huu Nhan, and Nguyen Dinh Tho; Keiichi Tsunekawa (Director, JICA Research Institute) the Ho Chi Minh City People's Committee; and the and Mr. Hiroto Arakawa (Senior Special Advisor, Department of Natural Resources and Environment JICA) under whose general guidance this report (DoNRE). For the Manila study, we are grateful to was prepared. Thanks to Perpetual Boetang for her vii assistance with formatting the report, to Robert Liv- cial support for the preparation of this report through ernash for editing, and to Jim Cantrell for managing the Trust Fund for Environmentally and Socially production of the publication. Finally, we thank the Sustainable Development and the Norwegian Trust governments of Norway and Finland for their finan- Fund for Private Sector and Infrastructure. viii | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Abbreviations and Acronyms ADB Asian Development Bank IRS3 Integrated research system for AOGCM Atmosphere-ocean general circulation sustainability science models IZ Industrial zones BMR Bangkok Metropolitan Region JICA Japan International Cooperation BAU Business as usual Agency CCA Climate change adaptation LGUs Local government units CoP Conferences of Parties MONRE Ministry of Natural Resources and DIVA Dynamic interactive vulnerability Environment assessment MP Master plan DRR Disaster risk reduction NESDB National Economic and Social ECLAC Economic Commission for Latin Development Board America and the Caribbean PC People's Committee GCMs Global climate models PCMDI Program for Climate Model Diagnosis GDP Gross domestic product and Intercomparison GEF Global Environment Facility SRES Special Report on Emissions Scenarios GHG Greenhouse gas UNDP United Nations Development Program GPCP Global Precipitation Climatology UNFCCC United Nations Framework Project Convention on Climate Change GRDP Gross regional domestic product VOC Vehicle operations cost HCMC Ho Chi Minh City WCRP World Climate Research Program 1DD One-degree daily WGCM Working Group on Coupled Modeling IPCC Intergovernmental Panel for Climate Change Note: Unless otherwise noted, all dollars are U.S. dollars. ix Executive Summary InTRoduCTIonAndRATIonAle study on the economics of adaptation to climate change, which estimates that the cost of adaptation Coastal areas in both developing and more industri- to climate change is likely to be the highest in this alized economies face a range of risks related to cli- region (World Bank 2010). In flood-prone cities such mate change and variability (IPCC 2007a). Potential as Ho Chi Minh City, Kolkata, Dhaka, and Manila, risks include accelerated sea level rise, increase in sea potential sea level rise and increased frequency and surface temperatures, intensification of tropical and intensity of extreme weather events poses enormous extra tropical cyclones, extreme waves and storm adaptation challenges. The urban poor--often living surges, altered precipitation and runoff, and ocean in riskier urban environments such as floodplains or acidification (Nicholls et al. 2007). The Intergovern- unstable slopes, working in the informal economy, mental Panel for Climate Change Fourth Assessment and with fewer assets--are most at risk from expo- Report (IPCC 2007a) points to a range of outcomes sure to hazards (Satterthwaite et al. 2007). under different scenarios. It identifies a number of Despite its importance, few developing country hotspots--including heavily urbanized areas situ- cities have attempted to address climate change sys- ated in the low-lying deltas of Asia and Africa--as tematically as part of their decision-making process. especially vulnerable to climate-related impacts. Given the risks faced by coastal cities and the impor- The number of major cities located near coast- tance of cities more broadly as drivers of regional lines, rivers, and deltas provides an indication of the economic growth, adaptation must become a core population and assets at risk. Thirteen of the world's element of long-term urban planning. The Mayor's 20 largest cities are located on the coast, and more Summit in Copenhagen in December 2009--and than a third of the world's people live within 100 follow-on efforts to institutionalize a Mayor's Task miles of a shoreline. Low-lying coastal areas repre- Force on Urban Poverty and Climate Change--sig- sent 2 percent of the world's land area, but contain nify much-needed attention to this issue. 13 percent of the urban population (McGranahan et In response to client demand and recognizing al. 2007). A recent study of 136 port cities showed the importance of addressing urban adaptation that much of the increase in exposure of population and major vulnerabilities of Asian coastal cities, the and assets to coastal flooding is likely to be in cities Asian Development Bank (ADB), the Japan Interna- in developing countries, especially in East and South tional Cooperation Agency (JICA), and the World Asia (Nicholls et al. 2008). Bank agreed to undertake an analysis in several In terms of population exposed to coastal flood- coastal megacities to address climate adaptation ing, for example, in 2005 five of the ten most popu- and prepare a synthesis report based on the city- lous cities included Mumbai, Guangzhou, Shanghai, level findings. The selected cities included Manila Ho Chi Minh City, and Kolkata (formerly Calcutta). (led by JICA), Ho Chi Minh City (led by ADB), and By 2070, nine of the top ten cities in terms of popula- Bangkok (led by the World Bank).1 tion exposure are expected to be in Asian developing 1 Kolkata is also one of the selected cities but is not included countries (Nicholls et al. 2008). The vulnerability of in the synthesis report as it was ongoing at the time of the prep- the East Asia region is also highlighted by the global aration of this report. A brief overview is included in Annex A. xi Why these three cities? The three developing using case studies of three cities that are different country cities selected for this study are all coastal in their climate, hydrological, and socioeconomic megacities with populations (official and unofficial) characteristics. Specifically, it draws on an in-depth ranging from 8 to 15 million people. Two are capital analysis of climate risks and impacts in Bangkok, cities and all three are centers of national and regional Manila, and Ho Chi Minh City to highlight to na- economic growth contributing substantially to the tional and municipal decision makers (a) the scale GDP of the respective countries. However, being low- of climate-related impacts and vulnerabilities at lying coastal cities situated in the deltas of major river the city level, (b) estimates of associated damage systems in the East Asia region, all three are highly costs, and (c) potential adaptation options. While vulnerable to climate-related risks and rank high in the report focuses on three cities in East Asia, the recent rankings of exposure and vulnerability. Ho policy implications resulting from the comparative Chi Minh City and Bangkok are among the top 10 analysis of these cities has broader relevance for cities in terms of population likely to be exposed to assessing climate risks and identifying adaptation coastal flooding due to climate-related risks in 2070, options in other coastal areas. according to the first global assessment of port cities (Nicholls et al. 2008). Further, Manila has been identi- fied as particularly vulnerable to typhoon damage, AppRoAChAndMeThodology and HCMC ranks fifth by population exposed to The approach to assessing climate risks and im- the effects of climate change (Nicholls et al. 2008). A pacts consists of the following sequential steps: (1) recent study also identifies Manila, Ho Chi Minh City, determining climate variables at the level of the and Bangkok among the top eleven Asian megacities city/watershed through downscaling techniques; that are most vulnerable to climate change (Yusuf (2) estimating impacts and vulnerability through and Francisco 2009).2 Devastating floods in Manila hydrometeorological modeling, scenario analysis, in 2009 only confirm the vulnerability of this city to and GIS mapping; and (3) preparing a damage/ extreme weather events. For instance, flooding in loss assessment and identification/prioritization Manila from tropical storm Ketsana in September of adaptation options. was the heaviest in almost 40 years, with flood waters As a first step, each of the city-level studies reaching nearly 7 meters. More than 80 percent of considered two IPCC scenarios, a high- and a low- the city was underwater, causing immense damage emissions scenario,4 and estimated climate risks to housing and infrastructure and displacing around to 2050. The 2050 time horizon for the study is ap- 280,000­300,000 people.3 All of this highlights the propriate given city-level planning horizons and need to better understand and prepare for such cli- the typical time frame for major flood protection mate risks and incorporate appropriate adaptation measures. The downscaling analysis allowed esti- measures into urban planning. mation of changes in temperature and precipitation While there is a growing literature on cities and in 2050. These parameters were used as inputs to the climate change, as yet there is limited research on hydrological modeling. In addition to this, assump- systematically assessing climate-related risks at the tions and estimates were also made about changes city level. This report aims to fill this gap. Further, in sea level rise and storm surge in 2050 based on it aims to provide evidence-based information to past historical data and available estimates. support urban policy and planning as these issues are debated at the local, national, and global levels. 2 Vulnerability in the scorecard was understood in terms of exposure, sensitivity, and adaptive capacity of objeCTIve the cities. See also http://www.idrc.ca/uploads/user- S/12324196651Mapping_Report.pdf. 3 http://edition.cnn.com/2009/WORLD/asiapcf/09/27/ The main objective of this report is to strengthen philippines.floods/index.html. our understanding of climate-related risks and im- 4 Different scenarios were considered to assess impact due pacts in coastal megacities in developing countries to the uncertainties in projecting future climate conditions. xii | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport For each city, complex hydrometeorological version of the study has been presented at several models were then developed using a whole host international forums. of local information. These included (a) climate variables such as changes in temperature, precipi- tation, sea level rise, and storm surge; (b) socio- unCeRTAInTIeS,lIMITATIonS, economic and developmental factors such as land AndInTeRpReTIngThe subsidence, land use, and population increases; and (c) local topographical and hydrological FIndIngSoFThISSTudy information. Flooding in the metropolitan areas Any study forecasting conditions four decades was chosen as the key variable to assess impact. hence will be faced with large uncertainties and The hydrological analysis allowed determination these need to be borne in mind in interpreting the of the area, depth, and duration of flooding under results of this study. One uncertainty concerns different scenarios. This information was used to the pathway of GHG emissions. To address that identify the scale of risks and vulnerability of sec- issue, the city case studies examined both a high tors, local populations, and districts (represented and a low GHG emissions scenario to bracket the in GIS maps), as well as estimate damage costs. likely future conditions. In the climate change Two of the three studies undertook cost-benefit downscaling methodologies, there are uncertain- analysis to prioritize adaptation options, while ties in forecasting the increase in extreme and the third approached the issue of adaptation more seasonal precipitation under the different sce- qualitatively. To understand the impact of climate narios. The techniques applied in the statistical change in 2050 in each city, an important assump- downscaling examined the results from sixteen tion made by all teams was that without climate atmosphere-ocean general circulation models change, the climate in 2050 would be similar to the (AOGCM). Robust relationships were identified 2008/ base-year climate. Various climate scenarios for temperature (with a ~ 10 percent internal are overlaid on this assumption. error) and precipitable water increases (with a ~ 10­20 percent error) (Sugiyama 2008). Hydrologic models can simulate flood events with relatively pRoCeSSoFpRepARATIon small errors (<10 percent) if sufficient data are The analysis was carried out over a period of one- available for good calibration. For future forecasts, and-a-half years. The synthesis team and the city- however, land use changes in the watersheds level teams met periodically and worked closely to and drainage areas can dramatically affect flood develop common terms of reference to guide the patterns and can be further examined in future city-level studies, as well as share methodological sensitivity analyses. issues and ongoing findings. These discussions and Further, cities in 2050 are likely to be vastly the analysis undertaken for each city have formed different from today's cities. Understanding how the basis of this report. Further, each city-level different is a huge task and there was no attempt team worked with their respective country/urban to model economic growth and link it to urban counterparts to build ownership and capacity for development. Instead, assumptions about cities in the analysis. For instance, the main counterparts in 2050 were based on best available data, government HCMC were the HCMC People's Committee and plans and projections which also introduced uncer- the Department of Natural Resources and Environ- tainties and errors. Despite these limitations, the ment (DoNRE). The study sought to inform the results presented in this report highlight the scale of preparation of HCMC's citywide adaptation plan. the likely risks and impacts facing coastal cities that In Bangkok, the main counterpart was the Bangkok appear to be robust to the assumptions about the Municipal Authority. In Metro Manila, the main climatic, spatial, and socioeconomic development counterpart was Metro Manila Development Au- of the cities by 2050. Key findings and lessons are thority (MMDA). At the global level, a preliminary summarized below. executiveSummary | xiii KeyFIndIngS City, and San Juan Mandaluyong City are likely to face serious risks of flooding. Frequencyofextremeeventslikelytoincrease All three cities are likely to witness increases in Increaseinpopulationexposedtoflooding temperature and precipitation linked with climate In all three cities, there is likely to be an increase in change and variability. In Bangkok, temperature the number of persons exposed to flooding in 2050 increases of 1.9° C and 1.2°C for the high and low under different climate scenarios compared to a situa- emissions scenarios respectively are estimated for tion without climate change. For instance, in Bangkok 2050 and are linked with a 3 percent and 2 percent in 2050, the number of persons affected (flooded increase in mean seasonal precipitation respectively. for more than 30 days) by a 1-in-30-year event will In Manila, the mean seasonal precipitation is ex- rise sharply for both the low and high emission sce- pected to increase by 4 percent and 2.6 percent for narios--by 47 percent and 75 percent respectively-- the high and low emissions scenarios. In HCMC, compared to those affected by floods in a situation future projections suggest greater seasonal variabil- without climate change. In Manila, for a 1-in-100-year ity in rainfall and increasing frequency of extreme flood in 2050, under the high emission scenario more rainfall related to storms. than 2.5 million people are likely to be affected (as- suming that the infrastructure in 2050 is the same as Increaseinflood-proneareaduetoclimate in the base year), and about 1.3 million people if the changeinallthreecities 1990 master plan is implemented. In HCMC, cur- rently, about 26 percent of the population would be In all three megacities, in 2050, there is an increase in affected by a 1-in-30-year event. However, by 2050, the area likely to be flooded under different climate it is estimated that approximately 62 percent of the scenarios compared to a situation without climate population will be affected under the high emission change. In Bangkok, for instance, under the condi- scenario without implementation of the proposed tions that currently generate a 1-in-30-year flood, flood control measures. Even with the implementa- but with the added precipitation projected for a high tion of these flood control measures, more than half of emissions scenario, there will be approximately a 30 the projected 2050 population is still likely to be at risk percent increase in the flood-prone area. In Manila, from flooding during extreme events. How to plan for even if current flood infrastructure plans are imple- such large percentages of population being exposed mented, the area flooded in 2050 will increase by 42 to future flooding needs to be seriously considered. percent in the event of a 1-in-100-year flood under the high emission scenario compared to a situation Costsofdamagelikelytobesubstantialand without climate change. In HCMC, for regular canrangefrom2to6percentofregional events in 2050, the area inundated increases from 54 gdp percent in a situation without climate change to 61 percent with climate risks considered under the high In Bangkok, the increased costs associated with emission scenario. For extreme (1-in-30 year) events, climate change (in a high emission scenario) from in 2050, the area inundated increases from 68 percent a 1-in-30-year flood is THB 49 billion ($1.5 billion), (without climate change) to 71 percent (with climate or approximately 2 percent of GRDP. These are the risks considered) under the high emission scenario. additional costs associated with climate change. The Further, there is a significant increase in both depth actual costs of a 1-in-30-year flood--including costs and duration for both regular and extreme floods resulting from both climate change and land subsid- over current levels in 2050 in HCMC. The analysis ence--are close to $4.6 billion in 2050. In Manila, a also highlights areas that will be at greater risk of similar 1-in-30-year flood can lead to costs of flooding flooding in each metropolitan area. In Metro Manila, ranging from PHP 40 billion ($0.9 billion)--given for instance, areas of high population density such current flood control infrastructure and climate con- as Manila City, Quezon City, Pasig City, Marikina ditions--to PHP 70 billion ($1.5 billion) with similar xiv | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport infrastructure but a high emission climate scenario. a high emission scenario in 2050. One out of eight Thus, the additional costs of climate change from of the affected inhabitants will be those living in a 1-in-30-year flood would be approximately PHP condensed housing areas where the population 30 billion ($0.65 billion) or 6 percent of GRDP. The primarily lives below the poverty line. Of the total HCMC study adopts a different methodology to ana- affected population, approximately one-third may lyze costs and its results cannot directly be compared have to encounter inundation of more than a half- to the costs of Manila and Bangkok. The HCMC study meter for at least one week, marking a two-fold uses a macro approach and estimates a series of an- increase in the vulnerable population. People living nual costs up to 2050. The flood costs to HCMC, in in the Bang Khun Thian district of Bangkok and the present value terms, range from $6.5 to $50 billion.5 Phra Samut Chedi district of Samut Prakarn will be The "annualized" costs of flooding would likely be especially affected. In HCMC, in some of the areas, comparable to the costs of Bangkok and Manila. both the poor and non-poor are at risk. However, in general, poorer areas are more vulnerable to flood- damagetobuildingsisanimportant ing. Thus city planners need to devise strategies componentofflood-relatedcosts that focus on the poorer sections of the city through improved access to housing, infrastructure and Damage to buildings is a dominant component of drainage, devising appropriate land use policies flood-related costs, at least in Bangkok and Manila. and improving the level of preparedness among In these cities, over 70 percent of flood-related costs the more disadvantaged social groups. in all scenarios are a result of damages to buildings. Cities are, almost by definition, built-up areas full landsubsidenceisamajorproblemand of concrete structures, so it is not surprising that canaccountforagreatershareofthe the main impact of floods is on these structures damagecostfromfloodingcomparedto and the assets they carry. In HCMC, 61 percent of climate-relatedfactors urban land use and 67 percent of industrial land use are expected to be flooded in 2050 in an extreme One of the main findings of this study is that non- event if the proposed flood control measures are climate-related factors such as land subsidence are not implemented. Potential flooding in HCMC also important and in some cases even more important has major implications for planning in key sectors than climate risks in contributing to urban flooding. such as transportation and waste management. For In Bangkok for instance, there is nearly a two-fold instance, the city's existing and planned transpor- increase in damage costs between 2008 and 2050 due tation network, wastewater treatment plants and to land subsidence. Further, almost 70 percent of the landfill sites are likely to be exposed to increased increase in flooding costs in 2050 in the city is due flooding under the high emission scenario even with to land subsidence. While data for land subsidence the implementation of the proposed flood protec- were not available for Manila and HCMC and this tion system, raising important issues for planners issue was not considered in the hydrological model- such as managing the environmental consequences ing for these two cities, available literature suggests of flooding. Thus, as cities develop over the next 40 that it is an important factor in all three cities and years, it will be important to consider climate risks should be considered in follow-up studies. Even in designing their commercial, residential, and in- though the megacities have already undertaken a dustrial assets and zones. number of measures to slow down land subsidence, further regulatory and market incentives are clearly Impactonthepoorandvulnerablewillbe required to stem groundwater losses. City govern- substantial,butevenbetter-offcommunities ments need to better assess factors contributing to willbeaffectedbyflooding land subsidence and consider options to reduce it. In Bangkok, the study estimates that about 1 mil- 5 The exchange rates used were the average exchange rates lion inhabitants will be affected by flooding under in 2008: 1 USD = THB 33.31, PHP 44.47 and VND 16,302.25. executiveSummary | xv ReCoMMendATIonS Climate-relatedrisksshouldbeconsideredas anintegralpartofcityandregionalplanning Coastal cities in developing countries face enor- mous challenges linked with current patterns While improved urban environmental management of population and economic growth, associated is important, the studies also show that given the environmental externalities, urban expansion and additional costs linked with climate change, cities existing climate variability. Climate change will need to make a proactive effort to consider climate- pose additional risks beyond those currently facing related risks as an integral part of urban planning coastal megacities. As the study shows, these risks and to do so now. First, city planners need to de- will also be associated with significant costs to local velop strategic urban adaptation frameworks for populations and infrastructure. Strong political will managing climate risks involving a range of tools is thus needed to strengthen the capacity to address such as policy and regulatory reforms, investments, both existing climate variability and additional risks and capacity building. Such a strategy can provide posed by climate change. Three main lessons stand an overarching framework for actions taken within out from the study. each sector at the regional, delta, and city levels. Second, much more emphasis needs to be given to bettermanagementofurbanenvironment improving the knowledge base regarding climate andinfrastructurewillhelpmanagepotential risks and related socioeconomic and development climate-relatedimpacts factors. Developing and updating scenarios and planning for a range of potential outcomes will be Analysis carried out in the city case studies show critical for urban planners. This can be accomplished that sound urban environmental management is by strengthening the collaboration between plan- also good for climate adaptation. As the Bangkok ning and sector agencies and research institutions, study shows, land subsidence, if not arrested, would thus giving municipal agencies the tools to make contribute a greater share of damage costs from decisions regarding risk management over the long floods than a projected change in climate conditions. term (Rosenzweig et al. 2007) Third, it is important Thus, addressing land subsidence and factors con- to strengthen the capacity of local urban govern- tributing to it is important from the perspective of mental institutions to adapt to climate change. urban adaptation. While the HCMC study has not Among other things, this involves strengthening the estimated the damage costs due to other environ- capacity to prioritize different adaptation options, ment-development factors--such as the presence of improving coordination between various urban solid waste in the city's drains and waterways, poor sector agencies and sector plans, and incorporat- dredging of canals, siltation of drains, deforesta- ing climate change considerations into the earliest tion in the upper watershed--it provides extensive stages of decision making. qualitative evidence to demonstrate the role these factors play in contributing to urban flooding. Targeted,city-specificsolutionscombining Collectively, the studies highlight the importance infrastructureinvestments,zoning,and of addressing existing environment-development ecosystem-basedstrategiesarerequired factors as a critical part of urban adaptation. They also show that given the high risks of continuing Given that cities are characterized by distinct cli- to urbanize according to current patterns, much matic, hydrological, and socioeconomic features-- more effort should be given to considering the but also that the urban poor in general are more environmental implications of urban growth and vulnerable to increased flooding due to climate expansion in the context of managing current and change--targeted, city-specific, and cutting edge future climate risks. approaches to urban adaptation are needed. First, xvi | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport these include strategies that focus on the more of the city. For instance, in HCMC, storm surges vulnerable areas of the city and the urban poor. and sea level rise are important factors contributing Second, as the studies show, hard infrastructure to flooding. However, in Bangkok these factors are interventions can also be usefully combined with relatively less important. The policy implication is ecosystem-based solutions. For instance, construc- that adaptation measures need to be designed based tion of dykes can be matched with management on the specific hydrological and climate character- and rehabilitation of mangrove systems, refores- istics of each city. Fourth, damages to buildings tation of upper watersheds, river and canal bank emerge as a dominant component of flood-related protection, and implementation of basin-wide flow costs, at least in Bangkok and Manila. Vulnerability management strategies. Urban wetlands provide mapping, land use planning and zoning could be a range of services, including flood resilience, al- used to restrict future development in hazardous lowing groundwater recharge and infiltration, and locations, ultimately retiring key infrastructure providing a buffer against fluctuations in sea level and vulnerable buildings in these areas. Similarly, and storm surges. Thus, rehabilitation of urban building codes aimed at flood-proofing buildings wetlands is critical. Third, as the city case studies (including the lowest habitable elevation in vulner- show, while a combination of climate-related factors able areas) could dramatically reduce damage costs. can contribute to urban flooding, some factors are Such targeted measures could go a long way in much more important than others in different cities helping coastal megacities to adapt to current and depending on location, elevation, and topography future climate risks. executiveSummary | xvii 1 Introduction bACKgRoundAndRATIonAle a shoreline. Low-lying coastal areas--defined as areas along the coast that are less than 10 meters As recent weather events have illustrated, coastal above sea level--represent 2 percent of the world's areas in both developing and more industrialized land area, but contain 13 percent of the urban popu- economies face a range of risks related to climate lation (McGranahan et al. 2007). A recent study of change (IPCC 2007a). Anticipated risks include an 136 port cities showed that the population exposed accelerated rise in sea level of up to 0.6 meters or to flooding linked with a 1-in-100-year event is more by 2100, a further rise in sea surface tempera- likely to rise dramatically, from 40 million cur- tures by up to 3° C, an intensification of tropical rently to 150 million by 2070 (Nicholls et al. 2008). and extra tropical cyclones, larger extreme waves Similarly, the value of assets exposed to flooding and storm surges, altered precipitation and run- is estimated to rise to $35 trillion, up from $3 tril- off, and ocean acidification (Nicholls et al. 2007). lion today. The study also shows that significant, The Intergovernmental Panel for Climate Change increasing exposure is expected for the populations Fourth Assessment Report (IPCC 2007a) points to and economic assets in Asia's coastal cities. a range of outcomes under different scenarios and In flood-prone cities such as Manila, potential identifies a number of hotspots--including heav- sea level rise and increased frequency and inten- ily urbanized areas situated in the large low-lying sity of extreme weather events poses enormous deltas of Asia and Africa--as especially vulnerable challenges on urban local bodies' ability to adapt. to climate-related impacts. For instance, by 2080, Apart from their location, the scale of risk is also the report points out, many millions more people influenced by the quality of housing and infra- may experience floods annually due to sea level structure, institutional capacity with respect to rise (IPCC 2007a). More frequent flooding and in- emergency services, and the city's preparedness undation of coastal areas can also result in various to respond. The urban poor are most at risk from indirect effects, such as water resource constraints exposure to hazards in coastal cities, as they tend due to increased salinization of groundwater sup- to live in riskier urban environments (such as plies. Human-induced pressures on coastal regions floodplains, unstable slopes), tend to work in the can further compound these effects. informal economy, have fewer assets, and receive The location of many of the world's major cit- relatively less protection from government institu- ies--such as Mumbai, Shanghai, Jakarta, Lagos, tions (Satterthwaite et al. 2007). and Kolkata--around coastlines, rivers, and deltas Despite its importance, few developing coun- provides an indication of the population and as- try cities have initiated efforts to integrate climate sets at risk. Thirteen of the world's 20 largest cities change issues as part of their decision-making are located on the coast and more than a third of process. Given the risks faced by coastal cities and the world's population lives within 100 miles of the importance of cities more broadly as drivers of 1 regional economic growth, adaptation must become flooding in Manila caused by tropical storm Ketsana a core element of long-term urban planning. in September was the heaviest in almost 40 years, Recognizing the importance of this issue, the with flood waters reaching nearly 7 meters. More World Bank, Asian Development Bank (ADB) and than 80 percent of the city was underwater, caus- the Japan International Cooperation Agency (JICA) ing immense damage to housing and infrastructure agreed to undertake an analysis in several coastal and displacing around 280,000­300,000 people.8 cities to address climate change adaptation and All of this highlights the need to better understand prepare a synthesis report based on the city-level and prepare for such climate risks and incorporate findings. The selected cities include Manila (led by appropriate adaptation measures into urban plan- JICA), Ho Chi Minh City (led by the ADB), Bangkok ning. While there is a growing literature on cities (led by the World Bank's East Asia and Pacific Re- and climate change, as yet there is limited research gion), and Kolkata (led by the World Bank's South on systematically assessing climate-related risks Asia Region). This synthesis report builds on the at the city/local level and assessing damage costs, analysis undertaken in three of these cities--Manila, particularly in cities in developing countries. This Bangkok, and Ho Chi Minh City (Figure 1.1).6 report aims to fill this gap. Further, it aims to provide The different cities were selected given the science-based information to support urban policy threats they face from increasing hydrometeorologi- and planning as these issues are being debated at cal variability driven by climate change. Bangkok, the local, national, and global levels. located in the Chao Phraya delta, was identified as a hotspot in a background report to the IPCC's AR4 (IPCC 2007b). Manila was identified in OECD's vul- objeCTIve nerable port cities report (Nicholls et al. 2008), par- The main objective of this report is to strengthen ticularly regarding typhoon damage. HCMC ranked our understanding of climate-related risks and im- fifth by population exposed to the effects of climate pacts in coastal megacities in developing countries change (Nicholls et al. 2008). A recent study also using case studies of three cities that are different identified Manila, Ho Chi Minh City, and Bangkok in their climate, hydrological, and socioeconomic among the top eleven Asian megacities that are most characteristics. Specifically, it draws on in-depth vulnerable to climate change (Yusuf and Francisco analysis of climate risks and impacts in three cit- 2009).7 Devastating floods in Manila in September ies--Bangkok, Manila, and Ho Chi Minh City--to and October 2009 only confirm the vulnerability of highlight to national and municipal decision makers this city to extreme weather events. For instance, (a) the scale of climate-related impacts and vulner- abilities at the city level, (b) estimates of associated damage costs, and (c) potential adaptation options. FIguRe1.1 AsianMegacityhotspots The comparative analysis carried out in this report IBRD 38067 shows the increasing climate risks faced by coastal SEPTEMBER 2010 megacities and the need to consider adaptation as part of long-term strategic planning. Even though the study is based on analysis in three cities, the THAILAND Bangkok VIETNAM Manila 6 The Kolkata study was not completed at the time of the Ho Chi Minh PHILIPPINES City preparation of the synthesis report and thus was not in- cluded in main report. Annex A provides a brief overview of the study. 7 Vulnerability in the scorecard was understood in This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on terms of exposure, sensitivity, and adaptive capacity of the cities. See also http://www.idrc.ca/uploads/user- this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. S/12324196651Mapping_Report.pdf 8 http://edition.cnn.com/2009/WORLD/asiapcf/09/27/ Source: Asia map IBRD 38067 philippines.floods/index.html 2 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport policy implications have broader relevance for HCMC, which considered the A2 and B2 scenarios), assessing climate risks and identifying adaptation and estimated climate risks to 2050. The 2050 time options in other coastal areas. horizon for the study is appropriate, given planning horizons in most cities and given that the typical time frame for major flood protection planning is pRoCeSSoFpRepARATIon about 30 years. Moreover, the uncertainty in cli- mate projections expands rapidly past roughly the The analysis was carried out over a period of one- mid-21st century, providing additional justification and-a-half years. The synthesis team and the city- for limiting the time horizon to 2050. Climate vari- level teams worked closely to develop common ables considered included changes in temperature, terms of reference to guide the city-level studies. changes in precipitation, estimated sea level rise, Further, while the synthesis team helped coordinate and estimated storm surge. In addition, non-climate the process, each city-level team worked indepen- factors--such as land subsidence, land use changes, dently with their respective country counterparts to salinity intrusion, and population increases--were build ownership and capacity for the analysis. The also considered. Flooding in the metropolitan ar- city teams were comprised of members with a range eas was chosen as the key climate variable to be of skills, including climate modeling, hydrological examined. The approach consisted of the following analysis, GIS mapping, economic analysis, and sequential steps: (1) downscaling climate variables urban planning. The city teams and the synthesis to the level of the city/watershed; (2) hydrometeo- team preparing this report also met periodically to rological modeling and scenario analysis, presented share methodological issues and ongoing findings in GIS maps; and (3) damage/loss assessment and and research. These discussions and the analysis identification/prioritization of adaptation options. undertaken for each city have formed the basis of These steps are discussed in more detail in chapter 2. the preparation of this synthesis report. At the level To support this analysis, each city team collected of each city, the teams have undertaken stakeholder extensive historical and city-specific data related consultations with city officials and government to past climate events such as storms and flooding, agencies at different levels, nongovernmental socioeconomic data, information about local topog- organizations, the private sector, and other con- raphy and hydrology, information on land use, and stituencies. For instance, the main counterparts in so forth. Data limitations were a major challenge, HCMC were the HCMC People's Committee and but each team worked with existing data from pub- the Ministry of Natural Resources and Environment lic sources, as well as data made available by city (MONRE); the study sought to inform preparation governments and institutions. There are numerous of HCMC's city-wide adaptation plan. In Bangkok, uncertainties at each step of the analysis. the main counterpart was the Bangkok Municipal While the main focus of this report is on assess- authority. In Metro Manila, it was the Metro Manila ing future climate risks at the city level, it builds Development Authority (MMDA). At the global on the recognition of strong links between climate level, preliminary findings have already been pre- adaptation and ongoing efforts toward disaster sented at several international forums to reach risk management. Despite the institutional differ- urban planners, municipal decision makers, and ences in terms of how these efforts have emerged, researchers. and differences in how climate change/variability and disasters manifest themselves, they both share common ground in striving toward strengthening oveRvIewoFMeThodology/ adaptive capacity of vulnerable communities, build- AppRoAChAndClIMATe ing resilience, and reducing the impact of extreme pARAMeTeRSSeleCTed The city-level studies considered two IPCC emis- 9 Different scenarios were considered to assess impact due sions scenarios,9A1FI and B1 (with the exception of to the uncertainties in projecting future climate conditions. Introduction | 3 events. The analysis undertaken in this report uses level through downscaling techniques, flood risk several methodologies that have long been used assessment through hydrometeorological models, in the context of disaster risk management--such and damage cost analysis. Chapter 3 presents as damage cost assessment and probabilistic risk the main findings from climate downscaling, analysis--illustrating the opportunities for cross- hydrological modeling analyses, and use of GIS fertilization in both areas. mapping.10 Chapter 4 presents the analysis and findings relating to damage cost assessment, as well as an analysis of adaptation options. Finally, STRuCTuReoFTheRepoRT Chapter 5 draws broad policy lessons and presents conclusions. Chapter 2 presents methodologies used to deter- mine climate change risks at the city/river-basin 10 For a broader set of GIS maps, please refer to city-specific reports. 4 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Methodologies for Downscaling, Hydrological Mapping, and Assessing 2 Damage Costs I n order to assess the impact of climate change Rangeofemissionsscenariosconsidered in terms of increased flooding in 2050 in each of the coastal cities, three main methodological The potential impact of climate change can vary steps were taken. These include (1) determining greatly depending on the development pathway climate-related impacts at the city/river-basin level that is assumed. Beginning in 1992, the IPCC has through downscaling; (2) developing flood risk provided various scenarios for the emissions of assessment hydrometeorological models for each greenhouse gases based on assumptions of differ- city to estimate flooding in 2050 under different sce- ent development pathways--namely, complex and narios; and (3) assessing damage costs. This chapter dynamic interactions among future demographic provides a summary of these methodologies. It changes, economic growth, and technological and en- highlights the climate change scenarios selected, ap- vironmental changes. These emissions scenarios are proaches to downscaling, assumptions underlying projections of what the future may look like and are hydrological analysis, and the approach to damage a tool to model climate change impacts and related cost assessment. Some of the methodologies used uncertainties. As described in the IPCC Special Report here--such as damage cost analysis and probabilis- tic risk assessment--are also used in disaster risk management.11 Uncertainties and errors involved 11 A probabilistic risk assessment provides an estimate in different steps of the analysis are also discussed. of the probability of loss due to hazards. It is commonly used in disaster risk management planning and provides a quantitative baseline for measuring the benefits (or losses avoided) of disaster management alternatives. In climate SeleCTIonoFeMISSIonS change impact and adaptation studies, it also provides a baseline for assessing the change in risks due to the in- SCenARIoS,downSCAlIng, creasing hydrometeorological hazards associated with AndunCeRTAInTIeS climate change. See, for instance, Earthquake Vulnerability Reduction Program in Colombia, A Probabilistic Cost-benefit Analysis (World Bank Policy Research Working Paper To measure the impact of climate change on the 3939, June 2006) for an example of a probabilistic risk as- cities in 2050, it was necessary to assume emissions sessment used in disaster risk management planning. The scenarios and as a first step, "downscale" climate process involves the development of several intercon- nected modules, which calculate in turn the hazard prob- change forecasts to local levels so that the meteoro- ability, exposure, vulnerability (or sensitivity to damage), logical parameters--such as changes in temperature damages, and losses. While the approaches used in the de- velopment of the modules and the calculation of the losses and precipitation--could be applied as inputs to the varied, each city case study did, however, follow a similar hydrometeorological models. analytical process. 5 on Emissions Scenarios (IPCC 2000), four storylines and impact assessments at national or regional level yield four different scenario families--A1, A2, B1, (Jones et al. 2004). Further, GCMs are not designed to and B2--that have allowed development of 40 differ- study hydrological phenomena, and GCM outputs ent scenarios, organized in six different groups (IPCC and hydrological inputs are not at the same temporal 2007a). Each of these scenarios is equally valid, with and spatial scales.15 Another related but important no probabilities of occurrence being assigned. Thus, consideration is to perform bias corrections on the the A1 storyline refers to assumptions of a future GCM results. To overcome some of these problems, world of rapid economic growth, global population downscaling techniques have been developed to that peaks in mid-century and declines afterwards, obtain local-scale surface weather (at resolutions of and the introduction of efficient technologies. The A1 10­50 kilometers), from regional scale atmospheric storyline is disaggregated into three groups--based variables that are provided by GCMs. Downscaling on alternative directions for technological changes in is the process of making the predictions from global the energy system--where A1FI refers to fossil inten- climate models (GCMs/AOGCMs) relevant to a spe- sive sources. The B1 scenario family also assumes a cific region so as to generate appropriate inputs for global population that peaks in mid-century but is other tasks (in our case to feed hydrological models). based on an assumption of a shift toward a service There are several methods, each with its own degree and information economy and the introduction of of challenges based on the quality/quantity of de- clean technologies.12 Together, these scenarios "cap- tailed local/regional meteorological data required ture the range of uncertainties" linked with different and computational complexity (Box 2.1). driving forces (Jones et al. 2004). For this study, the A1FI and B1 scenarios were Cascadingsetofuncertaintiesinusing chosen because they represent the high and low climatemodelsforassessingimpactsat brackets, respectively, of the estimated global temper- locallevel ature increases under the SRES storylines. In contrast to the Bangkok and Manila studies, HCMC used the There are a number of caveats about the use of A2 and B2 scenarios. The main reason for this is that climate models. First, there is a cascading set of the A2 and B2 scenarios have been adopted as the uncertainties, starting with the emission scenario official climate scenarios for Vietnam by the govern- chosen, uncertainties in future concentrations and ment of Vietnam under the national target program CO2 feedback cycles, uncertainties in the response to respond to climate change.13 There are numerous of the climate, the AOGCMs used, the downscaling uncertainties associated with projecting future cli- technique utilized, and the manner in which the im- mate; selecting different scenarios allows the case studies to account for some of these uncertainties. 12 For more details, see http://www.ipcc.ch/pdf/special- reports/spm/sres-en.pdf. downscalingfromglobalclimatemodels 13 They were also chosen because they were the scenarios available for the ECHAM model used in the study. See Projections of future climate change are usually de- HCMC study, Annex C. 14 AOGCMs are global climate models that couple togeth- rived from global climate models (GCMs). A GCM is er the interactions between atmosphere and ocean. For a a mathematical representation of the climate system nested RCM, there is generally no feedback process of the based on the physical attributes of its components, changes in the RCM cells into the parent AOGCM, leading to disparities in boundary conditions between the RCM and their relation, and various feedback processes. the parent AOGCM. Various emissions and concentration scenarios (dis- 15 Empirically, it has also been observed that only poor cussed above) are used as input into climate models quality hydrological models can be derived when using the GCM outputs directly, and that even simple downscaling to estimate global climate projections. GCMs--in- methods improve the quality of hydrological models. For cluding atmosphere-ocean general circulation mod- hydrological impact studies of climate change, the impor- els (AOGCMs)14 --are run at a course spatial resolu- tant climatic variables are temperature and precipitation, which are downscaled to provide inputs of precipitation tion (a few hundred kilometers) and cannot capture and evapotranspiration to hydrological models while main- the local detail needed for hydrological modeling taining the correlation between the downscaled variables. 6 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport box2.1 S trengthsandlimitationsofdifferentdownscalingTechniques SelectedforthisStudy The city studies utilized two different downscaling techniques--pattern scaling which is a kind of statistical downscaling, and dynamic downscaling techniques. The first was undertaken for all the studies. The HCMC study also used dynamic downscaling. Briefly, statistical downscaling techniques are "based on the construction of relationships between large scale and local variables calibrated from historical data" (Jones et al. 2004). These statistical relationships are then applied to large-scale climate variables from AOGCM simulations to estimate corresponding local/regional characteristics (like temperature and precipitation). Pattern scaling, one of the techniques used in this work, is a form of statistical downscaling that requires minimal regional meteorological data and is computationally not very complicated. In pattern scaling the change in the local variable of interest--like temperature or precipitation--is represented in terms of the GCM's change for the variable per unit change in global temperature (this is the scaling factor). This factor is then multiplied by the global change in temperature associated with a specific IPCC scenario (A1FI and B1). In this specific case, the scaling factor was obtained by performing a linear regression on a number of GCM models available (see Sugiyama 2008 for details). The main advantage of statistical downscaling is that it is computationally not very expensive and can provide information at point locations. However, among several limitations, one issue is that the statistical relationships may not remain the same in a future climate world and the method does not provide information on temporal and location linkages. In contrast, dynamical modeling techniques use physical models of the climate system allowing direct modeling of the dynamics of the physical systems that influence the climate of a region. They are rapidly becoming the most widely applied downscaling technique. New systems like PRECIS--used by the HCMC study--can be utilized from a PC platform, but they can still generate a broad range of daily, site-specific (i.e., 25 kms resolution grids) hydrometeorological data for time periods spanning a century. This allows examination of time slices like 2030 to 2070 to establish frequency relations for events in 2050 that can include both floods and droughts. Boundary conditions need to be derived from coupled GCMs. Nevertheless, RCMs also have a number of potential pitfalls that need to be understood. They require large amounts of boundary data linked with the parent GCM, acquire the errors of the parent GCM, and (as is the case with PRECIS) may take several months to run. An additional limitation is their potential exclusive reliance on only one AOGCM. Before applying the RCM, the user needs to understand whether it can model the types of events (e.g., typhoons) that are of particular interest in the study. Typically, for studies that involve downscaling analysis such as disaster risk assessments, when trying to incorporate climate change risk factors into the analysis, statistical downscaling or regional climate models are likely to be the methods of choice. Source: Authors' compilation. pact parameters (e.g., precipitation and temperature 24-hour precipitation increase factors, and (c) sea- increase) that are generated are applied in estimat- sonal mean precipitation increases for 2050 for the ing flood impacts at the city level. Trying to quantify two climate change scenarios A1FI and B1. These these uncertainties--and given a confidence interval factors were derived from16 AOGCMs models for for the outcomes--is extremely difficult, and some all four cities and are shown in Table 2.1. uncertainties--like future emission scenarios--can- The robustness of the relationships was exam- not be assessed. However, despite these uncertain- ined statistically by the IR3S team by examining ties, the IPCC AR4 recognizes that the AOGCM the scatter among the data points generated by the climate change models do allow global forecasts, downscaling of the 16 AOGCMs. They found ro- and that increasingly the downscaling techniques bust relations for global and local temperature and are providing information on the likely scale of dif- precipitable water (used as a proxy for extreme pre- ferent climate impacts at local levels (Giorgi 2008). cipitation) relationships, and viable but less strong relationships for the temperature/mean seasonal estimatesforchangesintemperatureand precipitation increases. How this was handled is precipitationin2050basedonstatistical explained in Box 2.2. downscalingapproach 16 This part of the analysis was carried out by Integrated The statistical downscaling technique used in the Research System for Sustainability Science (IR3S) at the study (see Sugiyama 2008) 16 provided estimates of University of Tokyo. The results are presented in Sugiyama (a) the expected temperature increases, (b) extreme (2008). Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 7 TAble2.1 ClimateChangeForecastsfor2050 City Manila Bangkok Ho Chi Minh City Kolkata Approx. longitude 121° E 100° E 106° E 88° E Approx. latitude 14° N 13° N 10° N 23° N Tlocal / Tglobal [unit less] 0.88 0.94 0.89 0.90 Tlocal (A1FI) [K] 1.8 1.9 1.8 1.8 Tlocal (B1) [K] 1.1 1.2 1.2 1.2 Pmean / P present mean 2.0 1.5 2.2 5.1 / Tglobal [% / K] Pmean / Pmean (A1FI) [%] (June-July-August) present 4.0 3.0 4.4 10 Pmean / P present mean (B1) [%] (June-July-August) 2.6 2.0 2.9 6.6 Pextreme / P present extreme 8 8 8 9 / Tglobal [% / K] Pextreme / Pextreme (A1FI) [%] present 14 15 14 16 Pextreme / P present extreme (B1) [%] 9.2 9.8 9.3 11 Source: Sugiyama (2008). box2.2 d ownscalingfrom16gCMs To develop the estimates, the study first downscaled global mean temperature and precipitation. The pattern scaling technique adopted ignores the differences between locations within the grid used in the GCM. The downscaled relationships were robust and the inter-model variation was small. To examine the increases in extreme and mean precipitation, the study examined the forecasts made by the models for 24-hr extreme and mean precipitation increases. They found significant variability among them in the simulation. For example, the factor increase for extreme precipitation in the tropics as drawn from the models ranged from 3 percent/°C to 28 percent/°C. For storms with a 1-in-30-year return period, the mean increase among the models was 10 percent/°C, but the standard deviation was 6.9 percent/°C , making the statistical relationships unusable. To overcome this problem, the study used the increase in precipitable water (i.e., the amount of water in a column of air that could fall as precipi- tation), for which there was a more robust relationship within the models, which researchers have indicated would serve as a good scaling factor for extreme precipitation (Allen and Ingram 2002) as a proxy for the increases in extreme precipitation. Such relationships have been established in the literature (see the report for the references). For mean precipitation, the team found a stronger correlation in monthly mean precipitation estimates and used these data sets to make the mean increase estimates. Source: Sugiyama, 2008 useofregionalclimatemodelinginhCMC B1 scenarios respectively. For the parent AOGCM study used, the A1FI and B1 forecast simulations were not available. Although HCMC utilized a dynamic In addition to the above, the HCMC study also uti- downscaling technique, they also compared their lized the PRECIS model nested in the low resolution precipitation forecasts with those derived from the (~2o) ECHAM Version 4 AOGCM model (European statistical method and found the results broadly Centre for Medium Range Weather Forecasts, Uni- consistent with their forecasts. versity of Hamburg) to generate daily hydrome- teorological data for the study.17 For this modeling, 17 The Southeast Asia (SEA) Regional Center (RC) for the A2 and B2 scenarios were used, which for the year Systems for Analysis, Research and Training (START) un- 2050 do not deviate dramatically from the A1FI and dertook the work. 8 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport uncertaintiesintheassignmentoffuture flooding in each coastal city under different sce- probabilities narios in 2050. Outputs from the downscaling analysis--including estimates and assumptions Hydrometeorological probability assessments (as for a range of climate variables (such as sea level used in this study) are based on the analysis of rise, storm surge) and non-climate variables (such historical data of extreme events and assume that as land subsidence, land use, population changes, the underlying population's statistics (e.g., mean, economic growth, existing and planned infrastruc- standard deviation, skew, etc.) do not change over ture)--were used as inputs into the hydrological time. In this regard, the discussion of probabilities analysis and GIS mapping. This part of the analysis associated with future extreme events presents is discussed below. a conundrum, which is approached differently depending on whether a statistical or nested RCM estimatingfloodinginurbanareasunder downscaling technique has been used. For the sta- differentscenarios tistical methodology used in this study, a factor in- crease has been computed for extreme 24-hour and The risk of loss from a hazard such as flooding mean seasonal precipitation events. These factors is based on exposure and, consequently, is site were applied to current events with the comparable specific. Thus, to estimate the extent of flood risk probability. For the RCM methodology, simulations in each city, hydrometeorologic models specific to over a 40-year time slice around the study date each city (Box 2.3) were developed. These models (2050) allowed calculation of the probability of the were based on a host of historical and city-specific extreme events. These approaches have some in- information--such as existing drainage and sew- herent limitations. For the statistical methodology, erage systems, soil characteristics, river flow, it assumes that the parent population's inherent canals, dams, land subsidence, siltation, existing variability remains the same, which could easily flood protection infrastructure, and so forth--to be wrong. For the RCM, decadal variations, which estimate future flooding under different scenarios. also affect the population sets' characteristics, may Floods can occur from a combination of factors, not be reflected. The statistical downscaling "factor including upstream storm runoff, storm rainfall increase" or RCM forecasts can have significant on the city, or sea level rise and storm surge acting errors even though they both reflect best practices alone or in conjunction with other storm events. in climate science. The specific question of the These factors were incorporated into the scenario influence of these errors on downstream impact analysis. assessments has not been studied in any detail. In The extent of flooding was estimated in terms addition to estimating changes in temperature and of the depth, duration and area flooded and over- precipitation under different scenarios in 2050, each laid on GIS maps with land-use mapping of the city team also made estimates for changes in storm population (disaggregated by poverty levels and surge and sea level rise. Further, where data was other vulnerability criteria), assets, infrastructure, available, estimates for land subsidence were also utilities, and environmental and cultural resources. considered in estimating flooding under different This was done both for the base case and for the scenarios. How each city study approached these year 2050 under different climate scenarios. For the issues is discussed in more detail in chapter 3. purposes of this study, the city studies assumed that without climate change, the climate in 2050 would be similar to the climate in 2008/base year. hydRologICAlModelIng This helped assess the scale of physical damage FoRdevelopIngSCenARIoS that can be caused to infrastructure, income, land, oFFloodRISK populations, etc, in the cities with and without climate change. The second sequential step of the analysis involved hydrological analysis to estimate the extent of Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 9 Returnperiodsconsidered18 Large watersheds with many sub-basins, res- ervoirs, diversions, etc. have necessarily complex In this study, for each of the cities, flood risk was models describing the water transfer within the estimated for events of different intensities or return basin. For the Chao Phraya River Basin, for ex- periods, namely a 1-in-10-year, 1-in-30-year, and ample, a complex combination of models needed 1-in-100-year floods in the base year and in 2050.19 to be interlinked to accurately simulate flooding in This was based on a statistical evaluation of historic Bangkok. The schematic shown in Figure 2.1 below storm events that allowed the assignment of prob- shows the interconnectedness of rainfall-runoff (on abilities to extreme precipitation events20 associated the upper watersheds), reservoir and river routing, with the floods. and flood routing (2D). The complexity has an im- pact on the time and depth of analysis. For example, hydrologicalmodelingundertakento a simulation run for Bangkok with one set of input estimatefloodingunderdifferentscenarios parameters took 24 hours of computer time, and the study required over 30 simulation runs not count- The hydrological models simulate the movement ing the runs required for calibration. HCMC used a of water on land after it falls as precipitation.21 A simulation model that had been previously devel- key initiating input parameter for such modeling oped and calibrated by the government. Manila's is precipitation (derived for each city from the hydrologic models were far simpler and the period downscaling analysis). The manner and timing of of flooding modeled was only a few days, so the the precipitation's arrival in the rivers is driven by length of the simulation runs in terms of computer the precipitation intensity and duration, the terrain time was far shorter than for Bangkok. slopes and land use cover (which can drive infil- tration and evapotranspiration), and the channel Calibrationofhydrometeorologicalmodels characteristics. Computer models, in which the rel- evant information is digitized, allow the otherwise As noted above, to ensure the accuracy of the mod- daunting calculation of the movement of water els, rigorous calibration is necessary. The process re- through the watershed and channel. In addition quires significant trial and error in getting the model to the input parameter of precipitation and spatial to accurately reflect historic movements of floods as data affecting the flow of the runoff, the boundary runoff, through river channels and overland flood (or tail-water) conditions where the flood exits a events. In Figure 2.2 below for Manila, the dotted system into the sea can create a backwater effect, line shows the observed discharge and the blue line raising the water level profile of the river upstream from the sea and increasing the overbank spillage 18 This refers to the recurrence interval or the return period and flood inundation levels. Storm surge can also of a flood and is an estimate of the time interval between cause the movement of the water upstream from the two flood events of certain intensity. It is a statistical mea- sea into the coastal cities. The hydrometeorological sure of the average recurrence interval over a long period of time and is the inverse of the probability that the event will models were designed to model these interactions be exceeded in any one year. For example, a 10-year flood and are based on general principles outlined below has a 1 / 10 = 0.1 or 10 percent chance of being exceeded in any one year and a 50-year flood has a 0.02 or 2 percent (Box 2.3). chance of being exceeded in any one year. Thus, a 100-year Manila and Bangkok applied commercially flood has a 1 percent chance of occurring in any given year. available software to develop the hydrologic mod- However, as more data becomes available, the level of the 1/100 year flood will change. els and HCMC utilized hydrologic models that had 19 HCMC restricted its analyses to 1-in-100 and 1-in-30-year already been developed and calibrated by the gov- floods. 20 An extreme event that causes flooding could be from ernment for their catchment area. Outputs from the extreme 24-hour rainfall on a small watershed, a high sea- HCMC hydrologic models can also be linked to com- sonal rainfall on a large watershed, or a typhoon or cyclone mercial or open-sourced software models to refine and associated storm surge. 21 Storm surge and SLR have been incorporated into the inundation information. Typically a combination models as boundary tailwater conditions for the hydrologic of models was used in the city studies (Table 2.2). models. 10 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport box2.3 S omebasicprinciplesforhydrologicalMapping Hydrologic models use a mass balance approach in developing the empirical relationships, where- CONSTANT INTENSITY RAINFALL in the volume of inflow (or precipitation) minus the losses equals the volume of outflow. For the RATE, DEPTH PER UNIT TIME rainfall-runoff relationship, the rainfall excess, which appears as runoff in the river channels, is RAINFALL EXCESS equal to the precipitation minus the evaporation, infiltration, depression storage, and interception ION AT TR as depicted in the figure below. The empirical IN FIL models must be calibrated against observed DEPRESSION STORAGE EVAPORATION events in order to provide useful estimates. + INTERCEPTION The empirical relationships generally capture physical information (e.g., land use information, soil types, antecedent rainfall, TIME area, and gradients for rainfall-runoff models, or river gradient, cross-sections, and roughness estimates for river-channel routing models) that are used with parameters to generate the input-output relationships. The parameters are adjusted in an iterative approach to obtain the best fit to the actual data as part of the calibration process. Source: American Society of Civil Engineers, Hydrology Handbook (1996). shows the simulated discharge using the model S TAble2.2 ummaryofCityCase and the actual input precipitation data, indicating StudyhydrologicModeling the robustness of the model. This type of "ground- City Study Hydrologic models used truthing" was done for the hydrological modeling Bangkok Rainfall-runoff modeling of the upper Chao Phraya in the Bangkok and HCMC studies as well. watershed entailed dividing the basin into 15 sub- basins and using NAM model; the MIKE FLOOD software model was used for the lower Chao Phraya coupling the h FIguRe2.1 ydrometeorological MIKE 11 and MIKE 21 for 1D and 2D flows.* ModelSchematicfor HCMC Rainfall in the upper watershed of the Saigon and Chaophrayawatershed Dong Nai rivers are captured by the Dau Tieng and Tri An reservoirs; outflows from these were modeled by the Southern Institute of Water Resource and Planning based on historical (2000) flooding. The HYDROGIS model was used for the modeling of HCMC. It has been specifically designed and calibrated by Vietnam's Ministry of Natural Resources and Environment's Institute of Meteorology, Hydrology and Environment and models both 1D and 2D flows. Manila The NAM software was used to model the rainfall-run- off relations linked to the MIKE FLOOD; MIKE 11 and 21 were used to model floods in Manila. Note: * When the water is flowing in a channel, the model is frequently referred to as one-dimensional (1-D) and overland flow models are referred to as two-dimensional (2-D) models. Source: Panya Consultants. Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 11 FIguRe2.2 ManilaRainfall-RunoffCalibrationhydrographs Source: Muto et al. (2010). Infrastructurescenariosandadditional of goods and services to the economy (ECLAC 2003; climateinputsintohydrologicalmodeling Messner et al. 2007).23 These costs are in terms of traffic delays, production losses, or emergency ex- Since the extent and impact of flooding in 2050 in penditures. Some delays may be immediate, while different cities will depend not only on climate and others may be more medium term because of the weather variables but also on the infrastructure secondary effects of flood-related transportation available in the cities to drain water and prevent bottlenecks and re-directed traffic, for example. flooding, the hydrological models also assumed Table 2.3 summarizes the different direct and indi- certain infrastructure and flood protection scenarios rect and tangible and in-tangible costs associated to estimate flood occurrences. Further, the studies with flooding. also made estimates and made assumptions regard- The Bangkok and Manila case studies follow the ing storm surge and sea level rise as inputs to the approach outlined above and estimate direct and hydrological modeling. These assumptions and indirect costs for four areas: (1) buildings, indus- estimates are discussed in more detail in chapter 3. try, and commerce; (2) transportation and related infrastructure; (3) public utilities such as energy and water supply and sanitation services; and (4) AppRoAChToASSeSSIng people, income, and health. The HCMC study also dAMAgeCoSTS discusses the impacts of extreme events on natural ecosystems. However, the HCMC study does not The main impact of climate change on the four cities provide any detailed monetization of impacts. The in 2050 is assumed to be in the form of increased more macro approach it takes to damage cost assess- flooding. Flooding has direct and indirect effects. ment is discussed in detail later in this chapter. The The direct impacts have to do with immediate physical harm caused to "humans, property and 22 One of most difficult of the direct costs to value is loss of the environment" (Messner et al. 2007) and involve life. Manila and HCMC have both witnessed loss of lives costs from loss in the stock of infrastructure, tangible during major storms in the past but do not attempt to value assets and inventory, agricultural and environmen- such losses from future storms. 23 It is important to be careful not to double count damages tal goods, and injuries and life loss.22 The indirect since the value of the sum for flows from an asset would impacts of flooding are a result of a loss in the flow equal the value of the asset. 12 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble2.3 directandIndirectCostsfromFlooding Tangible Intangible Direct Costs Repair, replacement, and cleaning costs associated with physical damage to assets Loss of human life Public infrastructure Loss of ecological goods Commercial and residential buildings Loss of archaeological resources Inventory Vehicles Loss of productive land and livestock Crop loss1 Indirect Loss of industrial production or revenues Long-term health costs from pollution or flood Costs Increased operational costs for commercial entities injuries Lost earnings or wages Post-flood recovery inconvenience and Time costs from traffic disruptions vulnerability Post-flood flood proofing investments by the private sector Emergency flood management costs to the public sector 1 Income loss from crop destruction is conventionally treated as a direct cost. If land is salinized or there is permanent inundation, then the loss in the asset value of land is the cost to include. Care needs to be taken to distinguish between one-time crop loss and land value losses that are of longer duration. Source: Modified from Messner et al. (2007). methods used by the Manila and Bangkok study to damagestobuildings,industry,commerce, develop cost estimates are discussed below. Figure andresidentsinbangkokandManila 2.3 summarizes the main steps taken in estimating costs and then estimating the net benefits from A key impact of floods is on existing buildings. investments to reduce flooding. Floods are assumed to cause damage but not destroy FIguRe2.3 estimatingImpacts--AFlowChart Current and 2050 Valuation of damages Adaptation Floods (all cities) (Bangkok and Manila): Identification of adaptation 1 Area flooded 1 Repairs of buildings strategies (all cities) 2 Duration of flood 2 Losses in household assets and Estimation of adaptation 3 Depth of flood commercial inventory and assets infrastructure costs and 3 Transport delay and infrastructure annualized benefits based on different cli- costs (Bangkok and Manila) mate changes, land sub- 4 Revenue losses to the public sector Estimation of discounted sidence, and assumptions (electricity utilities, hospitals, net benefits (Bangkok and about flood control infra- Impacts in flooded transportation authorities, water Manila) structure areas on: and sanitation utilities) 1 Industry and Commerce 5 Wage and income losses 2 Land, agriculture and ecosystems 3 Transportation Vulnerability assessment (HCMC) 4 Energy, water and and valuation of: sanitation 1 Infrastructure at risk 5 Income, population, 2 People at risk and health 3 Land at risk based on flood duration 4 Likely GDP losses and depth 5 Ecosystems at risk (no valuation) Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 13 e FIguRe2.3 stimationofdamagetobuildings,Assets,andInventoriesinthe bangkokandManilaCases 1 2 3 4 5 Identify buildings in Classify buildings into Identify assets and Establish average val- Apply a damage rate flooded areas using commercial, industrial, inventories in build- ues of buildings and to different buildings building /infrastruc- residential or other ings based on surveys assets/inventories and assets/inventories ture surveys. sub-classifications. or on the type of based on tax informa- to establish actual building. tion, industrial sur- damage at different veys, household levels of flooding. The durable wealth data damage rate is etc. obtained from previ- ous flood information or from secondary sources and varies with the level of floods. buildings. The steps used in estimating damage logical analysis), the value of floor areas of flood- costs in the case of Bangkok and Manila are identi- affected buildings was identified. fied in Figure 2.4. In order to estimate the actual cost of damage to buildings, the case studies used different damage Identifying,classifying,andvaluingdamage rates depending on the level of the flood.24 The dam- tobuildings age rate or the percent of the value lost from floods is dependent on the depth of inundation. A different The Bangkok and Manila case studies used informa- rate is applied to assess damage to buildings and tion from building censuses and other infrastruc- damage to assets. The Bangkok study used a 1997 ture-related studies to classify buildings first into survey to establish damage rates; the rate was used residential, commercial, and industrial buildings. for all types of buildings and housing (residential, Based on national statistical organization classifi- commercial, and industrial). The Manila study used cations, these categories of buildings were further damage rates, which vary across buildings and assets. subdivided. For example, commercial buildings Assets such as machinery, office furniture, alone were subcategorized into nine groups in the and inventory damaged by floods also need to be case of Bangkok. valued. The loss of assets from damaged buildings Once buildings were classified into different was evaluated based on "representative assets" subgroups, they were valued based on available in different types of buildings. For example, in government data. The Bangkok case study identi- the case of the Bangkok case study, average asset fied the book values of buildings of different types values obtained from census data for buildings are based on construction costs when the buildings THB 328889 ($9,873) for Bangkok and THB 220180 were legally registered. These values were then ($6,609) for Samut Prakarn. In the Manila study, depreciated to reflect aging. In Manila, the base the value of stocks, assets, and inventories of com- unit construction cost of residential and commer- mercial and industrial enterprises was obtained cial buildings was obtained from the government from the National Statistical Organizations' data of assessor's office and the median value of different types of buildings estimated. Then, based on the 24 The flood damage rates used in the case studies were ob- estimated depth of floods (output from the hydro- tained from either local surveys or from secondary sources. 14 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport establishments. A damage rate (Table 2.4) was ap- F Table2.4 looddamageRatebyType plied to assets and inventories to determine losses. ofbuildinginManila The Bangkok study used a similar approach. Flood Level To estimate losses to residential assets, the value Type of Building Affected Assets 100­200 cm 200­300 cm of household goods first needs to be established. In the Manila case, the value of household effects was Residential Finishings 0.119 0.58 estimated to be 35 percent of the value of residential Household Effects 0.326 0.928 construction costs of flooded buildings. Similarly, fin- Business Entities Assets 0.453 0.966 ishings were estimated to be 25 percent of construc- Stocks 0.267 0.897 tion costs. In the Bangkok study, average household Source: Adapted from Manual on Economic Study of Floods, Japan assets or durables were obtained from the Population Note: Damage rates are available for less than 50cm and more than 300 cm as well and Housing Census of 2000. The value of these assets was based on market prices. Once the asset values were obtained, damage rates were used to was identified. The length of major and minor cur- estimate the losses from different levels of flooding. rent and future roads likely to be affected by floods was then established. Based on unit maintenance cost data from the Department of Public Works Assessingimpactsonroadsand and Highways, the cost of maintaining these flood- transportationnetworks affected roads was estimated. The Manila study also Flooding affects road and train networks in all the estimated indirect costs from flooding related to cities. There are direct and indirect costs associated time delays from traffic disruptions and problems with loss of transportation infrastructure. Direct with vehicle operations. In the Manila case, the costs are (a) construction or repair costs; and (b) vehicle operations cost (VOC) is obtained from the any losses of vehicles or damaged trains and so on. Department of Public Works and Highways and in- These costs are estimated where applicable, based cludes fixed costs of operating vehicles and running on average repair costs. Indirect costs associated costs. The costs of floods or the change in VOC per with damage to transportation infrastructure include km was estimated as the difference between VOC increased vehicle operations costs (VOC) resulting on flooded roads and on roads in good condition. from damages to roads from flooding or because of This difference in VOC is multiplied by the number changes in traffic pattern (increased petrol costs), or of affected roads in 2008 and 2050 to obtain damage time costs from losses in productivity (ECLAC 2003). costs in these two periods. To estimate time-delay More than 90 percent of inhabitants of Bangkok costs, data on flood related transport delays and and Samut Prakarn travel by roads and highways. the number of road commuters were obtained from However, the road network in Bangkok, Samut secondary sources and transportation surveys.25 Prakarn, and the BMR are set at an elevation of The number of commuters delayed by flooding 1.5­2.0 meters above ground, and most of them was then multiplied by average hourly income as are covered with reinforced concrete pavement or a proxy for time costs. 7­10 cm of asphaltic pavement. Therefore, they are not expected to incur flood damages. Rail lines are Assessingdirectandindirectcostsin generally set at about 1 meter above the surround- energy,watersupplyandsanitation ing ground level. The Bangkok rapid transit system, both elevated and underground, has been designed Energy, like the transportation sector, may sustain to be protected from overflow flood water. Thus, direct and indirect damages from floods. Direct the costs associated with transport networks from damages to electricity generation plants, transmis- flooding in Bangkok are expected to be minimal. 25 The Manila study draws heavily on the government's The Manila study undertook a detailed analysis Transport Master Plan, which lays out potential transporta- of damages to the transport sector. First, based on tion changes up to 2015, and on the Metro Manila Urban the transport master plan, future road construction Transportation Integration Study (MMUTIS). Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 15 sion and distribution lines, and energy distribution tor because of the position of the pipes and extent centers depends on whether these facilities are in of pressure in them. the flood-affected area of each city. If they are, then the extent of damage is estimated and replacement estimatingincomelosses and repair costs are used to value these damages based on cost information available with govern- As previously noted, each city estimates the popu- ment departments. In most cases, electric utilities lation likely to be affected by 2050 floods. Where have strict rules for cutting electricity during certain possible, survey data and other secondary sources types of floods. This leads to revenue losses, which were used to identify the characteristics of the af- were estimated in the Bangkok case study. In terms fected population, particularly to examine whether of direct costs, future power generation plants for floods will affect the poorest communities. Bangkok and Samut Prakarn are planned outside the cities and there will be no direct impact. The Incomelossestothecommercialsector Manila study did not estimate energy-related costs. As with the case of the energy sector, costs to In order to estimate income losses from floods, the water and sanitation systems are evaluated only if Manila and Bangkok case studies examined losses it is established that there is a likelihood of flood to firms and individuals. During the duration of the damage to specific subcomponents such as water flood, income losses are borne by the private com- intake facilities, pumping stations, water treatment mercial sector because of a halt in their activities. plants, main lines to storage tanks, storage tanks, Since it is difficult to directly estimate these costs distribution networks, and so on. In the case of without detailed production information, where Bangkok, it is assumed that water supply will be- available, national statistics on income per day as- come dysfunctional if water floods reach 2 meters sociated with different kinds of commercial struc- above the ground surface. This will disrupt water tures were used. services to the area and result in revenue losses to The Bangkok study identified the number of the utilities as well as other costs, including health buildings in flood-affected areas and then, based costs. The calculation of sales loss for the water sup- on the duration of the flood, estimated the rate of ply system due to flood damages in Bangkok was damage to these buildings. To obtain business losses, estimated as follows: the Bangkok case study then identified the average income per day from commercial establishments watersupplysalesloss=no.affectedusers based on business surveys. The loss in value added xwaterdemandperuserxwatersalesrate per day (net income) was adjusted due to savings in xfloodduration business expenditure when an operation is closed down. The net commercial income was estimated The data for this estimation came from the five-year to be THB 4930 ($149) per establishment per day. record of the Metropolitan Waterworks Authority, A similar accounting was used to estimate average which indicates water sales for a residential unit are income to industrial establishments. These average 0.48 m3 per day per household and water revenue is values are multiplied by the number of days of flood- 2.06 baht per m3. Water sales for a nonresidential unit ing and the number of buildings affected by floods to are 3.71 m3 per day per customer and water revenue estimate income losses from commerce and industry. is 2.83 baht per m3. In the Bangkok study, direct dam- The Manila case study took a similar approach. age to the water supply and sanitary infrastructure The analysis was based on a 2008 survey on busi- is not assessed since they are protected from the ness income losses to flood-affected firms. The study worst possible flood in the future. The calculation used data on sales losses as a substitute for income of income losses to sanitation service providers or profit data, which were unavailable. Buildings because of the shutdown of sanitation services was that were affected by floods were categorized ac- similarly undertaken. The Manila study did not find cording to different commercial activities; each floor major costs associated with the water delivery sec- was assumed to host one firm. By multiplying the 16 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport sales loss data (categorized by type of economic and expert advice. The main diseases associated with activity) by the number of affected buildings in each floods are diarrhea, conjunctivitis, athlete's foot, economic category, the study established the value malaria, cholera, and typhoid. Ideally, we would of income losses to the commercial sector. require information on different outbreaks and cost per episode per person for specific diseases in order estimationofincomelossesforpoorand to value health impacts. However, this is difficult non-poorhouseholds to obtain and some approximations are made. The Bangkok study makes an attempt to examine health Households who work in flood-affected areas may costs associated with floods. In this case study, the lose income during the duration of the flood. A number of individuals affected by floods is estimated general assumption made in the Bangkok case is and multiplied by the average per person health costs that salaried workers will not see income losses associated with hospital admissions. This is clearly a from floods. However, the informal sector needs first approximation for what might be the real costs further accounting. In Bangkok, the low-income on health from flooding. However, lack of data, and housing developments, called condensed hous- difficulties in establishing dose-response informa- ing, in flooded areas are identified and--based on tion related to different flood-related diseases, lead estimates of household size--the number of daily to this more simplistic analyses. The Manila study wage earners in this area is estimated. Then based does not estimate the health costs from floods even on the daily wage rate, the income loss to poor and though it makes an attempt to estimate some of the non-poor households is calculated. health impacts of flooding. A more detailed analysis The Manila case study examines income losses of health impacts needs to be carried out as a follow- to workers in the formal and informal sectors. Be- up to this study. cause of lack of data on number of households, the case study, like Bangkok, relied on the number of buildings affected by floods. Non-poor residents ASSeSSMenToFdAMAgeCoSTS were assumed to occupy residential buildings at the InThehCMCSTudy rate of 1.5 households per building. This allowed the Manila study to estimate the total number of non- The HCMC study used a macro-approach to assess poor resident households in flood-affected "formal" damage costs. Two different methods were used to residential buildings. The total number of affected calculate the cost of climate change at an aggregate households was multiplied by the average income macro-level: (1) cost estimates using land values; and per household (based on government statistics) to (2) cost estimates based on aggregate GDP loss. The obtain income losses to the formal sector. methodology applied in each of these approaches To establish losses to the poor, the Manila study and the results obtained are described below. first estimated the number of informal structures affected by floods. It was assumed that at least two Thelandvalueapproach households live in each structure. Thus, the total number of affected households was estimated. Economically speaking, the value of any asset is Again, using government statistics on the poverty the sum of the value it is expected to generate over level per day of PHP 266 per household, the total time, discounted to reflect people's preferences for losses to the informal sector was then established. consumption in the present. The same is true for land. In principle, the current value of the land stock limitedanalysisofhealthimpactsinthe in HCMC is the value of future production that can citystudies be expected to be generated with this asset. Land value can be a particularly good guide to the cost of An important question concerns the impact of floods climate change as land values capitalize most values on public health. Each city is different and has health that need to be captured in any assessment of cost impact data based on previous experience with floods (such as roads, railways, water supply systems, Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 17 FIguRe2.5 possibleRelationshipsbetweenFlooddurationandlandvalueloss 1 0.9 0.8 0.7 Loss of land values 0.6 Linear relationship between 0.5 percentage of time the area is 0.4 ooded and % loss of land value 0.3 Quadratic relationship between 0.2 percentage of time the area is ooded and % loss of land value 0.1 0 0.01 0.06 0.11 0.16 0.21 0.26 0.31 0.36 0.41 0.46 0.51 0.56 0.61 0.66 0.71 0.76 0.81 0.86 0.91 0.96 (Days of ood/365) Source: ADB (2010). drainage systems, energy infrastructure, hospital lationship between the economic value of land and and schools). An estimation of how the impacts of the number of days in any given year that an area is climate change may affect the value of the land stock flooded is not known, and must be assumed. in HCMC thus provides a first-order approximation In the HCMC report, two types of relationship of the economic costs of climate change for the city. have been assumed: The loss of land value is directly The first step in the analysis was to determine proportional to the proportion of time the land is land prices. Land price data was aggregated by inundated; this is the linear relationship in Figure 2.5. district and an average land price determined for For example, if the land is flooded for 182 days per each district. The administratively determined year, (about half of the time), this assumption yields price for land is published by HCMC People's an estimated loss of 50 percent of the land value. A Committee annually, as required under the 2004 proportional relationship would overestimate the Land Law. These prices are used to determine the loss in land value. Indeed, if land is flooded only a level of compensation for state appropriations of few days a year, we might suppose that this would land and for taxation purposes. In principle, these have a limited impact on the value of the land, if any. administrative prices are based upon the potential On the other hand, the proportion of the loss of land productive value of the land, rather than market value may increase at an increasing rate as the length prices (although following the 2004 Land Law, the of time the land is flooded increases, reaching 100 administrative price for land has moved closer to percent of land value lost when the land is flooded land market values). These are current land prices, for 100 percent of the time (the quadratic relationship not 2050. These are also not market prices. is shown in Figure 2.5). The quadratic relationship The area subject to flooding both in extreme is considered to represent a more realistic relation- events and regular flooding was then determined ship between flood duration and land value loss as under future scenarios using the HydroGIS mod- small periods of flooding (e.g. for one or two days) eling. Once the average land price and flooding are unlikely to influence land value much, whereas extent and duration were estimated, the relation- longer term flooding (e.g. for 300 days a year) is ship between the flooding and the decline in the likely to reduce the value to near zero. economic value of the flooded land was determined to allow the calculation of the cost of flooding due Thegdpapproach to climate change. The study assumes that there is some positive relationship between the duration of The second approach to estimating the costs of the flooding and the loss in land value: that is, the climate change is to calculate costs on the basis of longer the period of flooding, the greater the loss in expected losses in production (proxied by GDP) due land value. However, the precise nature of that re- to climate-change-induced flooding. Like land, GDP 18 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport measures capture a number of important values (level gross domestic product, urbanization, flood and of economic development, extent of industrial activi- transportation infrastructure, and spontaneous ties, transportation system, water supply, drainage, adaptation and migration. Where possible, they etc). To calculate costs using this method, a number built on existing plans and studies undertaken by of important assumptions were integrated into the city governments related to urban growth, public analysis: Areas subject to flooding lose all their utilities and transportation networks, and so forth. productivity for the duration of the flood. The GDP Population in 2050 is estimated based on the best produced in a particular area is directly proportionate available local projections for each city. The popu- to population (i.e. if an area accounts for 1 percent of lation distribution within each city in 2050 builds the population it will also account for 1 percent of the on available estimates of population density and GDP). GDP is produced evenly over time (i.e. with distribution and some understanding of future 1/365th of annual GDP being produced everyday). population dynamics. Economic growth and pov- Based on GSO statistics, GDP growth will average erty projected in 2050 varies across the case studies. 11.5 percent between 2006­10; 8.7 percent between These assumptions are made explicit in the follow- 2011­25; and 8 percent between 2026­50. Population ing chapters with respect to each city. growth will take place in line with the high estimate Human spontaneous adaptation to the possi- discussed in chapter 4. Future values are discounted bility of increased flooding is considered but is not at an annual rate of 10 percent.26 fully incorporated into the studies. This is partly Given these assumptions and the information because of the difficulties in predicting adaptation on flooding derived from the HydroGIS models, an behavior and partly because populations, while estimation of the costs due to flooding is relatively likely to respond to incremental change, may not straightforward. For each district and for each year have a measured response to the type of extreme until 2050, the following calculation was performed: events considered in this study. 27 The case studies do take into account human migration in identifying Annual cost Number of days the district is flooded x population distribution in 2050. In general, popula- = of flooding number of people affected x GDP/capita/day tions exposed to flooding are likely to continue to be exposed to flooding except in cases where there The annual cost of flooding was calculated is improved flood infrastructure put into place. for each year between 2006 and 2050, and the dis- counted costs summed for the whole period. The Assumptionsaboutprices HCMC study presents results from both the land values and GDP measures. The differences in the Given difficulties in estimating future prices, all results and approaches are discussed in chapter 4. the three case studies largely keep the real value of goods and services unchanged. The Bangkok study estimates the costs of damage in 2008 using current ASSuMpTIonSAbouT TheFuTuReoFCITIeSIn 26 Because in the case of land, the land value already rep- eSTIMATIngdAMAgeCoSTS resents the discounted future income stream capitalized in land value, so there is no need for discounting when using Cities in 2050 are likely to be vastly different from land values to estimate the cost of climate change in the pre- vious analysis. today's cities. Understanding how different is a 27 Yohe et al. (1995), for example, argue that if permanent huge task and there is no attempt to model economic inundation is expected due to SLR, it might be appropri- growth and link it to urban development. Instead, ate in cost-benefit analyses to treat the value of inundated buildings as zero or very low. If there is full information and existing data and official plans and projections are efficient adaptation, economic agents can be expected to let used to understand and develop GIS-based maps. these buildings fully depreciate before they are inundated. Further, in assessing damage costs, each city study In our study, we do not assume full information. Nonethe- less, rational households who can afford to move are likely made assumptions about population size and distri- to move away from flood-prone areas, which may be then bution, poverty, economic growth and city-specific populated by poorer communities. Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 19 prices and the same prices are applied to 2050 dam- form of recreation. In these circumstances, the costs ages. The Manila study assumes a 5 percent annual of flooding may be very different. growth in the value of goods and services lost from Impacts on populations working in flooded floods; that is, damages grow annually at 5 percent areas are another area of concern. It is hard to estab- up to 2050. However, for comparison, the Manila lish what categories of people may actually inhabit study deflates its 2050 values and presents the 2008 flood-prone areas in 2050. While the studies incor- values of all damages whether they occur in 2008 or porate trends in population distribution, it is not 2050. The HCMC study uses a macro approach to clear that these trends will remain unchanged in the estimating costs and examines current land values next forty years. It is rational to consider a scenario as well as annual income loss based on annually where increased flooding will lead to commercial projected GDP.28 establishments moving away from flood-prone The interest rate is an important price variable areas. Thus, these areas could become recreational for assessing adaption investments. The case studies areas as previously discussed, or they could become use discount rates that are normally used to dis- a refuge for the poor and vulnerable because of low count public investments in each city. The Bangkok land values. These complex issues cannot be teased study presents results using 8 percent, 10 percent, out in the damage costs framework that is used. and 12 percent rates. The Manila study uses a 15 There are two confounding issues to consider percent discount rate and the HCMC study uses a in examining health impacts: (a) climate change is 10 percent rate. likely to change the incidence and geographic range of different diseases such as dengue or malaria, for example, and it is difficult to predict changes in the ConCluSIon: general prevalence of these diseases in the three cit- MeThodologICAllIMITATIonS ies in 2050; and (b) economic growth and increased AndunCeRTAInTIeSIn education will potentially reduce the prevalence of certain diseases such as diarrhea. Each city has InTeRpReTIngReSulTS made its own assumptions about health impacts oFTheSTudy and provides qualitative and some quantitative information on costs, but this information is subject difficultieswith2050projections to significant uncertainty. While the studies attempt to establish some under- estimatingphysicaldamagesrequiresa standing of how the cities may look in 40 or so years varietyofassumptions from now as is discussed in the following chapters, this is really very difficult to do. Hydrological Establishing exactly how and what will be affected models in all three studies incorporate flood control by natural disasters is not easy. The Bangkok and infrastructure and the type of land use that may ex- Manila studies examine a variety of impacts on ist in the future. Future transportation networks are buildings and income, but a series of assumptions incorporated into damage assessments undertaken underlie this analysis. In order to estimate income by Manila and Bangkok, and official population losses, for example, both case studies have to make projections are used. Yet there are numerous as- assumptions about the number of households and sumptions underlying each of these projections and firms residing in flood-affected buildings. Without a the future may look different from what is assumed. careful survey of each flood-affected area, it would The road networks may not be built, new land use plans may be put in place, and exogenous shocks 28 In order to examine adaptation-related investments, some could change the population distribution. We can additional assumptions are made. The Bangkok study esti- mates net benefits under two scenarios: (a) prices stay the imagine a situation where flood-prone areas are same, and (b) net benefits grow at an annual rate of 3 percent . made into wetlands or parks and set aside for some The Manila study uses a 5 percent annual rate of growth. 20 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport be very difficult to know how accurate these as- when service is interrupted. In Bangkok, water and sumptions are. energy-related flood costs, however, are estimated by examining revenue losses to the government. valuationconsiderations Another important issue is the role of prices. It is very difficult to estimate the value of assets There are numerous challenges to valuing flood and incomes damaged by flooding in 2050. Given impacts that have to do with current and future data difficulties in predicting prices, most prices are availability. Flooding damages to building stocks held constant or in some cases some reasonable and machinery, for instance, are best valued at their growth rates are assumed. However, urban real depreciated replacement cost. In practical terms, estate prices and the prices of a variety of goods the valuation exercise was carried out in different and services damaged by floods can change signifi- ways in each study depending on data availability. cantly and this would affect the evaluation in 2050. For example, in Bangkok the value of an entire as- There are also difficulties in estimating income or set or building was assessed and a percentage of productivity losses as a result of floods. The Manila the value was treated as the damage cost. Further, study and Bangkok (for low income households) in Manila and Bangkok, buildings are evaluated at estimate the number of households affected by book-values, or the cost of the buildings when they floods and multiply this by an average value of were initially established. In HCMC, land is evalu- household income based on the type of houses ated at government-established prices, but in some these families live in. However, average values cases the market value of this land may be higher. used may not apply for flood-prone areas. Further, Also, the cost of repairing flooded components of a the families may have a way of earning income building may be very different from the value ob- even under flood circumstances if they can leave tained by applying a damage rate to the entire value the area for work. In the HCMC study, a macro of the building. Another issue is how losses related approach is applied and GDP per capita is used to electricity and water supply have been evaluated as the value of productivity lost in flooded areas. because of lack of appropriate information. The costs However, if these areas may not be producing the of supply disruptions should be assessed by estimat- "average GDP per capita," this could lead to an ing the economic value of public water/electricity overestimation of impacts. These uncertainties and supply, or, for example, by estimating the costs limitations need to be borne in mind in interpreting to water consumers of securing alternative water the results of this study. Methodologiesfordownscaling,hydrologicalMapping,andAssessingdamageCosts | 21 Estimating Flood Impacts and Vulnerabilities 3 in Coastal Cities I n this chapter, the main findings of the hydro- l FIguRe3.1 ocationofbangkokinthe logical analysis--in terms of potential increase ChaoprayaRiverbasin in flooding in 2050 under different scenarios-- IBRD 37477 100° 102° are presented. For each city, the analysis presents THAILAND CHAO PHRAYA RIVER the main drivers of flooding, how climate change PROJECT RIVER LAO P. D . R . 20° is likely to influence flooding, and the potential SELECTED CITIES NATIONAL CAPITAL physical consequences on the city and its residents INTERNATIONAL BOUNDARY in 2050. The findings from the hydrological model- Chiang Mai Nan ing were overlaid on GIS maps, thus relating flood depth/duration and area flooded (estimated from 18° Phrae 18° M YA N M A R the hydrologic modeling) with maps of future Yo Uttaradit m land use; location of residential, commercial, and industrial areas; and essential infrastructure and Tak THAILAND environmental resources. Use of GIS enables a Andaman Kamphaeng Sea Phet Phichit 0 100 Pin visual mapping of the area inundated and districts g Nan KILOMETERS 16° This map was produced by the Map Design Unit of The World Bank. and sectors likely to be at risk from flooding under The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any Nakhon Sawan endorsement or acceptance of such boundaries. different scenarios in 2050.29 For the purposes of this MYANMAR NAYPYIDAW LAO HANOI Uthai Thani Chainat P.D.R. VIETNAM report, only a selection of GIS maps is included; a VIENTIANE Sing Buri Thahanbok Lop Buri THAILAND more extensive set of maps can be obtained from Ang Thong Chao P h r a 15°N Supham Buri BANGKOK Phra Nakhon Si Ayutthaya T h a C hin Area of Map the city-level studies. CAMBODIA a y Andaman 14° Pathum Thani 14° PHNOM Sea PENH Nonthaburi Gulf of Bangkok 15°N Thailand Samut Prakan Ratchaburi Samut Sakhon Gulf of eSTIMATIngFuTuReClIMATe- Thailand 15°N 100°E MALAYSIA 105°E 100° 102° JANUARY 2010 RelATedIMpACTSInbAngKoK Source: Thailand map IBRD 37477. The Bangkok Municipal Region covers an area of about 1,569 sq kms and is located in the delta of the level due to land subsidence. Bangkok straddles Chao Phraya River basin--the largest river basin in the Chao Phraya River approximately 33 kilometers Thailand, covering an area of 159,000 sq kms, or 35 above the Gulf of Thailand. percent of the total land area of the country (Figure 3.1). The basin area is flat at an average elevation 29 For 2050, projected land use changes--where available-- of 1­2m from mean sea level. Some areas are at sea were incorporated into the GIS to assess the future impacts. 23 ImportanceofbMRtotheregional unit and population was used as a base for the as- economy sessment of income loss of the poor. The Bangkok Metropolitan Region (BMR) is the ClimateandprecipitationintheChao economic center of Thailand. It is the headquarters phrayaRiverbasin for all of Thailand's large commercial banks and financial institutions. The area to the east of Bangkok Bangkok has a tropical monsoon climate. The aver- and Samut Prakarn is also an important industrial age annual rainfall over the basin is 1,130 mm, and is zone. In 2006, the gross domestic product (GDP) of higher in the northeastern region of the basin (Table the Bangkok Municipal Region was 3,352 billion 3.2). About 85 percent of the average annual rainfall baht, or 43 percent of the country's GDP (7,830 billion occurs between May and October (Panya Consul- baht). The annual average growth rate was 7.04 per- tants 2009). Tropical cyclones occur between Sep- cent, and per capita GDP was 311,225 baht (Bangkok tember and October. In this case, rainfall continues case study report). The current official population of for a long period of time in a relatively wide area. Bangkok is estimated to be about 10 million people The peak river discharge is registered in October at (based on 2007 estimates) with an estimated growth the end of the rainy season. Severe flood damage rate of .64 percent between 2003­07.30 may arise with high tide in this period (ibid.). BMA's maximum temperature is in April and its minimum natureofurbanpoverty temperature is in December. The mean temperature ranges from 26°C to 31°C. As stated in the Bangkok study, in 2007, 0.6 percent or 88,361 people in the BMR were poor (Table 3.1). Mainclimate-relateddriversofflooding The poverty line for the Bangkok municipal region was 1,638 baht ($49) per person per month in 2007.31 Severe flooding in Bangkok is associated with The number of poor is an official estimate and does heavier than normal rainfall occurring over several not include unregistered people. Most of the poor live in condensed housing and are unregistered. Statistics from the Office of National Economic and 30 This figure does not include unregistered migrants. 31 Poverty incidence is measured at the household level by Social Development Board (NESDB 2007) show comparing per capita household income against the pov- 768,220 people living in 133,317 housing units of erty line, which is the income level that is sufficient for an the condensed housing area. Therefore in the as- individual to enjoy society's minimum standards of living. If an individual's income falls below the poverty line, he or sessment of flood impacts on the poor, this housing she is classified as poor. TAble3.1 povertylineandthepoorinthebMR 1 Poverty Line (Baht/person/month) Proportion of the Poor (%) No. of the Poor (person) Province 2006 2007 2006 2007 2006 2007 BMA 2,020 2,065 0.51 1.14 28,692 64,422 Samut Prakarn 1,647 1,712 -- 0.78 -- 9,961 Samut Sakhon 1,511 1,564 0.76 0.42 4,313 2,436 Nonthaburi 1,529 1,561 0.30 0.06 4,124 845 Pathum Thani 1,409 1,458 0.56 0.20 5,376 1,939 Nakhon Pathom 1,434 1,466 0.45 0.98 3,918 8,758 BMR 1,592 1,638 0.43 0.60 46,422 88,361 Source: The Office of National Economic and Social Development Board (NESDB), 2007 as cited in the Bangkok city study. 1 Most of the poor live in condensed housing and are unregistered. So the figure of 768,220 does not align with the number of poor in the table. 24 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble3.2 bangkokMonthlyAverageTemperatureandprecipitation Month Temperature Average Low (oC) Temperature Average High (oC) Average precipitation (mm) January 23° 32° 7.2 February 24° 33° 13.7 March 26° 34° 32.5 April 27° 35° 58.3 May 27° 34° 170.8 June 26° 33° 112.5 July 26° 33° 118.1 August 26° 33° 160.0 September 25° 33° 262.5 October 25° 32° 207.3 November 24° 32° 32.4 December 22° 32° 3.3 Total 1,171.4 Source: http://weather.uk.msn.com/monthly_averages.aspx?wealocations=wc:THXX0002 months during the monsoon period over the Chao urbanization, and removal of natural attenuation Phraya watershed. The flooding can last several basins (like wetlands) also contribute to flooding. weeks to over a month. Intense short-duration rain- Land subsidence is a critical factor in flooding and fall over the city can cause localized flooding that is discussed later in the section. normally lasts for less than 24 hours. Bangkok is not subject to direct hits from tropical typhoons or existingfloodprotectioninbangkok-- cyclones. Large floods have occurred in 1942, 1978, diversioncanalsandflooddikeswith 1980, 1983, 1995, 1996, 2002, and 2006. A number pumpeddrainage of dams in the upper watershed--including the Bhumibol Dam (1964) and Sirikit Dam (1971)--have In response to the flooding, Bangkok has imple- helped reduce flooding in Bangkok. In 1995 a seri- mented a large-scale flood protection system that ous flood occurred due to a series of tropical storms includes a flood embankment surrounding the reaching the upper watershed from the end of July city and a pumped drainage system. In 2006 there to September. The resultant runoff exceeded the stor- was another significant flood event in the upper age capacity of the Sirikit Dam, which had to release Chao Phraya watershed. Flood peaks were again 2,900 million m3 of storm runoff. Despite efforts to attenuated by allowing the agricultural lands north attenuate the flood wave by allowing flooding of of Bangkok to flood, and flood flows were also di- agriculture lands upstream of Bangkok, the flood verted to channels to the east of Bangkok. The flood wave crest reached Bangkok at the same time as the embankment system and pumped drainage system high spring tide, flooding 65 percent of BMA with protected the city. Nevertheless, areas surrounding inundation depths up to 2 meters and left some areas Bangkok continue to experience inundation during flooded into December. The flood event of 1995 is peak flood flow. The city and national planners estimated to have a return frequency of 1-in-30 years. have plans to add further flood protection, which has been planned and budgeted through 2014. The non-climatedriversofflooding long-term land use plan until 2057 for the greater Bangkok region provides for development areas In addition to climate-related drivers, non-climate and environmental zones that will be used to allow drivers such as land subsidence, deforestation, flood flows and drainage. Most of the flood control estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 25 activities that are being considered in the Royal Ir- watershed using spatial and temporal patterns from rigation Department master plan are infrastructure historic flood years. The forecasted temperature activities (such as construction and improvement of increases of 1.9°C and 1.2°C for the A1FI and B1 canals, dikes, pumping stations, etc.). Flooding of scenarios respectively for 2050 were used in the hy- agricultural lands upstream of Bangkok is utilized drological modeling. The SLR estimates used were during major floods to help attenuate the flood peak, 0.29 m and 0.19 m for the A1FI and B1 scenarios lowering the flood profile in Bangkok. respectively. This was based on IPCC AR4 (Working Group 1, Chapter 10, Table 10.7) estimates. It was Currentandplannedlanduseandurban decided to use half of the 2099 increase for the 2050 planninginbangkok--bangkokinthefuture estimates; namely, 0.29 m and 0.19 m for the A1FI and B1 scenarios respectively. The Bangkok study projected what the city would look like in the future based on a wide range of available data about population and growth and a T box3.1 hebangkokMetropolitan number of sectoral and urban development plans Region(bMR):Some (Box 3.1). Modern urban land use planning in Assumptionsaboutthe Bangkok started in the 1950s. Bangkok's inner city Future contains the Grand Palace, government offices, ma- jor universities and educational establishments, and Bangkok is changing rapidly and some of these changes have been taken into account in modeling the future. Bangkok is 2-to-4-story row houses that are used as commercial increasingly "suburbanizing." Between 1998 and 2003, BMR saw and residential units. The inner city is a national a 74 percent increase in urban development. The city's housing historic conservation area where construction of stock has doubled in the last decade. This trend is expected to high-rise buildings is prohibited. Urban growth in continue. The Government's Land Use Plan for 2057 is used to Bangkok, however, has progressed in a sometimes project the number of commercial, residential, and industrial ad hoc manner and without a unified plan linking buildings in Bangkok in 2050. Available government plans are taken into account in establishing future road and transportation it to the surrounding areas. To remedy this problem, networks. Plans suggest tightened urban areas within a network the government has developed a 50-year regional of expressways bounded by large environmental protection areas spatial plan (finished in 2007) that foresees growth to divert any floods from the centers. of the surrounding areas. This plan also provides Based on existing trends, the city of Bangkok is not expected for significant environmental protection and wet- to grow a lot, while neighboring suburbs will continue to grow. land areas. For instance, suburban Nonthaburi increased its population by 47 percent during 2002­07. Bangkok City is expected to have a population of 10.55 million (4.65 million households) and BMR a Climatechangeparameters--statistical population of nearly 16 million by 2050. To forecast the population downscaling in 2050 for BMR, information on population projections for Thailand 2003­30 prepared by the government (NESDB) was used as a base. For the 2050 simulations, the Bangkok City case The NESDB report projects population to the year 2030 for the whole study applied the statistical downscaling factors kingdom and to 2025 and 2020 for Bangkok and other provinces respectively. In projecting the population to 2050, a regression estimated by the University of Tokyo's Integrated function was applied. Research System for Sustainability Science (IRS3) In terms of the BMR economy, regional GDP is projected for two discussed in chapter 2. Given the size of the Chao areas (Bangkok and Samut Prakarn) of BMR most likely to be affected Phraya watershed and the long time it takes upper by flooding. GDP for Bangkok and Samut Prakarn in 2050 is expected watershed rainfall to reach Bangkok as stream flow, to be six-fold higher than the current GDP. In terms of poverty in the team applied the mean increase in seasonal the cities, in Bangkok in 2007 only 88,361 people (0.6 percent of the total) in the BMR were officially considered poor. However, this precipitation (3 percent and 2 percent for A1FI and number excludes some 768,220 people living in 133,317 condensed B1 respectively) to the three-month rainfall events housing units. In the flood impact assessment, the higher, more for the corresponding 1-in-10, 1-in-30 and 1-in-100- accurate number was used. year precipitation events for the watershed. These Source: Adapted from Panya Consultants (2009). total precipitation levels were distributed over the 26 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport estimationofstormsurgeinthecaseof suming that current efforts to reduce groundwater bangkok extraction continue. As a consequence, the team's estimate of land subsidence by 2050 varied from 5 to For storm surge, the team examined historic storm 30 cm depending on the location. Figure 3.2 shows surges in the Gulf of Thailand. There have been land elevations in 2002 versus projected land eleva- three events since 1962 along the coast of southern tions in 2050 as a consequence of land subsidence. Thailand. No recorded typhoon or cyclone has had a track that approached the mouth of the Chao Phraya Simulationeventsmodeled at the north of the Gulf of Thailand. Nevertheless, one typhoon created wind patterns that drove 2 to 3 The study ran simulations for 1-in-10, 1-in-30, and meter waves near the mouth and drove up sea levels. 1-in-100-year flood events for 2008 as the baseline As a result, the team used a storm surge of 0.61 m for for the inundation hazard. The study then ran simu- the study. It needs to be noted that the time frame for lations applying only the land subsidence expected the storm surge would be on the order of 24 hours, by 2050 without applying the climate change factors and only one event has been recorded in 30 years. for precipitation or SLR. This was done to allow dif- The hydrometeorological processes driving these ferentiation of the contribution of land subsidence events have little or no correlation. To have a joint to the flood hazards and risks. Then simulation runs impact, the storm surge effect would have to occur were made for 1-in-10, 1-in-30, and 1-in-100-year near the peak storm discharge for the Chao Phraya flood events in 2050, applying either the A1FI or B1 watershed. The joint probability of occurrence, al- forecasted variables (precipitation and SLR). Finally, though not calculated in detail, is less than 1/1000. simulation runs were made for the impact of storm Table 3.3 summarizes the climate change parameters surge on the 1-in-30-year A1FI and 1-in-100 A1FI used to model the 2050 impacts in Bangkok. and B1 simulations. Infrastructure scenarios were assumed as fol- landsubsidence­asignificantcontributor lows. For the 2008 base year calculations, it was tofutureflooding assumed that the existing and nearly completed flood protection infrastructure was in place. For the Although land subsidence is not driven by climate future 2050 calculations, it was assumed that the change, its impact on flooding in Bangkok needs to planned flood protection infrastructure will have be considered for the 2050 simulations. The team been implemented (Panya Consultants 2009). undertook a comprehensive assessment of historic and forecasted land subsidence for Bangkok, and Conservativevalueswereusedinforecasts held technical consultations with the concerned agencies. Historic land subsidence in Bangkok has The use of the mean increase in seasonal precipitation reduced from highs of 10 cm/year to 1 to 2 cm/year for the Chao Phraya watershed is appropriate, as it over the period of 1978 to 2007, and from 2002 to matches the time to concentration for the flooding 2007 the rate had declined to 0.97 cm/year. The team in Bangkok. Mean seasonal increases were applied predicted that due to government efforts to control to the 1-in-10-year, 1-in-30-year, and 1-in-100-year groundwater pumping, the average subsidence rate precipitation events and distributed across the water- would continue to decline by 10 percent/year, as- shed spatially and temporally using historical rainfall C TAble3.3 limateChangeandlandSubsidenceparameterSummaryforbangkok Temperature increase Mean Seasonal IPCC Scenario (oC) Precipitation Increase (%) Sea Level Rise (m) Storm Surge (m) Land subsidence (m) B1 1.2 2 0.19 0.61 0.05 to 0.3 A1FI 1.9 3 0.29 0.61 0.05 to 0.3 estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 27 2.1.2 Land SubsidenceLand Subsidence 2.1.2 From average data, the average land the Bangkok Metropolitan Region (BMR) From observed data, theobserved land subsidence rate insubsidence rate in the Bangkok Metropolitan Region (BMR) has from 10 reduced from 10 cm per year to 1-2 1978 year from 1978 to 2007. Furthermore, has gradually reducedgradually cm per year to 1-2 cm per year fromcm perto 2007. Furthermore, during the last 5 years (2002-2007), the average land subsidence rate has reduced to 0.97 cm per to 0.97 cm per during the last 5 years (2002-2007), the average land subsidence rate has reduced is expected that is land subsidence land subsidence by would reduce Thus, year. Therefore, ityear. Therefore, itthe expected that therate would reducerate 10% per year. by 10% per year. Thus, the subsidence land 2002-2050 (48 years) would spatially vary from 5 to 30 the accumulated landaccumulatedduringsubsidence during 2002-2050 (48 years) would spatially vary from 5 to 30 l FIguRe3.2 andelevations,2002versus2050landSubsidence cm depending on in Figure shown cm depending on location as shownlocation as 2.1-3. in Figure 2.1-3. 640000 660000 640000 680000 660000 700000 680000 640000 700000 660000 640000 680000 660000 700000 680000 700000 Pathum Thani Pathum Thani Pathum Thani Pathum Thani ! ( ! ( ( ! ! ( 1540000 1540000 1540000 1540000 1540000 1540000 1540000 1540000 Nonthaburi Nonthaburi Nonthaburi Nonthaburi ( ! ! ( ! ( ( ! Bangkok Bangkok Bangkok Bangkok ! ( ! ( ( ! ! ( 1520000 1520000 1520000 1520000 1520000 1520000 1520000 1520000 Samut Prakan Samut Prakan Samut Prakan Samut Prakan ( ! ( ! ( ! ( ! 1500000 1500000 1500000 1500000 1500000 1500000 1500000 1500000 1480000 1480000 1480000 1480000 1480000 1480000 1480000 1480000 640000 660000 640000 680000 660000 700000 680000 640000 700000 660000 640000 680000 660000 700000 680000 700000 Legend Legend Legend Legend ! ( Province Province Boundary River/Canal Network -1 - 0 0-1 1-2 8 - 10 ( ! 10 - 15 15 - 20 Province Province Boundary River/Canal Network -1 - 0 0-1 1-2 8 - 10 10 - 15 0 1.5 3 15 - 20 Ê 6 9 12 ! ( Province 0 1.5 3 Ê Province Boundary 6 River/Canal Network 9 -1 - 0 0-1 121 - 2 8 - 10 ! ( 10 - 15 15 - 20 Province Province Boundary River/Canal Network -1 - 0 0-1 1-2 8 - 10 10 - 15 0 1.5 3 15 - 20 Ê 6 9 12 0 1.5 3 Ê 6 9 12 Elevation (m.MSL) 2-3 20 - 25 Elevation (m.MSL) 2-3 20 - 25 Kilometers Elevation (m.MSL) 2-3 20 - 25 Elevation (m.MSL) 2-3 20 - 25 Kilometers Kilometers Kilometers -7 - -5 -5 - -3 3-4 4-5 25 - 35 -7 - -5 -5 - -3 Land Elevation 25in35 2002 Land Elevation in 2002 3-4 4-5 - -7 - -5 -5 - -3 3-4 4-5 25 - 35 -7 - -5 -5 - -3 Land Elevation in -2050 Land Elevation in 2050 3-4 4-5 25 35 -3 - -1 5-8 -3 - -1 (Surveyed8 benchmark) (Surveyed benchmark) 5- -3 - -1 5-8 (Forecasted land subsidence) -3 - -1 5-8 (Forecasted land subsidence) Source: Royal (RTSD) and Panya Consultants' forecast Source: Royal Thai Survey DepartmentThai Survey Department (RTSD) and Panya Consultants' forecast Source: Panya Consultants (2009). Figure 2.1-3 Land Elevation due to and 2050 due to Figure 2.1-3 Land Elevation in 2002 and 2050in 2002 Land SubsidenceLand Subsidence 2.2 CLIMATE 2.2 CLIMATE 2.2.1 General 2.2.1 General distribution patterns (1995). There is an implicit remain under water for over a month. The increase The climate of the Chao Phraya River Basin belongs to the The average annual The climate of the Chao Phraya River Basin belongs to the tropical monsoon.tropical monsoon. The average annual variability (timing to 1,600 mm and to be mm and registering areas, where existing assumption that the underlyingover the basin is 1,130 mm, varyingis likely to 1,600in higherwestern higher in the rainfall over the basin is 1,130 mm, varying from 1,000 rainfall from 1,000 registering the in the According basin. is northeastern region of the basin.in 2050. to the rainfall pattern, about 85% of the average annual average annual to be insuf- to the rainfall pattern, about 85% of the northeastern region of the This According flood protection infrastructure is likely and location) will remain the same rainfall occurs between May and October. Tropical cyclones occur between September and September and rainfall occurs between May and October. Tropical cyclones occur between October and mayOctober and may In this the basin. strike the basin. strike case, rainfall continues for aThere is alsolong increase in a ficient. continues of a an a conservative in the sense that increased variabilityIn this case, rainfalllong period for time in period of time in the maximum relatively wide area. The peak registered in October, the end October, the end of relatively wide area. The peak river discharge is river discharge is registered in of the rainy season, the rainy season, and droughts) around the mean would this high tide in this temperature ranges base ranges (worse floods andsevere flood and severe flood damagehigh tide in with period. The meanperiod. The mean temperatureyear and 2050 for damage may arise with may arise water depth between the from 26oC to 31oC. Its 26oC to 31oin Its maximum April and its in April and its minimum The periods. Figure 3.3 for from maximum C. the study. events of is in December. is in generate greater losses than estimated temperature is in temperature isminimumdifferent return December. The evaporation the basin is normally at its highest in at its highest in in October evaporation (Class-A Pan) in (Class-A Pan) in the basin is normallyApril and lowest April and lowest in October with an average annualan average annual value of about 1,700example, shows a comparison of the maximum with value of about 1,700 mm. mm. depths of flooding for the 1-in-30-year flood under MAInFIndIngSFRoM 2008 and the 2050 A1FI scenario. The implication is hydRologICAlAnAlySISAnd that the people living in Bangkok will be facing more gISMAppIngFoRbAngKoK 2-2 frequent events that significantly disrupt daily life. 2-2 A flood frequency graph obtained by plotting the inundated area versus return frequency for the Flood-pronearealikelytoincreasein Bangkok City case study results for the 2008 and 2050 bangkokinthefuture The hydrological analysis shows that for events of b TAble3.4 angkokInundatedArea different return periods, the area inundated will underCurrentConditions increase in 2050 for both the B1 and the A1FI sce- andFutureScenarios narios (Table 3.4). For instance, the current (2008) estimated annual Scenario inundated area in Bangkok and Samut Prakarn is 2008 2050 B1 2050 A1FI about 550 km2, which will increase to 734 km2 in 2050 Frequency (km2) (km2) (km2) under the A1FI scenario for a 1-in-30-year flood. That 1/100-year 737 893 927 is an increase of 184 kms, or approximately a 30 per- 1-in-30-year 550 719 734 cent increase in the area inundated. The inundation 1-in-10-year 359 481 518 could be for varying depths and varying number of days, but about 7 percent of these provinces could Source: Panya Consultants (2009), Appendix k, Table K2.4-1. 28 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Climate Change Impact and Adaptation Study for Bangkok Metropolitan Region Final Report Chapter 3: Model Development and Simulation M FIguRe3.3 aximumwaterdepthfor1-in-30-yearevent,2008and2050,A1FI 620000 640000 660000 680000 700000 620000 640000 660000 680000 700000 1580000 1580000 1580000 1580000 1560000 1560000 1560000 1560000 Pathum Thani Pathum Thani ! ( ( ! 1540000 1540000 1540000 1540000 Nonthaburi Nonthaburi ( ! ! ( Nakhon Pathom Nakhon Pathom ( ! ( ! Bangkok Bangkok ( ! ( ! 1520000 1520000 1520000 1520000 Samut Prakan Samut Prakan ( ! ! ( Samut Sakhon Samut Sakhon 1500000 1500000 1500000 1500000 ! ( ( ! 1480000 1480000 1480000 1480000 620000 640000 660000 680000 700000 620000 640000 660000 680000 700000 Legend ( ! Province -3 - -1 5-8 0 - 10 Ê 0 2 4 8 12 16 Legend ( ! Province Province Boundary -3 - -1 -1 - 0 5-8 8 - 10 0 - 10 10 - 50 0 2 4 Ê 8 12 16 Province Boundary -1 - 0 8 - 10 10 - 50 Kilometers Kilometers River/Canal Network 0-1 10 - 15 50 - 100 River/Canal Network 0-1 10 - 15 50 - 100 Dike 1-2 15 - 20 100 - 200 Dike 1-2 15 - 20 100 - 200 Elevation (m.MSL) 2-3 20 - 25 200 - 300 Max. Water Depth Elevation (m.MSL) 2-3 20 - 25 200 - 300 -7 - -5 3-4 > 25 300 - 400 Max. Water Depth -7 - -5 3-4 > 25 300 - 400 C2050-LS-SR -5 - -3 4-5 Max. Water Depth (cm) > 400 C2008-T30 -5 - -3 4-5 Max.Water Depth (cm) > 400 -A1FI-T30 Source: Panya Consultants' calculation Source: Panya Consultants (2009). Figure 3.2-1 Maximum Water Depth of Case C2008-T30 and C2050-LS-SR-A1FI-T30 Figure 3.2-2 illustrates the maximum water surface profile along the Chao Phraya River from the FIguRe3.4 return periods. Comparing A1FIriver mouth to Bang Sai district, Ayutthaya at 10, 30 and 100-year floodangkokFloodhazard further demonstrates how the flood hazard will b Case C2008-T100 Bangkok. As the graph in water level will increase dramatically inand C2050-LS-T100, it reveals that the maximum Relationship be reduced due to land subsidence flood event in 2008 Figure 3.4 below illustrates, aof about 0.20 m, considering the sea level at the river mouth is not 1,000 an inundated area of Case 2 has a frequency with increased. Comparing625 kmC2050-LS-T100 and C2050-LS-SR-A1FI-T100, the maximum water 900 level return increased 1-in-50 years. By 2050 of 0.02 or will be period of due to the increasing of flood water from the upper basin and the sea level 800 ~39% increase C2050-LS-SR-A1FI-T100 rise. Comparing Caseflood events inundating and C2050-LS-SR-SS-A1FI-T100, the maximum Inundated area (km2) 700 Increased 625 under the A1FI scenario, water level will be increased at the river mouth due to storm surge, but the effect will appear up to 600 frequency Increased areas of 625 km2 would be expected to occur with an 500 about 50 km from the river mouth. 400 450 Intensity increased frequency of 1-in-15 years. The corollary of 300 this is that the intensity of flooding for a 1-in-15-year 200 event would increase from about 450 km2 to 625 km2. 100 1/50­year 1/15­year 0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Flood frequency populationexposedtofloodingwillincrease 2008 2050 A1FI for1-in-30-yearfloodforbothA1FIandb1 scenarios Source: Panya Consultants (2009). Severe floods inundate the lower Chao Phraya floodplain in which Bangkok is situated. The huge a "bottleneck" to the discharge reaching the Gulf volume of water (over 31,000 million cubic meters of Thailand. This can leave many areas flooded for for a 1-in-30-year event) associated with these more than 30 days. The flood protection schemes for floods can take months to drain as Bangkok acts as Bangkok are generally designed to protect against estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 29 1-in-30-year floods. There are steep increases in Impactofthefloodsonpeoplelivingin persons affected as inundation levels exceed the condensedhousing 2008 1-in-30-year design standards. For example, the number of persons who would be affected by Figures 3.5a and 3.5b show the impact of flooding 1-in-100-year flood event in 2008 is nearly double under the 1-in-30-year A1FI flood in 2050 in com- the number affected by a 1-in-30-year flood, as parison to 2008, with an overlay of poor housing shown in Table 3.5. Similarly, the number of persons in Bangkok. It shows increased flooding in several affected in 2050 by a 1-in-30-year event will rise sharply condensed housing areas where the poor live. As for both B1 and A1FI scenarios, with 47.2 percent and the study points out, " about 1 million inhabitants 74.6 percent increases, respectively. of Bangkok and Samut Prakarn will be affected by the A1FI climate change condition in 2050. One in eight of the affected inhabitants will be from the condensed Areasinbangkokvulnerabletoflooding housing areas where most live below the poverty level. Although Bangkok's flood embankment and One-third of the total affected people may be subjected to pumped drainage will continue to offer substantial more than a half meter inundation for at least one week. protection for the interior area, land outside the em- This marks a two-fold increase of that vulnerable bankment to the north and the southwest is likely to population. The impact will be critical for the people experience significantly worse flooding. These are living in the Bang Khun Thian district of Bangkok areas which are undergoing housing, commercial, and the Phra Samut Chedi district of Samut Prakarn" and industrial development. Within BMR, Bangkok (Panya Consultants 2009). and Samut Prakarn are the two provinces that are most affected by climate change. In a C2050-LS-SR- Keysectorslikelytobeaffectedbyflooding SS-A1FI-T30 scenario, almost 1 million people in in2050butbuildingsandhousingmost Bangkok and Samut Prakarn would be impacted affected by floods. The impact will be profound for people living on the lower floors of residential buildings. The most significant impacts will be felt by the Nevertheless, people living on higher floors in the residential, commercial, and industrial sectors with Bang Khun Thian district of Bangkok and the Phra more than a million buildings being affected by a Samut Chedi district of Samut Prakarn might also 1-in-30-year event in 2050. About 300,000 build- be impacted. District-wise, Don Muang district in ings in areas to the west of Bangkok--like Bang north Bangkok has the highest number of people Khun Thian, Bang Bon, Bang Khae, Phra Samut, affected by floods (approximately 90,000) owing to and Chedi districts--will also be impacted. For its higher population density. In the western part of the 1-in-30-year event, approximately 1,700 km the Chao Phraya River, about 200,000 people in Bang of roads would be exposed to flooding. The Lat Khun Thian, Bang Bon, Bang Khae, and Nong Kham Krabang water supply distribution station and districts might be impacted. The maps in Figure 3.5a the Nongkhaem solid waste transit station would and 3.5b below show the differences in flooding for both be subject to 50­100 cm inundation, although the 1-in-30-year event, currently and in 2050. the main water and wastewater treatment facilities are protected. In addition, 127 health care facilities would be subject to flooding, with inundation levels e Table3.5 xposureofbangkok ranging from 10 to 200 cm. populationtoFlooding Population affected for >30 days Stormsurgeandsealevelrisehave Frequency 2008 2050 B1 2050 A1FI relativelysmallroleincontributingto 1-in-100 year 1,002,244 1,187,803 1,271,306 floodinginbangkok 1-in-30 year 546,748 805,055 954,389 Analysis carried out in the Bangkok study found Source: Panya Consultants (2009). a linear relationship between future precipitation 30 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport FIguRe3.5a A ffectedCondensed(poor)CommunityofCaseC2008-T30 A FIguRe3.5b ffectedCondensed(poor)CommunityofCaseC2050-lS-SR-SS- A1FI-T30 Source: Panya Consultants (2009). and flood volume in the Chao Phraya River. How- Samut Prakarn will increase by about 2 percent ever, it found that flood peak discharge in the Chao due to a storm surge striking the western coast of Phraya River will increase by a larger percentage the Gulf of Thailand. than precipitation, due to unequal travel times of floods from upstream catchments. Further, while storm surges and sea level rise are important, they eSTIMATIngClIMATe-RelATed will have less effect on flooding in Bangkok. While IMpACTSInMAnIlA storm surges do occur in the Gulf of Thailand and contribute to flooding the BMR area, it was esti- The Manila study (Muto et al. 2010) focuses on mated that the flood-prone area in Bangkok and Metro Manila and includes Manila and 16 munici- estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 31 palities. It is a low-lying area crisscrossed by the tion of 11 million, it is ranked as one of the most Pasig River and its tributaries, which flows north- densely populated cities in the world. The Pasig westward from Laguna de Bay, the largest lake in River bisects the city in the middle before draining the Philippines, to Manila Bay in the west. Metro into Manila Bay. Metro Manila, the broader urban Manila lies on a swampy isthmus and is marked agglomeration, has the highest per capita GDP in by three quite diverse hydrological characteristics. the country. In 2007, its population was estimated to These include the Pasig Mariquina area, Kamanava be over 20 million people. In 2008, it was ranked as area, and the west of Managhan area facing the the 40th richest urban agglomerations in the world, Laguna de Bay (Figure 3.6). The Pasig Mariquina with a GDP of $149 billion, indicating the economic area has several river systems and a catchment area importance of Metro Manila. Manila is also a major of 651 sq kms. It includes 10 cities/municipalities tourist destination, with over 1 million tourists visit- of Metro Manila. The Kamanava area along the ing the city each year. low-lying coast is flat, prone to typhoons, and has elevations ranging from around sea level to 2­3 unregulatedexpansionintofragileareas meters above sea level. Before the 1960s, it mainly andinformality consisted of lagoons, but has been filled up and cur- rently comprises commercial districts, residential With large in-migration and rapid population areas, and fishponds. The West of Managhan area growth, the city expanded to the suburbs, sur- is 39 square kms and covers five cities. There are rounding municipalities, and to areas risky for a number of drainage channels draining into the habitation (e.g., swampy areas, near or above este- Laguna Lake or Napindan River. ros or water canals, along the river or earthquake fault lines, etc.). A large part of the developments ImportanceofMetroManilatotheregional in informal settlements are unregulated. Many economy structures are built on dangerous and risky areas, such as near the seashore or flood zone, or on Like Bangkok, Manila is also a capital city and a ground prone to landslides. Socioeconomic factors major economic center. It is located on the eastern like land use practices, infrastructure development, shore of Manila Bay, on the western side of the building standards/codes and practices, and urban National Capital Region. It is a central hub of the development policies and programs have greatly thriving Metropolitan Manila area. With a popula- shaped the settlement and building patterns of the city. These forces generate an environment that poses high risks to residents and infrastructure FIguRe3.6 M etroManilaandits alike, especially in the low-lying flood-prone areas. watershed As in other cities in developing countries, there are significant wealth disparities in Manila that are 121°00' 121°20' 0 KILOMETERS 5 PASIG ­ MARIKINA RIVERS PHILIPPINES reflective of the country as a whole, with 97 percent San Jose Del Monte PROJECT RIVERS of GDP controlled by 15 percent of the population WATERSHED SELECTED CITIES REGION CAPITAL (WWF 2009). Within Metro Manila, 10.4 percent of NATIONAL CAPITAL Meycauayan Rodriguez the population in lives below the poverty line, ac- San Mateo cording to 2006 official statistics.32 Approximately R iver This map was produced by the Map Design Unit of The World Bank. a The boundaries, colors, denominations and any other information kin shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any ari 14°40' 61 percent of the population in Metro Manila does endorsement or acceptance of such boundaries. 121°20' M QUEZON Caloocan CITY Marikina Bosoboso 120°E 125°E Pasi g not have access to basic services, according to the REGION CAPITALS Tuguegarao River NATIONAL CAPITAL REGION BOUNDARIES San Juan San Fernando INTERNATIONAL Baguio BOUNDARIES MANILA Antipolo 2008 Philippine Asset report report card (cited 15°N San Fernando 15°N MANILA Quezon Mandaluyong Pasig Area of Map Legaspi Philippine Manila Sea in Manila study). There is an estimated housing Pateros Bay Tacloban Iloilo Taguig Cebu 10°N 10°N Parañaque Pililla Sulu Sea Iligan Butuan Cagayan de Oro backlog of almost 4 million. Laguna Binangonan Zamboanga Laguna Cotabato JANUARY 2010 IBRD 37476 Bacoor Bay Davao Bay M A L AY S I A 121°00' 120°E Celebes Sea 125°E 5°N 5°N 32 http://www.nscb.gov.ph/poverty/2006_05mar08/ta- Source: Philippines map IBRD 37476. ble_2.asp 32 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Manila'scurrentclimate--Tropicalwith TAble3.6 M anila:MonthlyAverage distinctwetanddryseasons Temperatureand precipitation Figure 3.7 shows the four main climate regimes in the Philippines and the relative frequency of Temperature Temperature Average typhoons making landfall. Manila falls into Type Average Low Average High precipitation 1, which is defined as having two distinct seasons, Month (oC) (oC) (mm) with the dry season running from November to January 24° 30° Negligible April and the wet season occurring in the other February 24° 30° 12.1 months. About 16 percent of the typhoons crossing March 25° 32° 9.1 the Philippines pass in the general vicinity of Manila April 27° 33° 15.9 and its nearby areas. The average temperatures May 27° 33° 133.0 and precipitation in Manila are shown in Table 3.6. June 26° 32° 150.8 Manila has an average annual precipitation of about July 26° 31° 292.9 1,433 mm, but there is significant variability. August 26° 31° 305.8 September 26° 31° 237.5 Mainclimate-relateddriversofflooding-- October 26° 31° 137.2 drivenbysingle-stormevents,usually relatedtotyphoons November 25° 31° 81.3 December 24° 30° 58.0 Unlike Bangkok, where floods are generated by Total 1433.6 heavy seasonal precipitation over 1 to 3 months, Source: http://weather.uk.msn.com/monthly_averages. extreme flood events in Manila are caused by heavy aspx?wealocations=wc:RPXX0017 precipitation events over 1 to 3 days generally associ- ated with typhoons and storm surge. The watershed area for Manila's Pasig-Marikina River System is ment. This is relatively small in comparison to the 651 km2, which includes the San Juan River catch- watersheds of other cities in the study. The precipi- tation pattern causing flooding is often associated with strong winds and flash flooding, which can be devastating for informal housing located along FIguRe3.7 d ifferentClimatic drainage ways. Seeking refuge during such events is Regimesinthephilippines extremely difficult and hazardous to the population. Other climate-related factors contributing to flooding include a combination of high tide, excess runoff from rivers, heavy rains, and sea level rise. A schematic of the watersheds around Manila is shown in Figure 3.8. non-climaterelateddriversofflooding Even though land subsidence is an important issue ! in Metro Manila, it was not possible to consider it for the hydrological simulation for the Manila study because of lack of availability of reliable data.33 The 33 One difficulty in Metro Manila is that the station that is supposed to measure land subsidence of Manila de Bay is located on one of the old piers of Manila Port, which itself Source: Muto et al. (2010). is sinking gradually. estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 33 severe flooding caused by Typhoon "Ondoy" in measures undertaken as part of these projects are September 2009 has raised public awareness of the primarily infrastructure projects involving attenuat- underlying causes of flooding. Recent analysis points ing peak flood flows in Laguna de bay, improving to mostly anthropogenic causes behind the extreme conveyance of flood waters through Manila to Ma- flood event, including (a) a decrease in river channel nila Bay, and preventing overbank spillage where the capacity through encroachment of houses, siltation flood profile is higher than the adjacent river bank. from deforestation, and garbage; (b) disappearance of 21 km of small river channels; (c) urbanization accel- Climatechangefloodsimulationparameters erating runoff concentration and reducing infiltration losses; (d) loss of natural retention areas; and (e) land To simulate the future climate-change-induced flood- subsidence. Among these, land subsidence is the least ing, the city study applied the statistical downscaled understood but important cause. It is being driven results from the University of Tokyo's Integrated by groundwater pumping and possibly geologic Research System for Sustainability Science (IRS3) processes associated with the West Marikina Valley for temperature and extreme precipitation increases. Fault (Siringan 2009). Land subsidence continues For SLR in 2050, the study applied approximately decades after the groundwater pumping stops, as 50 percent of the increase in sea level rise in 2100 illustrated by the Bangkok city case study. projected in the IPCC AR4.35 Historic storm surge in Manila Bay was calculated at 0.91 m during major Floodcontrolandmitigation typhoons; for 2050, a 10 percent factor increase was projects--1990Masterplan applied based on Ibaraki University methodology,36 bringing the storm surge estimate for 2050 to 1.0 m. Flood control and mitigation projects have been planned and implemented in Metro Manila over the years.34 A key feature of flood control in Manila 34 See Manila study for a list of planned and ongoing flood is the Mangahan Floodway, which was constructed control projects and studies. 35 In terms of boundary conditions for the hydrological in 1985 and diverts flood flows from the Marikina model, the SLR values were added to high tidal level; that River into Lake Laguna as a method of attenuating is, mean spring high water level in Manila Bay was set as the base line water level for the flood simulation, to which peak discharges and protecting Manila's urban ar- was added an SLR increase in accordance with the two eas from excessive inundation. JICA supported the IPCC scenarios. 36 The university provided a detailed technical methodol- development of a Flood Protection Master Plan in ogy for calculating the increase in storm surge based on an 1990, which still serves as a basis for current and fu- analysis of historic surges and forecasted increased typhoon ture flood protection planning. The flood protection intensity. FIguRe3.8 M ajorwatershedanddrainageAreasofManila Pasig-SanJuan Marikina West Mangahan Source: Muto et al. (2010) 34 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport M TAble3.6 anilaClimateChangeparameters Temperature Rise (°C) Sea Level Rise Increased Rate of Rain- Storm Surge Height Simulation Case (downscaled) (cm) (global) fall 24-hr event (%) (m) at Manila Bay 1 Status quo climate (no change in climate 0.0 0.0 0.0 0.91 dependant variables) 2 B1 with no change in storm surge 1.17 19 9.4 0.91 3 B1 with strengthened storm surge level 1.17 19 9.4 1.00 4 A1FI with no change in storm surge 1.80 29 14.4 0.91 5 A1FI with strengthened storm surge level 1.80 29 14.4 1.00 Source: Muto et al. (2010). The factors applied are summarized in Table 3.7. The tation of the Master Plan would be halted in 2008, 24-hr precipitation factor increase is reasonable as it the base year--referred to as "existing" infrastruc- is close to the time of concentration of flood flows for ture and abbreviated as "ex." The other alternative the Manila watershed. Similarly, SLR is conservative, was that the full plan would be implemented by since it is taken from the IPCC estimates and does 2050--referred to as "business as usual" and abbre- not include polar ice cap melt. viated as "BAU." To do this and include the climate change factors noted above, the team developed 22 Infrastructurescenariosassumedinbase simulation runs (Muto et al. 2010). yearandin2050 Assumptionsmaderegardingpopulation For the simulation runs, the study examined two increase,landuse,urbanizationin2050 flood infrastructure alternatives based on the 1990 Master Plan. The first was to assume the implemen- Numerous assumptions regarding what Metro Ma- nila might look like in 2050 compared to the base year were made in the study (Box 3.2). w box3.2 hatdoesMetroManila looklikeintheFuture? FIndIngSFRoMThe For the hydrological mapping, the team used 2003 topography data hydRologICAlAnAlySIS for the structure of the city. It was assumed that in terms of number AndgISMAppIngFoRMeTRo of buildings and road network, Metro Manila in 2050 would look very similar to what it is today. Further, while the team intended MAnIlA to use future land use in the hydrological and GIS mapping, this was not possible due to logistical reasons. Regarding population Significantincreaseinareaexposedto growth, Manila uses a linear extrapolation of current trends to project floodingbyextremeevents(1-in-100-year population to the future. The city is expected to be a megacity with event)in2050underb1andA1FIscenarios a population greater than 19 million by 2050. Commercially grow- ing sections of Manila are expected to see a decline in population, The flood pattern in Manila for 2050 is likely to be with growth occurring in other parts of the city. Areas such as the Pasig-Marikina river basin are seeing a decline in population, which significantly affected by the extent to which the 1990 is expected to continue; growth will likely continue to occur in West Master Plan is implemented. Table 3.8 shows that the Mangahan and Kamanava. Future regional GDP is not projected, but area exposed to flooding from a 1-in-100-year event a real growth rate of 5 percent is applied to a variety of economic under current conditions will increase from 82.62 variables and damage costs. km2 to 97.63 km2 under the 2050 A1FI 1-in-100-year Source: Muto et al. (2010). event, an increase of 18.2 percent. This represents a significant increase in the inundated area. With the estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 35 TAble3.8 M anila:ComparisonofInundatedArea(km 2 )with1-in-100-yearfloodfor 2008and2050ClimateChangeScenarioswithonlyexisting InfrastructureandwithCompletionof1990Masterplan No change to existing infrastructure All 1990 Master Plan infrastructure implemented 2008 (km ) 2 2050 B1 (km )2 2050 A1FI (km ) 2 2008 (km2) 2050 B1 (km2) 2050 A1FI (km2) Pasig-Marikina 53.73 63.19 67.97 29.14 40.09 44.14 West Mangahan 10.65 11.09 11.42 7.79 8.16 8.30 KAMANAVA 18.24 18.24 18.24 -- -- -- Total 82.62 92.53 97.63 36.93 48.25 52.44 Relative increase 12.0% 18.2% 30.7% 42.0% Source: Muto et al. (2010). implementation of the 1990 Master Plan, infrastruc- C FIguRe3.9 omparisonofpopulation ture in the inundated areas will be reduced consider- AffectedbyFlooding ably for the current flood hazard (i.e., an inundated underdifferentScenarios* area of only 36.93 km2 from a 1-in-100-year event in 3,000,000 2008), but the infrastructure will not afford the level 2,500,000 of protection in 2050 for the return periods on which 2,000,000 the designs are based. Climate change impacts under 1,500,000 the A1FI 1-in-100-year storm event would increase 1,000,000 the inundated area by 42 percent (from 36.93 km2 to 500,000 52.44 km2) compared to the base year 2008. The main 0 EX SQ EX B1 EX A1FI BAU SQ BAU B1 BAU A1FI point is that that the master plan will not offer the planned level of protection. Source: Muto et al. (2010). * Scenarios include 1-in-100-year event (P100) in 2008 (SQ) with the 2050 B1 and 2050 A1FI scenarios with only the existing infrastructure (EX) and with the full Increaseinpercentageofpopulation implementation of the 1990 Master Plan (business as usual or BAU). exposedtofloodingunderhighandlow emissionsscenarios population density that are likely to face serious As would be expected from the significant increase flooding risk (Figure 3.10). These include Manila in inundated area under the B1 and A1FI climate City, Quezon City, Pasig City, Marikina City, San change scenarios, the number of people affected Juan, and Mandaluyong City. by the floods increases significantly as well, as illustrated in Figure 3.9. Thus, for a 1-in-100-year AreasinMetroManilalikelytobeathigh flood in 2050, under the AIFI scenario more than riskfromfloodingduetoextremeeventsin 2.5 million people are likely to be affected, assum- 2050 ing that the infrastructure in 2050 is the same as in the base year. In a scenario where the flood protec- The hydrological analysis shows that some local gov- tion infrastructure under the 1990 Master Plan is ernment units (LGUs) will be much more affected implemented, the level of protection expected by its than others under different scenarios. As Figure 3.11 implementation will not be achieved under the A1FI shows, municipalities in the Pasig Marikina River 1-in-100-year storm event, where approximately 1.3 basin (namely Manila, Mandaluyong, and Marikina) million persons would be affected. and Kamanava areas (namely Malabon and Navotas) An overlay of population density maps with are more likely to be vulnerable to flooding com- GRID data on inundation shows areas of high pared to those in the West Managhan area. 36 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport FIguRe3.10 A reasofhighpopulation Impactonroads,railnetwork,powerand densityandwithhigh watersupplyfacilitiesandtransportation RiskofInundationunder A1FIScenario The hydrological analysis shows the length of roads that are likely to be flooded (in terms of depth and duration) under different scenarios. Thus for instance, under both emissions scenarios--A1FI and B1, more roads are flooded under "existing infrastructure scenario" where no additional infra- structural improvements are made, compared to a scenario where there is continued implementation of flood protection infrastructure in 2050 (Table 3.9). As a reference, it is useful to note that a depth of more than 26 cms is considered impassable for most vehicles in Manila. The hydrological analysis also shows specific locations where the power distribution system in Manila is vulnerable. It notes that the elevated rail system in Manila, although above the flood levels, is vulnerable to loss of power from the power utility Merelco, and that a number of the stations would be inaccessible during extreme floods. The impact of high winds in these situations was not detailed. The consultants held discussions with the water supply providers who indicated that their facilities were not vulnerable to floods. Source: Muto et al. (2010). FIguRe3.11 A reasathighRiskfromFloodingunderdifferentScenarios 80 70 60 50 40 % 30 20 10 0 City Mandaluyong Manila Marikina Quezon San Juan Passay Makati Pasig Taguig Pateros Kalookan Malabon Navotas Region Pasig ­ Marikina River Basin West Mangahan Area KAMANAVA Area Ex­SQ Ex­B1 Ex­A1F1 BAU­SQ BAU­B1 BAU­A1F1 Source: Muto et al. (2010). estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 37 TAble3.9 A ffectedlengthofRoadbyInundationdepth Road Length by Inundation Depth (kms) 8­20 cm 21­50 cm Above 50 cm Flood Scenario (Ex) Major Minor Major Minor Major Minor Total Status Quo 4.5 3.9 22.1 23.8 31.9 39.8 125.9 B1 5.4 9.7 13.6 15.1 47.9 55.6 147.3 A1FI 5.3 6.9 14.6 18.2 53.6 60.3 158.9 Road Length by Inundation Depth (kms) 8­20 cm 21­50 cm Above 50 cm Flood Scenario (BAU) Major Minor Major Minor Major Minor Total Status Quo 3.78 4.33 6.40 10.45 7.45 13.42 45.82 B1 7.24 8.15 9.54 15.73 12.07 20.82 73.55 A1FI 9.45 9.05 12.62 16.28 14.97 25.63 87.99 Source: Muto et al. (2010). Impactofexistingflood-relatedeventson additional levels by households to escape rising thepoor37 flood waters, use of styrofoam boats, installation of pumps or "bombastic" to help drain flood water In Metro Manila, there is a strong link between (done in some municipalities by the mayor). where the urban poor reside and their vulnerability The survey details many of the public health to flooding. Interviews with 300 poor households hazards associated with flooding . However, for some (in 14 communities in three river basins) to assess of the respondents, the consciousness of health risks the impact of floods (conducted as part of the posed by flooding is not very high. They say "Hindi Manila study) show that they primarily live in ka naman namamatay dahil sa baha!" (You do not environmentally fragile, low-lying areas and/or die from floods or rising waters here!). For example, swamps or wetlands that are highly vulnerable to the risk of catching an infection like the deadly ef- storm and tidal surges. Households interviewed fects of rat's urine (i.e., leptospirosis) is not high in have a median monthly income of about 44 pesos/ their consciousness. Following Typhoon "Ondoy" day (less than $1/day). Further, most of them live in September 2009, however, there was an epidemic in slum or squatter settlements and do not have of leptospirosis that infected over 2,000 people, with tenure security or access to basic services such as over 160 people dying and many requiring dialysis. clean water, electricity, sanitation, and drainage. About two-thirds reported suffering regular losses due to typhoons, floods, and storm surges. Twenty- eSTIMATIngClIMATe-RelATed seven percent of the households interviewed have IMpACTSInhoChIMInhCITy, substandard latrines (dug latrines) or none at all vIeTnAM (and use neighbor's latrine or direct to the river/ sea). Those who do have toilets complained about HCMC is a tropical coastal city located on the toilets overflowing and garbage being carried dur- estuary of the Saigon­Dong Nai River system of ing floods contributing to waterborne diseases. Vietnam. While not a capital city, Ho Chi Minh City About 65 percent of the households buy water from neighbors or suppliers. During floods, household 37 A detailed analysis of the impact of future flooding on the expenditures on water and transportation increase. poor was not undertaken as part of the Manila study and Coping strategies include addition of stilts or needs to be carried out as a follow-up to this report. 38 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport is also a key economic and financial center and a between 2000 and 2007. During the same period, ag- core part of the Southern Economic Focal region in riculture and related activities have grown relatively Vietnam. The city's official population is estimated slowly at a rate of 4.8 percent annually (ADB 2010). to be about 6.4 million people. The actual current population, which includes temporary and unreg- natureofpovertyinhCMC istered migrants, is estimated to be 7­8 million. The population of HCMC has been growing at a rate In 2006, HCMC had an overall poverty rate of 0.5 of 2.4 percent per year, which is a faster rate than percent (GSO Vietnam 2006).38 While this rate is the anywhere else in the country. The city accounted for lowest in the country, the absolute number of poor over 23 percent of the country's GDP in 2006 (ADB in the city is still high at between 30,000 and 40,000 2010). Rapid economic growth--11.3 percent annu- persons. In addition, the number of households liv- ally between 2000 and 2007--has been the central ing in inadequate housing and poor environmental driver behind the city's expansion, as the increasing conditions is likely to be much higher than the of- number and magnitude of income earning oppor- ficial poverty rate suggests. If unregistered migrants tunities attract migrants from throughout Vietnam. are primarily poor, then the poverty rate would also Key economic growth sectors include industrial be higher due to the large numbers of migrants miss- and service sectors, which have grown at a rate of ing from official statistics. Spatially disaggregated 11.9 percent and 11.1 percent annually respectively data for the city shows that there is significant varia- tion in poverty levels in different parts of HCMC. In general, rural districts have higher poverty levels compared to more urban districts. The highest pov- TAble3.10 h CMCdistrictpoverty erty levels are in the relatively sparsely populated Rates,2003 rural districts of Can Gio and Nha Be to the south of the city toward the coast. These are also the two Urban districts Rural districts districts that are the most prone to flooding in the Poverty Rate Poverty Rate city (Table 3.10). District (%) District (%) District 1 2.4 Cu Chi 6.9 Climateandprecipitationinthedongnai District 2 4.5 Hoc Mon 6.9 Riverbasin District 3 2.8 Binh Chanh 4.4 District 4 5.6 Nha Be 12.9 HCMC has a tropical monsoon climate with pro- District 5 3.8 Can Gio 23.4 nounced wet and dry season variations in precipita- District 6 5.5 tion. There is a large variation in monthly precipita- District 7 3.0 tion, but relatively constant daily temperatures with District 8 7.4 an annual average in the range of 26­27oC and a variation between months in the range of 4° to 5oC. District 9 5.0 As shown in Table 3.11, precipitation is highest in District 10 3.0 the months from May to November, accounting for District 11 4.6 90 percent of the annual rainfall. Typhoon events District 12 6.3 can pass over HCMC. The dry season runs from Go Vap 6.9 December to April and can cause drought condi- Tan Binh 5.5 tions regularly. Extreme droughts have occurred in Binh Thanh 5.0 1993, 1998, and 2002. Phu Nhuan 3.7 Thu Duc 7.8 38 This is calculated on the basis of the expenditure typical- HCMC average poverty rate 5.4 ly required to meet a minimum daily calorific requirement of 1,700 Kcals. A number of qualifications need to be made, Source: Inter-ministerial Poverty Mapping Task Force, 2003, cited in ADB (2010). as the official figure does not include migrants. estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 39 TAble3.11 oChiMinhCity:Monthly h it include seasonal monsoonal rainfall and tides. AverageTemperatureand Extreme flooding occurs when tropical storms and precipitation storm surges combine with tidal influences and monsoon rainfall to create extreme weather condi- Temperature Temperature Average tions. Storm surge has been identified as a key driver for Average Low Average High precipitation extreme events in HCMC. Further, warmer tempera- Month (oC) (oC) (mm) tures in the South China Sea are expected to increase January 22° 32° 5.6 the frequency of tropical storms and typhoons that February 23° 33° 4.1 land in southern Vietnam. SLR is likely to have an March 25° 34° 14.2 important influence on the inland reach of tidal flooding April 26° 35° 42.4 in HCMC. Currently 154 of the city's 322 communes May 26° 34° 138.9 and wards have a history of regular flooding, af- June 25° 33° 209.8 fecting some 971,000 people or 12 percent of the July 25° 32° 204.7 HCMC population (ADB 2010). With an extensive August 25° 32° 186.7 area subject to regular flooding, it is not surprising that more extensive flooding can be induced by September 25° 32° 178.3 tides, storm surges, heavy rains on the city directly October 24° 31° 222.0 or in the upper watershed, or a combination of those November 23° 32° 88.9 events associated with a typhoon. December 22° 31° 23.6 Total 1313.7 Source: http://weather.msn.com/monthly_averages.aspx?&wealocations=wc%3aV FIguRe3.12 h CMC:Frequently MXX0007&q=Ho+Chi+Minh+City%2c+VNM&setunit=C FloodedAreasunder CurrentConditions lowelevationtopographycontributesto flooding Much of HCMC is located in low-lying lands that are prone to frequent flooding associated with heavy rains or even just high tides. As reported in the HCMC case study (ADB 2010), about 40­45 percent of land cover in HCMC is at an elevation of between 0 and 1 m. The mangrove forests of Can Gio district in HCMC are a key natural resource and provide considerable storm protection. Figure 3.12 shows the areas of HCMC that flood frequently; areas colored in blue and red indicate areas flooded by tides (and hence is salty) and rain (freshwater) respectively. Mainclimaticfactorscontributingto flooding HCMC is subject to both regular and extreme flooding. Regular floods refer to floods that occur throughout the year on a daily and seasonal basis. Some of the main climatic factors that contribute to Source: ADB (2010). 40 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport box3.3 hCMCin2050 Future land use and development projections for HCMC were guided by a number of sectoral and urban development plans and strategies. As per the adjustment plan, the population of HCMC in 2025 is expected to be about 10 million; economic growth in HCMC is expected to increase 8 percent annually in the period from 2026 to 2050 (ADB 2010). Land use for residential purposes is expected to grow by 28 percent and for interior traffic by 215 percent. The vision underlying the adjustment plan includes distributing a population of about 10 million in several satellite cities. It assumes growth in both the industrial and service sectors until 2020. In accordance with this vision, a substantial amount of land in HCMC is being reallocated for industrial purposes. An industrial master plan prepared by the Ministry of Industry estimates this to be around 11,000 ha by 2025. While the adjustment plan recognizes the importance of green space, it also estimates that open areas will decline by about 25 percent. The adjustment plan refers to climate change but does not address measures needed for adaptation. Population projections for 2050 were developed by the study team based on government data and Master Plan. This resulted in two population estimates, one low estimate of 12 million people, and a high estimate of 20.8 million people by 2050. The high population estimate is likely to be more realistic. Future population distributions were also estimated based on current patterns and trends and are expected to be influenced by the hollowing out of the city center, higher growth and increased population densities in peripheral areas, and the development of satellite cities. In 2006, HCMC had an overall poverty rate of 0.5 percent. While this rate is the lowest in the country, the absolute number of poor in the city is still high, at between 30,000 and 40,000 people. The HCMC poverty rate was projected to remain significant as climate change impacts take hold. Rapid economic growth (11.3 percent between 2000 and 2007) has been the central driver behind the city's expansion, as the increasing number and magnitude of income earning opportunities attract migrants from throughout Vietnam. The HCMC study bases its GRDP numbers on an "Adjustment Plan" developed by the government. The master plan, which was developed for 2025, contained projections to 2050 for GDP. Accordingly, economic growth in HCMC is expected to be 8.7 percent between 2011 and 2025, and 8 percent between 2026 and 2050. Industrial zones and export promotion zones are expected to increase and the service sector, including financial and tourism services, is expected to grow rapidly. Source: ADB (2010). Importanceofnon-climate-relatedfactors considered to be a major factor affecting flooding incontributingtoflooding in HCMC.39 However, the watershed of the Dong Nai River basin has been subject to considerable The HCMC study shows that a number of non- deforestation. Deforestation leads to increased run- climate-related factors also contribute to and exacer- off, erosion, and sediment release and exacerbates bate the impacts of flooding. This includes domestic flooding in HCMC; this needs to be addressed. solid waste that is often dumped in canals and wa- terways, as well as poor dredging of canals. Canals urbanplanningandassumptionsaboutthe are often choked with water weeds and garbage, cityinthefuture so that in addition to accumulating sediment, their capacity to drain storm waters is severely limited As with the other studies, analysis carried out for (ADB 2010). Further, land subsidence is also an HCMC built its understanding of the future city important problem in HCMC, but was not included on the basis of a vast amount of data relating to in the flood modeling simulations because like Ma- current conditions and a number of sectoral and nila, land subsidence data were not reliable. This urban development plans and strategies devel- is an important limitation of this study. The upper oped by the government about the future (Box 3.3). watershed of the Saigon River and Dong Nai River, Briefly, the planning system in HCMC is shaped both of which drain HCMC, are well-regulated with by several strategic plans, including the urban dams and reservoirs, including the man-made Dau Tieng Lake and Tri An Lake, which are able to reduce 39 The Dong Nai and Saigon rivers are heavily regulated by peak flood flows associated with extreme precipita- hydropower and irrigation dams. They are the main source tion events. Given the intense management of the of energy and water supply for HCMC. However, climate change has not been factored into their construction or op- HCMC upper watershed, upstream rainfall is not eration. estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 41 box3.4 overviewofdownscalingandhydrologicalAnalysisCarriedoutfor hCMCStudy Low-resolution outputs (~2o) from ECHAM's (European Centre for Medium Range Weather Forecast-University of Hamburg) atmosphere-ocean-coupled global circulation model (GCM) Version 4 were downscaled by PRECIS (Providing Regional Climates for Impacts Studies, and a dynamic modeling tool developed by the UK Met Office Hadley Centre for Climate Prediction and Research, precis.metoffice.com). Two IPCC SRES greenhouse gas emission scenarios were used as inputs to the GCM. SRES A2 is a high-emission scenario, while B2 represents the low-emission world. ECHAM is considered a "moderate" GCM in that it does not give any extreme temperature or precipitation projections, but tends to be in the middle ranges. However, other GCM's are also available and need to be included in subsequent studies. PRECIS was used because it is the regional modeling tool that has been most widely used globally, as well as being the tool that the UNDP National Communication Support Unit recommends for countries receiving GEF funding. PRECIS gridded outputs at resolution 0.22o (~25km) for the Sai Gon-Dong Nai Basin were provided by SEA START RC. Daily outputs for maximum/ minimum temperature and rainfall for the baseline decade (1994­2003) and a future decade (2050­59), which represent the year 2050, were taken as a key input for the hydrodynamic model--HydroGIS--to calculate the water and flood regime of Ho Chi Minh City. PRECIS gave the total and seasonal rainfall anomalies that are comparable, though with a slightly wider range than the JBIC/IR3S ensemble, which was used by the other city studies in this collaborative initiative (see table below). By using PRECIS, the daily outputs as well as spatial distribution of climate variables were also obtained to get more in-depth information on temporal and geographic variables for the study area. The table below provides a comparison of 10 years averaged June-July-August rainfall from PRECIS and the IR3S rainfall ensemble for Ho Chi Minh City for the year 2050 (%change relative to baseline) Monsoon-driven seasonal rainfall years with maximum and mini- GHG Emission Scenarios PRECIS IR3S mum annual rainfall for each decade were selected to represent "wet" and "dry" years of the baseline and future A2 and B2 decades. High +7.5 (A2) +4.4 (A1FI) Event-based rainfall extremes for the baseline and future time Low +1.5 (B2) +2.9 (B1) slices were derived from the 1-, 3-, 5-, and 7-day maximum rainfall from PRECIS over the HCMC urban area. Extreme rainfalls of 30- and 100-year return periods were extrapolated from the log-log (power) regression between the frequency of occurrence (years) and amount of each rainfall extreme. The 30-year and 100-year extremes and 1-, 3-, 5-, and 7-day rainfall data were rescaled to the observed extreme rainfalls at Tan Song Nhat Airport. The scaling factors for each extreme period were used for scaling the respective future extreme rainfall. Source: Adapted from ADB (2010). master plan prepared by the Department of Plan- In terms of land use, key trends include a declin- ning and Architecture, the land use plan prepared ing population in central HCMC; doughnut-shaped by DoNRE, and the socioeconomic development urbanization, with higher population growth within plan prepared by the Department of Planning 10 kms from the center of the city; and industries be- and Investment. These are prepared by different ing relocated and established in industrial zones. Ag- agencies under different time scales with little riculture and forest land (primarily made up by the coordination between them, thus posing a difficult Can Gio biosphere reserve) situated in the rural and problem in estimating future infrastructure and suburban areas, has declined from 64 percent in 1997 urbanization scenarios. In terms of a development to 59 percent in 2006. According to DONRE's land scenario for 2050, HCMC has an urban master plan use plans, open space is expected to decline by 25 that was approved in 1998. In 2007, a study was percent by 2020 (ADB 2010). Rivers and canals cover completed to adjust the HCMC master plan up close to 15 percent of the total land use, reflecting the to 2025 (commonly referred to as the Adjustment Plan).40 The city case study used the Adjustment Plan 40 Even though it is in the process of being formally ap- as the basis of the most likely development scenario for proved by the government, in the interim, it is the recog- 2050 (ADB 2010). nized update for the 1998 Master Plan. 42 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport original swamp land on which the city was settled. calculated. Thus, it was estimated that future extreme Roads and other transport facilities occupy around 3 rainfall under the high emissions scenario will increase by percent; land for industry and commerce is expand- more than 20 percent for a 1-in-30-year flood and by 30 ing rapidly and now represents about 5 percent of percent for a 1-in-100-year flood. This information was the city's land use. There are 90 parks in HCMC. In then used as input into the hydrological modeling. addition to the urban master plan and the HCMC land use plans, the department of transportation has Inputsandoutputsforthehydrological prepared a transport master plan for HCMC until modeling 2020, which has been used for modeling impacts of climate change on the transport sector (ADB 2010). The main drivers of the water regime in the HCMC HCMC also has a power development plan and a include the following: seasonal monsoon-driven health sector master plan, both to 2020. Similarly, rainfall, extreme rainfall due to typhoons and tropi- a water supply master plan until 2025 is currently cal storms in the vicinity of the city, local SLR, storm under preparation (ADB 2010) and has been used to surge, upstream-downstream inflow as a function of model impacts on the water sector in 2050. catchment basin hydrology, and various land uses such as water management infrastructure, hydrolog- Floodprotectioninfrastructuresassumed ic/hydrodynamic conductivity, and water demand forthehydrologicalanalysis by sectors and geographic locations. Some of these are related to regional climate change and are more a HCMC has a planned flood protection dyke and function of land use and development. For instance, sluice system to protect much of the city from flood- "over 75 percent of flooding points in HCMC have ing with an estimated cost of $650 million. This is to occurred following rainfall of 40mm even during be implemented in phases, which will allow lessons ebb tide," which indicates that surcharge from storm learned during implementation to be applied to the drains is a major factor contributing to flooding. For later stages. The 2050 modeling has developed sce- the hydrological simulations, it was assumed that the narios assuming implementation and no implemen- population and land use in 2050 were the same as the tation of the proposed flood protection measures. base year (see HCMC annex). For the HCMC study, a hydrodynamic modeling tool--Hydro-GIS--was dynamicdownscalingtechniqueappliedto used to integrate information about these drivers model2050climatechange and derive output variables (such as flood depth, duration, salinity distribution, etc) for an assessment The HCMC study used the SRES A2 and B2 scenarios of risks and impacts in 2050. as the enveloping case for high and low projections. Further, unlike the other cities in the study, which estimatesforsealevelriseandstormsurge used statistical downscaling, the HCMC case study used both statistical downscaling and also a PRECIS In addition to using dynamic downscaling to esti- dynamic downscaling technique to model future mate precipitation changes, the study used the DIVA climate change parameters (Box 3.4). The dynamic (Dynamic Interactive Vulnerability Assessment) tool downscaling approach allowed the study to address developed by the DINAS-COAST consortium.41 The the complex hydrometeorological and oceanograph- results from that tool yielded SLR of 0.26 m and 0.24 m ic changes. Specifically, PRECIS allowed simulation of mean (minimum and maximum temperatures) 41 In DINAS-COAST, the DIVA method was applied to for the base case and A2 and B2 scenarios, simula- produce a software tool that enables its users to produce tion of daily rainfall data in the base year (averaged quantitative information on a range of coastal vulnerabil- over 1994­2003) and for 2050 (based on average ity indicators, for user-selected climatic and socioeconomic of 2050­59). Based on this, the percent increase in scenarios and adaptation policies, on national, regional and global scales, covering all coastal nations. http://www.pik- rainfall in terms of average increase and increase for potsdam.de/research/research-domains/transdisciplinary- extreme events (1-in-30 year, 1-in-100 year, etc.) was concepts-and-methods/project-archive/favaia/diva estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 43 TAble3.12 C limateChangeparameterSummaryforhCMC Temperature Precipitation 24-hr Precipitation 3­5 Land subsidence IPCC Scenario increase (oC) event days Sea Level Rise (m) Storm Surge (m) (m) B2 +1.4 ­25% Insignificant change 0.24 1.08 (from 0.53 2008) Not considered A2 +1.4 +20% +20% 0.26 1.08 (from 0.53 2008) Not considered for the A2 and B2 scenarios, respectively. The study assessment of not only floods, but also drought also notes that the city's low topography creates a events. While this was not defined as a focus of the situation where a tipping point42 of around 50 cms study for the four cities, it is important in the context would considerably increase the impact of SLR, with of HCMC. Drought scenarios also were modeled. significant areas potentially becoming permanently flooded. For storm surge, the study assessed that the historic storm surge lasts about 24 hours and MAInFIndIngSFRoM reached about 0.5 m. For storm surge from intensi- hydRologICAlAnAlySISAnd fied typhoons in 2050, the study examined adjusted gISMAppIngFoRhCMC historic tracks that would increase the storm surge and estimated future storm surges associated with Floodhazardincreases,forbothregular extreme events at 1.08 m, which would be associated andextremeeventsandnumberofpersons with a track making landfall at HCMC. exposedtofloodingrisesdramatically Floodsimulationslimitedtoregularand Large areas of HCMC (1,083 km2)43 flood annually; extreme1-in-30-yearfloodeventsforcurrent extreme floods (1-in-30-year) inundate 1,335 km2. andthe2050A2highemissionscenario Similarly, much of HCMC's population already experiences regular flooding with about 48 per- In assessing the impact of extreme flooding in cent of the communities affected annually. Table HCMC, the study focused mainly on the 1-in-30-year 3.13 shows a comparison of regular (annual) and extreme event for 2008 and the 2050 scenario. Also, extreme (1-in-30-year) flooding under the current the study focused mainly on the A2 scenario because and the 2050 A2 scenario for inundated areas and the team found that the extreme precipitation events communities affected. The analysis shows that the under the B2 scenario were not significantly differ- area inundated increases for regular events from 54 ent than 2008 events. A summary of the simulations percent to 61 percent in 2050, and for extreme events conducted is provided in Annex B. A significant advantage of the dynamic downscaling technique 42 A point at which the changes or impacts rapidly increase, is that it gives temporal and spatial information on possibly irreversibly. 43 1 km2 = 100 ha. precipitation patterns on a daily basis, which allows TAble3.13 S ummaryofFloodingatpresentandin2050withClimateChange Present 2050 Regular flood Extreme flood Regular flood Extreme flood Number of communes affected 154 235 177 265 (from a total of 322) Area of HCMC flooded (ha) 108,309 135,526 123,152 141,885 % of HCMC area affected 54 68 61 71 Source: ADB (2010). 44 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport from 68 percent to 71 percent; that is, an increase h FIguRe3.13a CMCCityCaseStudy: of 7 percent and 3 percent for regular and extreme Comparisonof1-in-30- events respectively in 2050. This is also indicated yearFloodfor2008 through GIS maps (Figures 3.13a and 3.13b). Significantincreaseinbothflooddepthand durationforregularandextremeeventsin 2050 A key finding of the hydrological analysis is that there is a significant increase in both depth and duration for both regular and extreme floods over current levels in 2050. The average maximum flood depth in HCMC is about 35 cm and the average maximum flood duration is 64 days. As a rule of thumb, more than 50 cm of flood water is consid- ered the threshold for safe operation of vehicles and bikes. Flood duration of more than three days at 50 cm depth is considered to cause significant h FIguRe3.13b CMCCityCaseStudy: disruption in the city. Results from the hydrological Comparisonof1-in-30- analysis estimate that there will be a 20 percent and yearFloodfor2050A2 43 percent increase in average maximum depth for Scenario regular and extreme events respectively, and a 21 percent and 16 percent increase in average maxi- mum flood duration during regular and extreme events respectively in 2050. That is, the average maximum flood duration increases from 64 to 86 days for regular events, and from 18­22 days for extreme events. Similar to past events, the depth of extreme floods in 2050 throughout the city will be higher (72­100 cm) compared to the depth for regular flooding (35­44 cm). Increaseinpopulationsatriskfromflooding byregularand1-in-30-yeareventin2050 Source: ADB (2010). Results from the hydrological modeling and GIS mapping show that an increasing number and proportion of the population44 will be affected by 44 As mentioned earlier, the official population of HCMC flooding from regular and extreme events in 2050 in 2007 was 6.4 million people and about 8 million if un- (Table 3.13). For regular floods, about 15 percent registered migrants are included. The team considered two (high and a low) population growth scenarios, but assumed of HCMC's population is affected for the baseline a high population growth scenario for the vulnerability scenario (in 2007). This increases to 49 percent analysis. The high case includes unregistered migrants into the baseline; that is, it assumes current population to be without implementation of proposed flood control about 8 million. This seems more realistic compared to the measures. With the implementation of the proposed assumption in the low-growth scenario, which does not as- flood control measures, the percentage of affected sume unregistered migrants currently in the city. Further, the existing rate of 2.4 percent population growth assumed population is reduced to 32 percent (ADB 2010). in the low-case scenario is low compared to the experience Climate change will also impact daily life in HCMC of other large cities in the region. estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 45 by 2050 under the A2 scenario. As Table 3.14 shows, people are likely to be affected. However, even with currently about 26 percent of the population would the flood control measures being implemented, be affected by a 1-in-30-year event. However, this more than half of the projected 2050 population is is expected to rise to 62 percent of the population still at risk from flooding during extreme events. (about 12.9 million people) by 2050 under the A2 scenario without implementation of the proposed districtsandareasmostatriskfrom flood control measures (ADB 2010). With the con- floodingdueto1-in-30-yeareventin2050 struction of the flood control system, it is estimated that under the A2 scenario, for a 1-in-30-year flood, Another important finding from the hydrological 52 percent of the population--or about 10.8 million analysis is that the spatial distribution of floods people--will be affected. Thus with the implementa- is also likely to change in 2050 under extreme tion of the flood control measures, 2 million fewer conditions, depending on implementation of dif- TAble3.14 d istrictpopulationAffectedbyanextremeeventin2050 2050--without flood 2050--with flood control Population 2007 Population 2050 2007 control measures measures District (1,000) (1,000) No. % No. % No. % District 1 201 232 88 44 100 43 64 27 District 2 130 1,492 85 66 1,410 95 1,160 78 District 3 199 148 17 9 42 28 35 23 District 4 190 125 99 52 125 100 81 64 District 5 191 128 42 22 83 65 54 42 District 6 249 216 34 14 193 90 26 12 District 7 176 1,071 94 53 1,071 100 691 65 District 8 373 575 88 24 574 100 171 30 District 9 214 3,420 39 18 2,322 68 2,311 68 District 10 239 172 67 28 76 44 76 44 District 11 227 154 46 20 36 24 26 17 District 12 307 1,583 65 21 795 50 788 50 Go Vap 497 592 73 15 149 25 103 17 Tan Binh 388 671 9 2 20 3 20 3 Tan Phu 377 482 0 0 19 4 17 4 Binh Thanh 450 623 64 14 511 82 502 81 Phu Nhuan 176 146 0 0 4 3 0 0 Thu Duc 356 1,433 133 37 756 53 733 51 Binh Tan 447 1,557 64 14 776 50 318 20 Cu Chi 310 1,804 83 27 489 27 502 28 Hoc Mon 255 1,483 77 30 728 49 728 49 Binh Chanh 331 1,926 282 85 1,744 91 1,601 83 Nha Be 75 437 75 100 437 100 369 84 Can Gio 67 393 66 99 393 100 393 100 Total 6,425 20,863 1,690 26 12,851 62 10,766 52 Source: ADB (2010). 46 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble3.15 d istrictsAffectedbyFloodinginbaseyearandin2050 Present 2050 Regular Extreme (Linda) Regular Extreme Commune name Flooded % flooded Flooded %flooded Flooded %flooded Flooded %flooded Dist (means District) Area (ha) area (ha) area area (ha) area area (ha) area area (ha) area Dist.1 762 42 5.56 215 28.24 54 7.13 327 42.94 Dist 2 5,072 3,134 61.80 4,525 89.21 3,691 72.78 4,784 94.32 Dist 3 471 0 0.00 57 12.14 0 0.00 133 28.17 Dist 4 408 61 14.93 397 97.35 110 27.03 408 100.00 Dist 5 432 15 3.46 183 42.28 16 3.68 282 65.17 Dist 6 713 16 2.25 224 31.41 56 7.88 638 89.53 Dist 7 3,554 1,068 30.05 3,178 89.42 2,040 57.39 3,552 99.94 Dist 8 1,969 185 9.42 1,700 86.38 846 42.99 1,964 99.78 Dist 9 11,357 7,165 63.09 7,829 68.93 7,350 64.72 8,103 71.34 Dist 10 584 37 6.25 144 24.57 37 6.25 258 44.08 Dist 11 508 18 3.55 70 13.77 18 3.55 120 23.67 Dist 12 5,463 2,559 46.85 2,621 47.98 2,563 46.92 2,742 50.20 Dist Go Vap 2,010 171 8.50 343 17.05 219 10.92 506 25.19 Dist Tan Binh 2,226 0 0.00 41 1.86 0 0.00 65 2.92 Dist Binh Thanh 2,094 594 28.36 1,619 77.33 942 44.99 1,718 82.02 Dist Phu Nhuan 468 0 0.00 1 0.12 0 0.00 12 2.56 Dist Thu Duc 4,692 1,514 32.27 2,140 45.62 1,813 38.64 2,256 48.08 Dist Cu Chi 43,246 7,313 16.91 10,842 25.07 7,985 18.46 11,735 27.14 Dist Hoc Mon 10,838 3,657 33.74 5,311 49.00 3,855 35.57 5,322 49.10 Dist Binh Chanh 25,422 16,924 66.57 22,340 87.88 20,863 82.07 23,057 90.70 Dist Nha Be 10,005 8,272 82.68 9,984 99.79 9,876 98.71 10,005 100.00 Dist Can Gio 61,284 55,148 89.99 60,413 98.58 59,734 97.47 61,284 100.00 Dist Tan Phu 1,575 0 0.00 0 0.00 0 0.00 63 3.98 Dist Binh Tan 5,192 415 7.98 1,349 25.98 1,082 20.84 2,551 49.12 HCMC TOTAL 200,346 108,309 54.06 135,526 67.65 123,152 61.47 141,885 70.82 Source: ADB (2010). ferent adaptation options. As Tables 3.14 and 3.15 in flooding, while others (Districts 1, 7, 11) will see indicate, the districts are not evenly affected due to a reduction in flooding. flooding. Districts 6 and 8 and Binh Than district, for example, will experience a significant increase Thepoorareatgreaterriskfromflooding; in the proportion of population affected, as well as insomecases,boththepoorandnon-poor an increase in the area flooded compared to other areatrisk districts. Some areas--such as Nha Be and Can Gio--will be severely affected by extreme events, In general, poorer areas in HCMC are more vulner- with 100 percent of its area flooded. With the flood able to flooding. As Figures 3.14a and 3.14b show, control system implemented, some districts (District some of the districts that are more vulnerable to 9, Binh Than) will experience significant increases flooding (dark purple in Figure 3.14b, are also the estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 47 Figure3.14a h CMCpovertyRatesby rural poor living in the south of the city are directly district dependent on natural resources for their livelihood. For instance, 60 percent of agricultural land would be affected by saline intrusion during regular floods. Inundation by salt water can damage crops and reduce the productivity of agricultural land, thus impacting the livelihoods of the rural poor. In urban areas of the city, the poor typically live along canals and drainage ditches--for example, Nhieu Loc Thi, Nghe canal, Tan Hoa-Lo Gom canal, Doi- Te canal--or in slum areas near industries. These areas are among those at higher risk from flooding, in part due to underdeveloped infrastructure and poor drainage and sanitation facilities. economicactivityishighlyvulnerableto floodingnowandinthefuture The industrial sector is an important and growing part of the HCMC economy. Climate change will affect economic activity and industrial production FIguRe3.14b d istrictsvulnerableto both directly through inundation of production ar- Flooding eas, and indirectly through inundation of essential infrastructure linked to strategic economic assets and effects on the availability of key production inputs (e.g. water, primary resources). Fifty percent of industrial zones (IZs) are at risk45 of flooding from extreme events with the proposed flood control measures in place, and 53 percent without the sys- tem in place (ADB 2010). An additional 20 percent of IZs will be located within 1 km of likely inunda- tion if the proposed flood control measures are in place (22 percent if the flood control measures are not in place), meaning that they are likely to suffer indirect impacts. In terms of land area assigned to industrial activities, about 67 percent of the area potentially under industrial use in 2050 is likely to be affected (Table 3.16). The city's existing and planned transport network is also likely to be exposed to increased Source: ADB (2010). 45 Risk from impacts due to climate-related impacts--name- ly, flooding-- in the study was determined by calculating districts with higher poverty rates (indicated by red the distance of an infrastructure component or facility from and orange in Figure 3.14a). However, in some of the flood zone that is inundated and assigning it a risk fac- tor. Thus a distance of 0km from the flood zone shows it is the areas that are most at risk--such as Can Gio and inundated, <1km = very high risk, <5km = high risk, <10 Nha Be--the poor and non-poor are both at risk. The kms =medium risk, and > 10kms =low risk. 48 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble3.16 e ffectsofFloodingon Impactofsalineintrusiononnatural Futurelanduseunder resourcedependentsectors 2050A2extremeevent It is estimated that the impact of flooding on ag- % future land use % future land use riculture, forestry, and primary industries will be Future land affected without flood affected with flood severe. In 2050, it is expected that there will be a use type control system control system significant increase in saline intrusion. Saline intru- Urban 61 49 sion is mainly a problem during regular flooding, Industrial 67 63 which is dominated by tidal influences (ADB 2010). Open space 77 76 It is estimated that close to 60 percent of agricultural Source: ADB (2010). lands are expected to be affected by increased salin- ity in 2050. Increased storm activity and sea level rise are expected to increase the level of salinity in flooding. Under the A2 scenario, a 1-in-30-year surface waters and also groundwater resources. event would inundate 76 percent of the axis roads, However, since agriculture is of declining impor- 58 percent of the ring roads, and nearly half of the tance in HCMC, this has a more direct impact on city's main intersections, even with the flood dyke the poor and farmers directly dependent on these protection system. Further, 60 percent of the city's resources rather than the economy as a whole. Fig- waste water treatment plants and 90 percent of the ure 3.16 shows an overlay of the 2050 A2 1-in-30- landfill sites are at risk of flooding (Figure 3.15). year flood on projected land use patterns. It shows The environmental consequences of flood waters that flooding will impact broad areas and sectors, inundating landfill sites and carrying pollutants but particularly agricultural and protected areas. to nearby fish ponds is a potential risk that needs further consideration. FIguRe3.16 h CMC2050A21-in-30- yearFloodInundation FIguRe3.15 I mpactonwaste overlaidonprojected ManagementSector landusepatterns Source: ADB (2010). Source: ADB (2010). estimatingFloodImpactsandvulnerabilitiesinCoastalCities | 49 potentialincreaseindroughtin2050under increase under the low-emission scenario in both the lowemissionscenario dry and the wet season. These assessments require more in-depth analysis. However, they do provide An initial analysis of the data collected for the an indication of the order of magnitude of change study shows that "the frequency of dry season and illustrate which scenarios are more susceptible drought in 2050 is likely to increase by the order of to drought. 10 percent under the low-emission scenario with little change under high-emission scenario" (ADB 2010). Further, the incidence of drought is likely to ConCluSIon To sum up, all three megacities face considerable climate-related risks in terms of changes in tem- FIguRe3.17 h CMCdroughtsand perature and precipitation. These climate risks SalinityIntrusionin are compounded by risks posed by non-climate 2050 factors (such as land subsidence, poor drainage, and deforestation in upper watersheds), thus in- creasing the likelihood of urban flooding. Under different scenarios, in all three cities, there is likely to be an increase in the area exposed to flooding and an increase in the percentage of population affected by extreme events. Given that all three cities are megacities with high population growth rates, the results warrant serious consideration. Despite data limitations, it is apparent that both climate and non-climate factors are important and need to be considered in future adaptation efforts. While flood protection infrastructures are either in place or planned, the analysis shows that in all three megacities, while these are likely to reduce the impact of flooding, they are not sufficient to provide the expected level of protection from future climate events. Source: ADB (2010). 50 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Assessing Damage Costs and Prioritizing Adaptation 4 Options C hapter 3 described the scale of physical TAble4.1 S ummaryofdamages impacts in terms of area, population, and Assessedinthebangkok sectors that are likely to be affected under Study different climate and infrastructure scenarios in Damage Type Sector Damage Mechanism the coastal cities. In this chapter, we discuss (a) the nature of the damages and monetary costs borne Direct Residential units Damage to buildings and assets Damages as a result of these physical changes, and (b) adap- tation options considered by the city studies. The Commercial units Damage to buildings and assets overall approach to estimating costs was discussed Industrial units Damage to buildings and assets in chapter 2 and is further elaborated in this chapter. Transportation, No direct damage to these public health, en- infrastructures expected in the ergy, water supply future and sanitation bAngKoK:AnAlySISoFdAMAge Indirect Population Loss of income CoSTSRelATedToFloodIng Damages In2008And2050 Commercial units Loss of income Industrial units Loss of income In Bangkok, climate-change-induced flooding will Transportation Loss of revenue (negligible) likely result in physical damages to buildings and housing, income losses to individuals and firms, Public health Additional costs of medical care revenue losses to public utilities, and might also Energy Loss of net revenue adversely affect public health. In estimating costs, Water supply and Loss of net revenue the Bangkok study examines direct and indirect sanitation tangible damages (Table 4.1) associated with po- Source: Panya Consultants (2009). tential floods under 16 scenarios. Damage costs are evaluated for the base year of 2008 and for 2050. As previously described, the scenarios include three key inputs for the damage cost assessment. In addition, climate change scenarios (current climate, B1 and the damage cost assessment makes numerous as- A1FI), three levels of flood intensity (1-in-10 year, sumptions regarding prices, population growth, 1-in-30 year, and 1-in-100 year) and additional sce- GDP, and the location of major infrastructure such as narios that include land subsidence, sea-level rise, roads, electric utilities and so on related to Bangkok and storm surge. Estimates for area, depth, and duration in 2050, as discussed in chapter 2. While assessing of flooding derived from the hydrological analysis were damage costs, prices are held constant in real terms 51 over the period of analysis. These assumptions are d FIguRe4.1 amageCostAssociated described in chapter 2. Damages costs are evaluated witha1-in-30-yearFlood in Thai Baht and converted to US dollars, using an (C2050-lS-SR-SS-A1FI-T30) average exchange rate for 2008 of 1 US$ = 33.31 THB. 60,000 55,432 Damage costs (millions of THB) 50,000 45,945 bangkoklikelytowitnesssubstantial 40,000 damagecostsfromfloodingin2050,ranging 30,000 21,671 from$1.5billionto$7billionunderaseries 20,000 12,302 ofclimateandlandusescenarios 10,000 9,726 140 2,012 1,145 13 0 Table 4.2 presents the costs incurred from damages Residence Commerce Industry Public Health Energy Water Supply & Sanitation to buildings, income losses, health costs, and rev- enue losses to public utilities for a range of scenarios. A comparison across scenarios shows that the costs Indirect Impact Direct Impact are likely to significantly increase with climate and Damage to Residencial Building 55,432 MBaht Damage to Commercial Building 45,945 land use change. In the base case scenario (2050- Damage to Industrial Building 12,302 LS-T10), a 1-in-10-year flood is estimated to result Income Loss of Daily Wage Earner 140 Income Loss of Commerce 21,671 in damages of 52 billion THB ($1.5 billion), but a Income Loss of Industry 9,726 1-in-100-year flood with climate change (2050-LS- Health Care Cost 2,012 Income Loss of Energy 1,145 SS-SR-A1FI) could cost THB 244 billion ($7 billion). Income Loss of Water Supply and Sanitation 13 The increase in costs from flooding are best illus- Total Damage Cost 148,386 MBaht trated by looking at the implications of a medium- Source: Panya Consultants ( 2009). sized flood. A 1-in-30-year flood in 2008 (2008-T30), for instance, is estimated to result in damage costs of 35 billion THB ($1 billion). In 2050 with climate change (2050-LS-SS-SR-A1FI-T30), a similar 1-in- intensity in any one year. The difference in the areas 30-year flood would cost 148 billion ($4.5 billion)46. between different exceedance curves represents Thus, in a future climate change (A1FI) scenario, a 1-in- the incremental annual damages as a result of a new 30-year flood in 2050 could lead to a four-fold increase climate/land use scenario. These annualized values in costs to Bangkok. are discussed later in the chapter. To illustrate the nature of different costs associ- ated with floods, Figure 4.1 breaks down the costs associated with a 1-in-30-year flood in the future FIguRe4.2 l ossexceedanceCurves, (2050-LS-SS-SR-A1FI). As the figure shows, the bangkok direct costs to buildings are the highest costs borne 300,000 as a result of flooding, followed by income losses 250,000 Damage costs (mill THB) to commercial establishments. 200,000 It is useful to understand more carefully the 150,000 relationship between the costs of hazards such as 100,000 floods and the probability of their occurrence. Fig- 50,000 ure 4.2 plots total damage costs to Bangkok from 0 0.00 0.05 0.10 0.15 floods in different scenarios against the probability Probability of ood ocurrence of their occurrence. As this figure suggests, the costs 2008 2050-LS 2050-LS-SR-B1 of damage increase for higher intensity floods, but 2050-LS-SR-A1FI 2050-LS-SR-SS-A1FI this also means that there is a lower probability of Source: Based on calculations in Panya Consultants (2009). such floods occurring. The area beneath each flood exceedance curve represents the average total ex- 46 1 USD=33.3133 THB, which was the average exchange pected flood damage cost from floods of different rate in 2008. 52 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble4.2 S ummaryofFloodandStormdamages,bangkok(million2008Thb) 2050-LS- 2050-LS- 2050-LS- 2050-LS- 2050-LS- 2050-LS- 2050-LS- 2050-LS- SR-SS- 2050-LS- 2050-LS- 2050-LS- SR-SS- 2050-LS- 2050-LS- SR-SS- SR- SR-SS- Damaged Items 2008-T10 T10 SR-B1-T10 SR-A1FI-T10 A1FI-T10 2008-T30 T30 SR-B1-T30 SR-A1FI-T30 A1FI-T30 2008-T100 T100 SR-B1-T100 B1-T100 A1FI-T100 A1FI-T100 Damage building 12,166 40,277 54,231 63,982 70,058 27,281 75,500 97,761 105,575 113,679 61,037 142,432 170,368 176,328 179,695 188,215 Residence 7,462 21,500 27,813 31,907 34,465 15,894 37,739 48,746 52,072 55,432 33,557 68,420 80,591 83,312 84,631 88,628 Commerce 3,254 13,565 19,466 23,828 26,814 8,293 29,250 38,689 42,157 45,945 20,960 61,657 74,194 76,781 78,492 82,147 Industry 1,450 5,212 6,952 8,247 8,779 3,094 8,511 10,326 11,346 12,302 6,520 12,355 15,583 16,235 16,572 17,440 Income/revenue 2,761 11,005 15,875 18,132 20,188 7,069 22,717 29,311 31,533 32,695 14,901 39,854 48,699 50,009 50,223 51,629 loss Daily wage 27 47 71 80 89 92 102 128 137 140 167 176 205 205 205 205 earner Commerce 1,315 6,306 9,638 11,138 12,676 4,077 14,940 19,518 20,883 21,671 9,172 27,744 33,963 34,913 34,968 3,508 Industry 1,264 4,260 5,699 6,394 6,894 2,640 7,066 8,596 9,397 9,726 4,897 9,965 12,174 12,534 12,692 13,158 Energy 149 383 456 508 517 254 598 1,057 1,104 1,145 659 1,954 2,340 2,340 2,341 2,341 Water supply & 6 9 11 12 12 6 11 12 12 13 6 15 17 17 17 17 sanitation Health care cost 321 537 717 825 872 934 1,107 1,665 1,893 2,012 3,240 3,473 4,020 4,020 3,945 4,022 Total 15,248 51,819 70,823 82,939 91,118 35,284 99,324 128,737 139,001 148,386 79,178 185,759 223,087 230,357 233,863 243,866 Total USD 458 1,556 2,126 2,490 2,735 1,059 2,982 3,864 4,173 4,454 2,377 5,576 6,697 6,915 7,020 7,320 Source: Panya Consultants' calculation. Notes: LS= land subsidence, SS=storm surge, SR= sea-level rise, B1 =low climate emissions, A1FI =high climate emissions, T10, T30, T100 = flood intensities AssessingdamageCostsandprioritizingAdaptationoptions | 53 damagecostsin2050arelargely damagecoststotransportarelikelytobe attributabletolandsubsidence limited Land subsidence as a result of groundwater utiliza- The Bangkok study examined different types of in- tion is a major problem in Bangkok. While there is frastructure and sought to understand the impact of reason to worry about climate change, the impacts floods on public investments. In general, new public of land subsidence are even bigger. For example, infrastructure in Bangkok has taken the possibility if we look at the impacts of a 1-in-30-year flood in of flooding into account. For example, because of Table 4.2, with current (2008) climate conditions, the the height at which highways and the metro rail are costs are about THB 35 billion ($1 billion). However, built, transportation will generally be unaffected by this increases to THB 99 billion ($3 billion) in 2050, flooding. This does not mean that small streets in just from expected land subsidence (2050-LS). Thus, Bangkok will not be flooded, but the economic costs there is a nearly two-fold (181 percent) increase in to transport infrastructure is expected to be limited. flooding costs between 2050 and 2008 as a result of The Bangkok study was unable to estimate the traffic land subsidence. delay costs associated with flooding, and therefore If we add climate change to this projected this potentially important cost is not included. change in land subsidence and estimate damage There are unlikely to be direct water and sani- costs in the context of a 2050-LS-SR-SS-A1FI sce- tation infrastructure-related costs because of their nario, the costs of a 1-in-30-year flood are THB 148 location at higher than flood levels. However, there billion ($4.5 billion). Climate change results in a are likely to be revenue losses to water and sanita- further 49 percent increase in flood damage costs tion agencies if flood waters rise above 2 meters. in 2050 (relative to 2050-LS). However, the bulk of Under these conditions water supply may become the increase (67 percent) in flooding costs in 2050 dysfunctional and the agencies will likely incur rev- is attributable to land subsidence. Thus, in order to enue losses.48 The revenue loss to water supply and reduce the costs of flooding in 2050, a top priority should sanitation utilities is expected to more than double be policies to reduce land subsidence. to THB 13 million ($390, 235) in 2050, given climate change (A1FI), sea level rise, storm surge, and land damagecostsvarybysectors,withmore subsidence. However, the bulk of this increase in than75percentofthecostsattributableto costs is attributable to land subsidence. buildings If flooding is higher than 1 meter, the Metro- politan Electricity Authority will lose revenues to Damage to buildings dominates flood-related costs the extent of THB 1.1 billion ($33 million) in 2050 in Bangkok. In the context of a 1-in-30-year flood, (A1FI-SL-SR-SS-T30). The damages with climate in each of the five climate and land scenarios con- change are twice the damages that are likely to oc- sidered for Bangkok, 76­77 percent of total damage cur in a no-climate-change scenario (2050-LS-T30). costs are attributable to damage to buildings. As However, infrastructural damages are limited be- Table 4.2 shows, in 2008, a 1-in-30-year flood could cause energy to Bangkok is supplied by two power cause building damages to the extent of THB 27 bil- plants that are either being retired or will be retired lion ($800 million) in Bangkok. A similar 1-in-30-year by 2050. Future infrastructure is expected to be flood could cost up to THB 75 billion ($2.3 billion) in 2050 after taking into account land subsidence and economic development. If we add climate change 47 These values are all in 2008 THB and reflect simply the impacts, the damage costs further increase by 51 physical changes that would be wrought as a result of cli- percent to THB 110 billion ($3.4 billion).47 There is a mate change, economic development, and land subsidence. 48 The average losses are calculated by multiplying water steep increase in building costs between 2008 and charges by the average water used per day and scaling this 2050 (177 percent) simply because of land subsidence up for the duration of floods and number people affected if (Table 4.2). The increase in damages from climate flood waters were higher than 2 meters. Similarly, sanitation costs = No. of affected people x waste generation (solid and change is also significant, but less steep. liquid) per capita per day x treatment cost x flood duration 54 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport built outside the cities and are not expected to be TAble4.3 C hangesinIncomelosses impacted by floods. towageearners, The Bangkok study estimates that health care Commerce,andIndustry costs could double with climate change (Table 4.2). Changes in income losses* from extreme weather (1/30 flood) The costs from a 1-in-30-year flood could be about under different scenarios THB 2 billion ($60 million) in 2050 with climate 2050-LS-SR-SS- change (2050-LS-SS-SR-A1FI), sea level rise, storm Scenarios 2008-T30 2050-LS-T30 (%) A1FI-T30 (%) surge, and land subsidence. This is about twice Daily wage -- 11 25 the cost that would occur in 2050 without climate earner change and only land subsidence (2050-LS). This Commerce -- 266 45 increase in health costs is primarily triggered by the larger number of people likely to be affected Industry -- 168 38 by floods. *Each column represents the percentage change in income relative to the previous column. dailywageearnerswillseesignificant increasesinincomelosses However, as expected, the overall cost to low- Climate change will affect the population of work- income workers is very small relative to the esti- ers living in condensed housing areas and deliver mated costs to industry and commerce (Table 4.2). significant income losses to daily wage earners. As Commercial income losses come second only to Table 4.3 shows, losses to daily wage earners are building damages as a source of losses from floods. likely to increase by 25 percent when we consider For instance, in the case of a 1-in-30 -year flood in a an A1FI climate change scenario (2050-LS-SR-SS- climate change scenario (2050-LS-SR-SS- A1FI-T30), A1FI-T30) in 2050 relative to a scenario where there income losses to commerce are estimated at 21 bil- is only land subsidence (2050-LS-T30). lion THB ($630 million). box4.1 examiningbuildingdamages,Incomelosses,andhealthCostsin bangkok In order to estimate the damage to buildings, the Bangkok study developed GIS maps of areas likely to be flooded. Different categories of buildings likely to be impacted--identified using National Housing Authority data (2004)--were overlaid on these maps. Of particular concern was condensed housing, which refers to clusters of more than 15 houses in an area of 1 rai. These areas, where poverty is high, were separately identified. To value buildings, the Bangkok case study identified the book value of buildings of different types and then depreciated these values. Dam- age costs were assessed as a percentage of total value based on the extent and duration of floods. Added to this was the value of assets that may be damaged. Average asset values associated with residential buildings (obtained from census data) are THB 328,889 for Bangkok and THB 220,180 for Samut Prakarn. Building damages were estimated separately for residential, commercial, and industrial buildings. With extreme precipitation and flooding, income losses can be incurred in flooded and non-flooded areas. The Bangkok study estimated three types of income losses in flooded areas: (a) business income loss, (b) industrial income loss, and (c) losses to daily wage earners. To obtain business losses, the Bangkok study first identified the average income per day from commercial establishments based on business surveys. This was then adjusted to reduce operational expenses. The average net commercial income was estimated to be THB 4,930 per establishment per day. A similar accounting was used to estimate average income to industrial establishments. These average values are multiplied by the number of days of flooding and the number of buildings affected by floods to estimate income losses from commerce and industry. Income losses to daily wage earners living in condensed housing areas were estimated based on the population in these areas affected by floods of different intensity. Disease outbreaks in the form of diarrhea, cholera, typhoid, and other diseases are possible during times of flood, depending on the duration and nature of the areas affected. But few estimates are available of the likelihood of these outbreaks and what populations may be affected. In the absence of disease and cost information, the average cost of hospitalization in Bangkok and Samut Prakhan was used as a proxy for public health costs. These costs, which are THB 7,582 and THB 3,756 per person per admission in the two regions respectively, were multiplied by 50 percent of the affected population to get health damages. AssessingdamageCostsandprioritizingAdaptationoptions | 55 damagecostsconstituteapproximately2­8 and analyzed some of the more prominent practices percentof2008gdpforeventsofdifferent and potential adaptive interventions, such as im- intensities proving flood forecasting and preparing strategies for flood warnings, evacuation, transport during Since Bangkok and Samut Prakarn are the two floods, and post-flood recovery; implementing areas that are mainly affected, flood damages are protection (e.g. construction of dikes/seawalls) or compared to their gross regional domestic product retreat strategies for coastal development; changes (GRDP). Under assumptions of high emissions in coastal development and land use; and public (2050-LS-SR-SS-A1FI), the impacts of a 1-in-30-year information campaigns and training exercises. The flood in 2050 will amount to THB 148 billion (5 municipal agencies in BMR have also developed percent of current GRDP). A bigger 1-in-100-year several plans for flood mitigation and drainage flood would cost about 8 percent of current GRDP. works that include both structural and non-struc- 49 Similar sized floods would cost 1­3 percent of tural measures. However, these plans do not include GRDP under a no-climate-change scenario (2008), climate change considerations. as indicated in Table 4.4. The Bangkok study proposes a portfolio of ad- To carefully estimate the costs of climate change, aptation options in the form of structural measures Table 4.4 compares 2050 flood damages with climate that can mitigate the impact of 1-in-30-year or 1-in- change (2050-LS-SR-SS-A1FI) to a scenario where 100-year floods. there is land subsidence but no climate change (2050-LS). The net impact of climate change is then The results of the simulations indicate that the the difference in costs between these two scenarios. crest elevations of dikes around Bangkok and This amounts to THB 49 billion ($1.5 billion) in 2008 along both banks of the Chao Phraya River will values, or approximately 2 percent of 2008 GRDP. not be high enough to cope with flooding of more than a 10-year return period in the future. Moreover, the protected area to the west of the pRIoRITIzATIonoF Chao Phraya River has insufficient pump capac- AdApTATIonopTIonSIn ity to drain the floodwater into the Tha Chin bAngKoK River and the Gulf of Thailand. Thus, the first set of options includes dike and pumping capacity As discussed earlier, Bangkok has a long history of improvement. floods and its residents have developed numerous adaptation measures to cope with the risk of floods. 49 In terms of GDP, the Bangkok study identifies the current A wealth of well-tested experience, information, and 2008 GDP for Bangkok and Samut Prakarn to be 2,916 billion THB and estimates that GDP in 2050 will be 19,211 billion technology is already available to identify viable THB. All damage costs are represented in 2008 THB, thus to adaptation strategies. The Bangkok study reviewed estimate percent of GDP, 2008 GDP values were used. TAble4.4 d amageCostsinbangkokandRegionalgRdp Damage Costs Million THB (2008) % of 2008 GRDP c2050-LS-SR-SS- Costs of Climate c2050-LS-SR-SS- Costs of Climate Mill THB C2008 C2050-LS A1FI Change C2008 C2050-LS A1FI Change T 10 15,248 51,819 91,118 39,299 0.52 1.78 3.12 1.35 T30 35,284 99,324 148,386 49,062 1.21 3.41 5.09 1.68 T100 79,178 185,759 243,866 58,107 2.72 6.37 8.36 1.99 Source: Based on estimates in Panya Consultants (2009). 56 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport For the eastern part of Bangkok, pumping capac- investment costs of making structural adaptations ity to drain floodwater into the Bang Pakong River investments to be 35.3 billion THB ($1.06 billion) and the Gulf of Thailand needs to be increased. Based for a 30-year return flood protection project, and on simulations, the study considers investments 49.5 billion THB ($1.5 billion) for 100-year return that would increase pumping capacity from flood protection. The total annual operation and 737 to 1,065 m3/sec. The total capacity of canals maintenance cost for 30- and 100-year return floods would be improved from 607 to 1,580 m3/sec. is estimated at 584 ($17.5 million) and 874 million In western Bangkok, there are three major THB ($26 million) respectively. pumping stations at Khlong Phasi Charoeng, Sanam Chai, and Khun Rat Phinit Chai, which proposedadaptationoptionscanreduce drain the floodwater into the Tha Chin River floodedareasby51percent and the Gulf of Thailand with a total capacity of 84 m3/sec. This current capacity is inadequate The maximum inundation area corresponding to a to cope with future climate change. Based on 30-year return period flood with and without the simulation results, increased pumping capacities proposed structural adaptation measures is pre- and canal improvements are proposed. sented in Figure 4.3. The maximum inundation area The BMA has proposed coastal erosion protection, of Bangkok and Samut Prakarn will reduce with including the rehabilitation of mangrove forest along adaptation measures from 744.34 to 362.14 km2, a the shoreline of Bang Khun Thian. In the eastern decrease of 382 km2 or 51 percent. area of the Chao Phraya River, there are plans Table 4. 6 shows flood damage costs in Bangkok to construct rock-pile embankments along the shore- with and without an adaptation-related infrastruc- line to protect the industrial community area from ture investment project. The difference in flood coastal erosion and waves. These considerations damage costs "with" and "without" adaptation would cost 35 and 49 billion THB to protect investments represents the benefits of the adapta- against floods of a 30- or 100-year return period tion investment projects. The expected benefit of respectively. the project is the expected annual reduction in flood damage cost. Box 4.2 describes how the annual Table 4.5 presents the investment costs required benefits are estimated as the incremental benefits to protect Bangkok against a 1-in-30-year flood and a of having flood control projects of different return 1-in-100-year flood in the context of an A1FI climate periods. The average annual benefits (or reduction scenario. The Bangkok study estimates the total in flood damage cost) are estimated at 4.4 and 5.9 TAble4.5 I nvestmentCostsforAdaptationprojectsinbangkok(millionThb) Investment Cost for 30-year Return Period Investment Cost for 100-year Return Period Year FS & DD Civil Work Pump Total FS & DD Civil Work Pump Total 1 21 21 30 30 2 42 42 59 59 3 42 42 59 59 4 1,981 1,981 2,405 2,405 5 3,962 3,962 4,811 4,811 6 5,944 3,083 9,027 7,216 5,065 12,281 7 5,944 6,166 12,110 7,216 10,130 17,347 8 1,981 6,166 8,148 2,405 10,130 12,536 Source: Panya Consultants (2009). Notes: FS=feasibility study; DD=detailed design. AssessingdamageCostsandprioritizingAdaptationoptions | 57 FIguRe4.3 M aximumInundationAreawithoutandwiththeproposedAdaptation Source: Panya Consultants (2009). box4.2 expectedAnnualbenefitsfromAdaptationinbangkok Annual Flood Damage Cost with and without the Project 300,000 The figure above shows flood damage costs in Bangkok with and without an adaption-related infrastructure investment project. A climate change 250,000 scenario of A1FI is assumed. Flood damage cost 200,000 Expected annual benefits from flood control investments are based E 150,000 C on the probability of floods of different intensity (return period) occurring. 100,000 D B Thus, the expected annual cost of flood damage is the area under the flood A 50,000 exceedance curve, which plots the probability of occurrence against costs of damage. Thus, the expected annual benefit from a flood protection project 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 is the difference in the area without and with the project. Probability of ocurrence In the case of Bangkok, the expected annual benefit with an Without Project Without Project Without Project investment project for 30-year return period (or a flood infrastructure for 30-yr Return Period for 100-yr Return Period project that could handle a 30-year flood), is the sum of areas of A, B, and C. For a 100-year return period, this is the sum of areas of A, B, C, D, and E. In the context of a cost-benefit analysis of flood damages, the discounted value of the damages from a 30-year / 100-year return period project is compared to the costs of the project. This allows decision makers to look at net discounted benefits and make a decision on which investment to undertake. Source: Zhang and Bojo (2009). 58 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble4.6 F looddamageCostswithandwithouta30-yearReturnperiodFlood protectionproject(millionThb) Flood Damage Cost Flood Damage Cost by Return Return Period (Year) Probability of occurrence Without Project With Project Without Project With Project 10 0.100 91,145 45,465 9,115 4,547 20 0.050 123,308 59,337 6,165 2,967 30 0.033 148,412 71,811 4,947 2,394 50 0.020 148,412 148,412 2,968 2,968 100 0.010 243,902 243,902 2,439 2,439 Source: Panya Consultants (2009). billion THB ($132 to $177 million) for investments value of damage costs. The results indicate that the that would protect against floods of 30-and 100- flood protection was economically feasible for both year return periods respectively. These benefits are floods of 30- and 100-year return periods if the op- estimated for the case of an A1FI climate scenario. portunity cost of capital is not more than 10 percent. The viability of these investments was deter- From this preliminary evaluation and with the mined by estimating the net present value (NPV) of understanding that Thailand uses a discount rate benefits for two scenarios. A base case considered the of 8 percent for public investments, the Bangkok real value of annual benefits to be constant through- study proposes that flood infrastructure should out the analysis period. The second case incorporated be designed to protect against a 100-year return growth in the real value of infrastructure damage ; it period flood as it provides a higher net return was assumed that damages (or benefits from flood (NPV=13.4 billion THB, or $0.4 billion). Such an control) would grow at an average rate of 3 percent investment will be economically efficient given per year. Discount rates of 8 percent, 10 percent, and an A1FI climate change scenario. However, if a 12 percent were used to estimate the NPV.50 discount rate of 10 percent is applied, Bangkok should opt for the adaptation project aimed at a Investmentstoreducetheimpactsofa 30-year return period. 1-in-100-yearfloodeconomicallyviablefor bangkok 50 The period of analysis was 38 years (2012­50), of which the Table 4.7 presents the results of the NPV calculations first 8 years are for studying, designing, and construction, assuming that there is 3 percent growth in the real and 30 years is the economic benefit period of the project. TAble4.7 n etpresentvalueofAdaptationMeasurestoprovideprotectionAgainst a1-in-30and1-in-100-yearFlood(millionThb) Designed Flood Protection Improvement Project for Description 30-Year Return Period 100-Year Return Period Discount Rate (%) 8 10 12 8 10 12 Present Value of Costs (million Baht) 24,950 21,578 18,831 35,117 30,276 26,349 Present Value of Benefits (million Baht) 36,354 25,439 18,286 48,521 33,954 24,406 Net Present Value (NPV) (million Baht) 11,404 3,862 ­545 13,405 3,678 ­1,944 Benefit-Cost Ratio (B/C Ratio) 1.46 1.18 0.97 1.38 1.12 0.93 Source: Panya Consultants (2009). AssessingdamageCostsandprioritizingAdaptationoptions | 59 AnAlySISoFdAMAgeCoSTS FloodinginManilacancausevarying damagesrangingfrom$109millionto RelATedToFloodIngIn $2.5billionindifferentcurrentandfuture MeTRoMAnIlA scenarios As discussed in the previous chapter, Manila City Flood costs as a result of various climate and infra- is a semi-alluvial plain formed by sediment flows structure scenarios are presented in Table 4.8. The from four different river basins. Its drainage and table presents flood damages associated with floods location between Manila Bay in the west and a of three different intensities (1/10, 1/30, and 1/100) large lake, Laguna de Bay to the southeast, makes under current no-climate-change conditions (SQ), it "a vast drainage basin" that is subject to frequent under scenarios where flood control infrastructure overflowing of storm waters. Manila is vulnerable to is in place (MP) or not (EX), and under two climate different types of flooding, including overbanking, change scenarios (A1FI and B1). Flood-related costs storage-related floods, and interior floods. range from 5 billion PHP ($109 million)--in a sce- The government of the Philippines is well aware nario where this a 1-in-10-year flood, master plan of the flooding problems in Manila; a master plan for infrastructure is in place, and there is no climate flood control infrastructure was developed over a change (10-SQ-MP)--to 112 billion PHP ($2.5 bil- decade ago. If this plan is implemented, the city will lion) in a situation where the planned infrastructure see a significant decline in flood impacts. Thus, the is not in place and climate change contributes to a Manila study discusses the implications of climate 1-in-100-year flood (100-A1FI-EX). change in the context of the master plan as well as Figure 4.4 shows the effects of different intensity what would happen without it. Like Bangkok, the floods in different climate scenarios, given Manila's Manila study examines flooding impacts in three existing flood control infrastructure. It is useful to climate change scenarios (no change, B1, and A1FI). consider the impact, for example, of a medium-sized Overall, the Manila study identifies impacts in terms 1-in-30-year flood. A 1-in-30-year flood would cost of damage costs in relation to 18 different scenarios. $0.9 billion in a no-climate-change scenario (30-SQ- In evaluating costs in dollars, an average exchange EX). These costs would increase by 72 percent to $1.5 rate for 2008 of 1USD = 44.47 PHP was used. billion in the case of climate change (30-A1FI-EX). The Manila case study, like Bangkok, uses a In a B1 climate change scenario, the cost would sectoral approach to estimate costs. Flood-related increase by 55 percent to $1.4 billion. costs are a result of (a) building damages; (b) losses to Figure 4.5 presents the loss exceedance curves public infrastructure and utilities; (c) income losses for Manila. This shows the cost of floods increasing to firms and residents; and (d) income losses from transportation blockages. It undertakes a more de- tailed analysis of the transport sector and identifies increased costs associated with flood-related traffic FIguRe4.4 F loodCostsunderThree disruptions. The study assumes that the value of ReturnperiodsandTwo various costs in 2050 are much higher than 2008, and ClimateScenarios(php) that these costs generally increase at a rate of 5 per- 3,000,000,000 cent per year. However, in the paragraphs below, the 2,500,000,000 2008 values of all costs are reported in order to enable 2,000,000,000 a comparison across cities. Like the Bangkok study 1,500,000,000 and as noted in chapter 3, the Manila case study 1,000,000,000 makes a series of assumptions in valuing damages 500,000,000 and uses best available information to obtain cost 0 SQ EX B1 EX A1FI EX estimates. Thus, damage cost estimates are better 1/10 1/30 1/100 viewed as an indicator of damages rather than as point estimates of actual costs that may be incurred. Source: Based on estimations in Muto et al. (2010). 60 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble4.8 F looddamageCostsinManila(2008php) Cost in 2008 in Pesos P10 SQ EX P10 SQ MP P10 B1 EX P10 B1 MP P10 A1FI-EX P10 A1FI MP Damages to Buildings Residential 785,486,988 320,880,033 842,295,372 491,606,130 595,395,243 546,225,294 Commercial 8,641,501,748 610,789,400 9,658,314,207 1,611,046,487 13,750,520,244 2,326,289,074 Institutional 66,863,814 20,189,999 91,707,535 37,209,916 96,826,650 37,268,296 Industrial 2,890,401,496 1,173,449,757 2,414,697,965 1,461,799,749 1,756,641,760 1,346,409,219 Maintenance cost Current roads 1,162,100 346,199 1,587,787 463,132 2,632,955 543,277 Future roads 44,014 30,219 44,014 31,532 91,969 38,102 Vehicle operating costs 8,823,186 2,628,501 12,055,195 3,516,306 19,990,575 4,124,802 Travel time cost savings 33,199,847 8,380,787 45,754,992 11,655,307 71,672,669 13,646,330 Loss to Firms Sales 2,816,137,180 2,704,662,851 2,961,770,824 2,822,212,152 3,044,628,088 2,881,793,868 Income loss of settlers Formal settlers 32,629,500 20,444,625 49,763,250 26,401,500 49,437,000 29,098,125 Informal settlers 85,652 51,072 151,620 51,072 255,892 72,352 Total 15,276,335,523 4,861,853,444 16,078,142,760 6,465,993,284 19,388,093,046 7,185,508,737 Total USD 343,484,797 109,317,627 361,513,243 145,386,332 435,936,694 161,564,467 Cost in 2008 in Pesos P30 SQ EX P30 SQ MP P30 B1 EX P30 B1 MP P30 A1FI-EX P30 A1FI MP Damage to Buildings Residential 1,802,689,882 399,849,739 3,660,228,253 549,439,668 4,210,760,389 637,339,590 Commercial 22,710,938,518 2,273,492,105 35,692,199,142 7,069,333,943 39,538,199,655 10,143,817,110 Institutional 158,250,637 23,533,947 270,248,699 85,001,479 334,199,868 96,920,697 Industrial 4,216,676,982 1,330,430,240 9,932,796,023 2,657,311,465 11,606,388,976 3,456,942,255 Maintenance cost Current roads 5,286,655 1,102,956 6,846,841 1,937,811 7,482,737 2,313,418 Future roads 244,376 244,376 302,185 302,185 329,119 329,119 Vehicle operating costs 40,138,658 8,374,141 51,984,296 14,712,729 56,812,303 17,564,506 Travel time cost savings 374,633,321 31,760,926 421,032,785 74,184,136 573,888,428 85,170,808 Loss to Firms Sales 10,756,786,447 3,281,670,824 11,832,564,006 4,515,810,393 12,434,679,407 5,075,470,880 Income loss of settlers Formal settlers 93,848,625 39,640,500 184,246,875 49,636,125 196,321,500 51,926,625 Informal settlers 4,731,076 92,036 5,367,880 111,188 5,750,388 118,636 Total 40,164,225,177 7,390,191,790 62,057,816,985 15,017,781,123 68,964,812,770 19,567,913,643 AssessingdamageCostsandprioritizingAdaptationoptions Total USD 903,083,118 166,166,717 1,395,355,360 337,671,262 1,550,657,529 439,979,917 | 61 (Continued to next page) 62 TAble4.8 F looddamageCostsinManila(2008php) (Continued) | Cost in 2008 in Pesos P100 SQ EX P100 SQ MP P100 B1 EX P100 B1 MP P100 A1FI-EX P100 A1FI MP Damage to Buildings Residential 3,688,647,788 1,045,670,772 6,022,893,816 1,326,288,039 7,517,544,912 2,101,690,472 Commercial 37,699,327,245 15,298,341,749 63,871,514,594 25,506,211,401 68,021,524,157 37,713,082,264 Institutional 298,785,692 158,994,559 485,447,235 173,893,911 1,874,981,233 253,765,175 Industrial 8,650,623,155 5,694,313,706 16,556,719,073 5,532,356,399 17,850,618,995 9,193,023,327 Maintenance costs Current roads 8,143,240 3,010,272 9,677,159 4,831,659 10,443,791 5,780,183 Future roads 360,001 360,001 485,467 485,467 524,226 524,226 Vehicle operating costs 50,729,576 22,855,337 62,246,103 36,684,130 68,001,872 43,885,751 Travel time costs 706,986,380 277,477,558 1,082,134,984 197,675,748 1,420,426,406 340,173,579 Loss to Firms Sales 13,403,412,143 6,567,976,899 14,085,687,162 7,745,705,319 14,639,854,088 8,339,388,091 Income loss of settlers Formal settlers 214,933,500 67,473,375 230,942,250 95,140,125 481,092,750 105,586,875 Informal settlers 6,050,968 584,668 6,881,952 1,247,540 7,089,432 2,091,824 Total (PHP) 64,727,999,688 29,137,058,896 102,414,629,796 40,620,519,739 111,892,101,862 58,098,991,768 Total (USD) 1,455,393,788 655,139,889 2,302,768,766 913,342,794 2,515,867,487 1,306,342,110 1 USD=44.474561 PHP Source: Muto et al. (2010). ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport FIguRe4.5 l ossexceedanceCurves lion) without the master plan and PHP 19 billion forManila(php) ($427 million) with the master plan implemented, so 120,000,000,000 implementing the master plan would reduce flood 100,000,000,000 damages by over 70 percent, given climate change 80,000,000,000 and the possibility of a 1-in-30-year flood. Thus, a Damage costs 60,000,000,000 very good starting point for Manila, in terms of its 40,000,000,000 response to climate change, would be to reconsider 20,000,000,000 and evaluate the master plans that are already on 0 the books. 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Probability of oods SQ EX SQ MP B1 EX B1 MP A1FI EX A1FI MP buildingdamagesmakeupamajorportion offlood-inducedcosts Source: Based on estimations in Muto et al. (2010). The single most important contributor to total damage costs in each of the climate scenarios is as the probability of their occurrence decreases. As damage to buildings. Approximately 72 percent of previously discussed, the area beneath each curve the costs from a 1-in-30-year flood (averaged across represents the average total annual expected damage all scenarios), for instance, result from damage to costs from floods of all intensities (also see adapta- buildings. Figure 4.7 shows the damage to build- tion section). Note that the loss exceedance curve is ings in the case of a 1-in-30-year flood. The dam- highest for the A1FI emissions scenario (A1FI-EX). ages are significant. Implementing the master plan Climate change increases the extent of damages in will reduce damages significantly. In the case of a all types of floods. 1-in-30-year flood with an A1FI emissions scenario, implementing the master plan will reduce damages ImplementingexistingMasterplans to buildings by 74 percent (relative to A1FI-EX). relatedtoinfrastructuredevelopmentwill significantlyreducefloodingrelatedcosts Incomelossesfromclimatechange associatedfloodswillincreaseby9to16 As Figure 4.6 shows, under each climate scenario, percent implementing the master plan will result in a signifi- cant reduction in costs. In an A1FI climate change Floods will result in income or revenue losses to scenario, for example, a 1-in-30-year flood will result individuals and firms. A 1-in-30-year flood with in damages to the extent of PHP 69 billion ($1.5 bil- FIguRe4.7 d amagestobuildings FIguRe4.6 d amageCostsAssociated froma1-in-30-yearFlood withdifferentScenarios (2008php) (php) 60,000,000,000 120,000,000,000 50,000,000,000 100,000,000,000 40,000,000,000 80,000,000,000 30,000,000,000 60,000,000,000 20,000,000,000 40,000,000,000 10,000,000,000 20,000,000,000 0 0 P30 P30 P30 P30 P30 P30 SQ EX SQ MP B1 EX B1 MP A1FI EX A1FI MP SQ EX SQ MP B1 EX B1 MP A1FI EX A1FI MP 1/10 1/30 1/100 Damages to Buildings Source: Based on estimations in Muto et al. (2010). Source: Based on estimations in Muto et al. (2010). AssessingdamageCostsandprioritizingAdaptationoptions | 63 TAble4.9 I ncomeandRevenuelossestoIndividualsandFirmsAssociatedwith Floods(2008php) Income Costs of Climate Change % increase in income costs from Flood Intensity SQ EX A1FI-EX (A1FI) Climate Change T 10 2,848,852,332 3,094,320,980 245,468,648 9 T 30 10,855,366,148 12,636,751,295 1,781,385,147 16 T 100 13,624,396,611 15,128,036,270 1,503,639,659 11 Source: Muto et al. (2010). an AIFI climate scenario, for instance, will result within pipes is expected to be strong enough to pre- in income losses to the extent of 12.6 billion PHP vent infiltration by contaminated water. Pumping ($284 million) (Table 4.9). Significant income losses stations are also above flood level. Floods will affect are expected for residents of parts of Manila. The many roads in Manila and will render them not population in the Pasig-Marikina River basin in passable. A 1-in-100-year flood in a climate change Manila and Malabon and Navatos in Kamanava scenario, for example, is likely to inundate over 30 are most likely to be affected. km of roads. Some parts of Manila's rail system will Examining a with (A1FI) and without (SQ) be affected by flooding, particularly if power cuts climate change scenario, Table 4.9 shows that cli- occur during floods. Under some flooding scenarios, mate change is likely to contribute to a 9 percent the railway system, LRT1 will be affected by power increase in income losses given a 1-in-10-year cuts and could be stopped. flood, a 16 percent increase in income losses from a 1-in-30-year flood, and an 11 percent increase Increaseincostoffloodingonroad in income losses if a 1-in-100-year flood occurs. networks More than 95 percent of the flood-related income losses, however, will be a result of losses borne by Road and transportation delay-related damages firms.51 Given a 1-in-30-year flood, firms will see from a 1-in-30-year flood with climate change a 16 percent increase in costs in a climate change (A1FI-EX) would be 638 million PHP ($14 million). (T30 A1FI-EX) relative to a no-climate-change sce- This represents an over 52 percent increase in costs nario (T 30 SQ EX). Comparing the same scenarios, relative to a no-climate change scenario (SQ EX). formal settlers or residents 52 will see an over 100 Over 90 percent of the road and transport-related percent increase in costs. Informal settlers will see costs are related to time-cost delays from traffic their damage costs from flooding increase by 22 disruptions. percent as a result of climate change associated with a 1-in-30-year flood. 51 Because of lack of data, the Manila case study uses rev- enues instead of net revenue losses to firms. Thus, firm-re- SectoralImpactswillvary lated losses are likely to be overestimated. A 2008 survey on business income losses to firms is used to examine the costs Most of the power stations that distribute energy of flooding to business. The average income or sales data to Manila will not be damaged by floods because was obtained for firms that undertake different economic activities such as manufacturing, construction, hotels, and of their location on high ground. However, a few so on. Affected buildings were classified according to dif- substations in the flood-prone areas may be shut ferent uses and the sales losses from each of the buildings obtained, assuming one firm per floor. down for a few hours if flood waters reach the 52 The income of individuals directly affected by floods is ground level. This would mostly lead to some rev- difficult to measure. The Manila study assesses income by enue losses because of closure. In terms of water first estimating the number of households that are affected by flooded buildings. The income loss to these households and sanitation services, piped water supply is not is then estimated using average per capita income data expected to be affected by flooding. Water pressure from the National Statistics Office. 64 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport box4.3 IncreasedTimeCostsandhealthRiskfromFloodinginManila The Manila study aggregates two different estimates to establish the cost of flooding on road networks. It estimates the cost of maintenance on flood-affected roads and adds to this the cost of delays to private individuals that would result from flooding. The cost of delays in travel because of floods are categorized into vehicle operating costs and time costs. Vehicle operating costs are estimated by examining the increased operating costs as a result of flooding for public and private vehicles. For time costs, the Manila study relies on a survey of travelers and uses the average value of time to evaluate private individuals time costs--separate costs are obtained for individuals using private and public transportation. The Manila study assesses the increase in likely gastrointestinal diseases as a result of water ingestion and exposure to pathogens in flood water. Based on coefficients from a dose-response model and the average number of floods that occur per year, the study estimates the increase in the annual risk in gastrointestinal diseases in the areas that are likely to be flooded. The risk of gastrointestinal diseases from accidental ingestion of water rose from 0.0134 at inundation levels of 50 cm of flood waters to 0.187 for inundation levels above 200 cm of flood water. This analysis, while undertaken, was not directly used in the damage cost assessment. Source: Muto et al. (2010). Floodsofdifferentintensitycancost FIguRe4.8 F loodCostsasapercent between3percentand24percentofgdp of2008gdp 60,000,000,000 Table 4.10 compares a "without climate change" 50,000,000,000 (SQ EX) and "with climate change" (A1FI-EX) sce- 40,000,000,000 nario, indicating that the costs of flooding can range 30,000,000,000 from PHP 15 billion ($337 million) (SQ-EX-10) to 20,000,000,000 PHP 111 billion ($2.5 billion) (A1FI-EX-100). Since 10,000,000,000 these numbers are presented in 2008 values, we can 0 compare them to Metro Manila's regional GDP of P30 P30 P30 P30 P30 P30 468 billion PHP (National Statistical Coordination SQ EX SQ MP B1 EX B1 MP A1FI EX A1FI MP Damages to Buildings Board). Damage costs range from 3 percent of GDP (SQ-EX-10) to 24 percent (A1FI-EX-100). Source: Based on estimations in Muto et al. (2010). The costs of climate change range from PHP 4 billion ($89 million) (1/10 flood) to 47 billion PHP (approximately $ 1 billion) (1/100 flood). As Figure 4.8 shows, climate change costs represent 1 percent (1-in-10 flood), 6 percent (1-in-30 flood) and 10 per- TAble4.10 d amageCostsfrom1-in- cent (1-in-100 flood) of GDP. 10,1-in-30,and1-in-100- yearFloodsindifferent Scenarios(2008php) pRIoRITIzATIonoFAdApTATIon Climate Change opTIonSInMAnIlA Damage Costs (2008 PHP) with The Manila case study looked at a variety of adapta- Flood an A1FI Scenario tion options that could reduce the impact of 1-in-30 Intensity SQ EX A1FI-EX with EX and 1-in-100-year floods. The adaption options exam- 1/10 15,276,335,523 19,388,093,046 4,111,757,522.59 ined included improving current practices, capacity 1/30 40,164,225,177 68,964,812,770 28,800,587,593.15 building and better coordination among local gov- 1/100 64,727,999,688 111,892,101,862 47,164,102,174.61 ernment and national flood management agencies and structural measures such as dam construction, Source: Muto et al. (2010). AssessingdamageCostsandprioritizingAdaptationoptions | 65 raising dikes, and improved pumping capacity that Table 4.12 presents the estimated initial costs of would reduce and/or eliminate the impacts of floods. each adaptation option considered.54In this analysis, the construction period considered is 5 years start- Manilatargetsadaptationoptionsthat ing from 2010 and the project's life is expected to wouldalmosteliminatefloodsinthepasig- last until 2060. Investments after the completion of MarikinaRiverbasin the master plan are assumed to start from 2014. No maintenance cost was included. Several structural measures are examined in order to analyze the economic implications of these in- Theincrementalbenefitsofmovingfrom vestments. The Manila study focused on analyzing thecurrentsituationtofulladaptationand adaptation options that could eliminate floods to the fromimplementingthemasterplanand extent possible.53 These include: thenmovingtofulladaptationconsidered Pasig-Marikina River: In this area, in order to The investment costs (assumed to occur over a prevent overbanking and flooding from the 5-year period) are compared with the incremental river, investments in raising the embankment expected annual benefits from flood reduction asso- are considered. Embankment raising is con- ciated with a 1-in-10-year, 1-in-30-year, and 1-in- sidered both with the possibility of building 100-year flood protection projects. The incremental the Marikina Dam and without. The Marikina benefits emerge from adaptation investments that Dam was initially proposed as part of the 1990 would take Metro Manila from its: master plan and would be able to prevent a 1-in-100-year flood. existing infrastructure (EX) to full adaptation West of Mangahan and KAMANAVA area: In level (difference between EX and full adapta- these areas, improved pumping capacity and tion in Figure) storm surge barriers are required to reduce 1990 master plan level (MP) to full adaptation floods. The investment required for installing level (difference between MP and full adapta- pump capacity to control flooding under an tion in Figure) allowable inundation depth of 30 cm is consid- ered. For preventing overflows from Manila Bay The annual benefits are identified in Figure 4.9. caused by typhoons, the study looks at options The area between the existing infrastructure curve for increasing the height of storm surge barriers and full adaptation level are the annual benefits in coastal areas. from making investments in the current situation where the master plan has not been implemented. The Manila study examines adaptation options The area between the master plan curve and the full among a number of different scenarios involving adaptation level refers to the annual benefits from different types of construction in different areas. Ta- going beyond the master plan. ble 4. 11 identifies the options that were considered In the cost benefit analysis of the different under each scenario. Notably, while the Bangkok investments, flood control benefits are assumed to study looks at adaptation only in the context of an grow at an annual rate of 5 percent. The net present A1FI scenario, the Manila case study looks at adap- value (NPV) of benefits was obtained using a dis- tation options in the context of no-climate-change (SQ), B1 and A1FI scenarios. The Manila study also 53 For KAMANAVA and West of Mangahan areas, total explores scenarios in which a major investment in elimination is not possible because of their low height. Rather, pumping capacity improvement is considered to constructing the Marikina dam is undertaken (wD) minimize the duration of the flooding. and cases where the dam is not constructed (nD). 54 For each scenario only certain types of adaptation invest- The costs of adaptation are also considered in the ments are considered. For example, in a B1EXwD scenario, there is no investment considered for a 1/10 flood since this context of whether the existing master plan (MP) is scenario already includes a dam, which is expected provide implemented or not (EX). protection for a 1/10 flood. 66 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble4.11 A daptationInvestmentsConsideredfordifferentReturnperiodsand ClimateScenarios Adaptation Investment Options/Costs Considered Climate Scenario 1/10 flood 1/30 flood 1/100 flood SQ SQ EX wD Pasig Marikina River embankment + Marikina Dam Pasig Marikina River embankment+ Marikina Dam SQ EX nD Storm surge barrier Pasig Marikina River embankment SQ MP wD Marikina Dam in addition to investments already 0 costs because flood prevention costs, considered under the MP including Marikina Dam, area already incorporated into the MP SQ MP nD 0 costs because additional flood prevention invest- ments are already incorporated under the MP B1 B1 EX wD Pasig Marikina River Embankment + Additional Pasig Marikina River embankment + embankment in Marikina River + Marikina Dam + Marikina Dam +additional embank- storm surge barrier ment in Pasig and Marikina River basin B1 EX nD Storm surge barrier + Pump Pasig Marikina River embankment +additional em- capacity improvement (KA- bankment in Marikina River + storm surge barrier MANAVA +West of Mangahan) B1 MP wD Additional embankment in Marikina River + Additional embankment in Pasig and Marikina Dam Marikina River basin B1 MP nD Additional embankment in Marikina river A1FI A1FI-EX wD Pasig Marikina River embankment +additional Pasig Marikina River embankment + embankment in Marikina River + Marikina Dam + Marikina Dam +additional embank- storm surge barrier ment in Pasig and Marikina River basin A1FI-EX nD Storm surge barrier + Pump Pasig Marikina River embankment +additional capacity improvement (KA- embankment in Pasig and Marikina River + storm MANAVA +West of Mangahan) surge barrier A1FI MPwD Additional embankment in Marikina River + Additional embankment in Pasig and Marikina Dam Marikina River basin A1FI MP nD Additional embankment in Pasig and Marikina River TAble4.12 I nvestmentCostsandnetpresentvalueofbenefitsAssociatedwith differentFloodControlprojectsinManila(php)usinga15percent discountrate 1/10 Flood 1/30 Flood 1/100 Flood Climate Scenario Investment Cost NPV Investment Cost NPV Investment Cost NPV SQ SQ EX wD NA 14,121,102,133 809,000,801 13,501,553,721 3,763,340,285 SQ EX nD 42,887,291 209,952,438 10,943,489,020 2,939,371,200 NA SQ MP wD NA 3,177,613,113 1,199,401,356 0.00 4,755,005,784 SQ MP nD NA 0.00 3,329,771,755 NA B1 B1 EX wD NA 14,232,087,722 5,634,294,466 13,604,450,310 10,150,493,344 B1 EX nD 1,003,222,253 (349,951,115) 11,054,474,609 7,764,664,865 NA B1 MP wD NA 3,216,390,949 3,512,076,308 102,896,589 7,587,409,494 B1 MP nD NA 38,777,837 5,642,446,707 NA (Continued to next page) AssessingdamageCostsandprioritizingAdaptationoptions | 67 TAble4.12 I nvestmentCostsandnetpresentvalueofbenefitsAssociatedwith differentFloodControlprojectsinManila(php)usinga15percent discountrate (Continued) 1/10 Flood 1/30 Flood 1/100 Flood Climate Scenario Investment Cost NPV Investment Cost NPV Investment Cost NPV A1FI A1FI-EX wD NA 14,248,304,696 7,092,797,122 13,640,673,269 12,011,488,435 A1FI-EX nD 1,409,166,226 (581,704,127) 11,099,925,438 9,203,568,238 NA A1FI MP wD NA 3,216,390,949 5,509,802,569 139,119,548 10,411,715,022 A1FI MP nD NA 68,011,692 7,620,573,684 NA Source: Muto et al. (2010). SQ=Status Quo, EX=Existing Infrastructure, B1, A1FI=Climate Change Scenarios, wD=With Marikina Dam, nD=No Dam, NA=not applicable. Costs are 0 in certain cases because it is assumed that these costs are incorporated in the master plan. count rate of 15 percent, which is what is used by the FIguRe4.9 A nnualbenefitsfrom Philippines National Economic and Development AdaptationInvestmentsin authority to estimate project feasibility.55 MetroManila 120,000 damconstructionemergesasan Damage cost (PHP in million) 100,000 economicallyviableoptioninbothclimate 80,000 changescenarios,followedbyembankment 60,000 buildinginthepasig-Marikinabasin 40,000 20,000 Table 4.12 presents the present value of net benefits 0 from different investments considered. The NPV 0.00 0.02 0.04 0.06 0.08 0.10 at12 billion PHP ($269 million) is highest among all Probability Existing infrastructure 1990 master plan level Full adaptation the different scenarios. This suggests that construct- (EX) (MP) level ing the Marikina Dam would maximize benefits relative to other adaptation options in the case of Source: Based on estimations in Muto et al. (2010). an A1FI or B1 scenario. The investments that would be required in this scenario are building the Pasig- Marikina River basin embankment, the Marikina full adaptation to an A1FI climate. This is what is Dam, and some additional embankments along the currently being undertaken by the government of Pasig and Marikina rivers. These investments would the Philippines in implementing the Pasig-Marikina largely eliminate floods in this part of Metro Manila. Flood Control Project Phase II to avoid damages from However, given that constructing the dam is a P30 floods. The recommendations in the context of decision that may or may not be taken, it is useful a B1 climate change scenario are similar. In sum, to consider what alternative options emerge. In this the first priority is to control flooding in the Pasig- case, in an A1FI scenario it is recommended that the Pasig Marikina River embankment be built with 55 It is important to note that the total avoided damages some additional components along the Pasig and (gross) in chapter 3 do not directly feed into the NPV calcu- Marikina rivers, and that storm surge barriers be lations. This is because there are three geographical areas, constructed. This basically means that investments and depending on the return period and adaptation invest- under the current master plan in the Pasig-Marikina ments, only the relevant benefits are considered for each scenario. (For example, the embankment for Pasig-Mariki- River basin should be prioritized and continued and na River does not affect KAMANAVA coastal area and there some additional investments need to be made for is thus no benefit in that area). 68 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Marikina River areas through a dam, embankments, area subject to flooding both in extreme events and storm surge barriers. While adaptation to climate and regular flooding was then determined under change would require significant investments, in the future scenarios using HydroGIS modeling. Once case of Metro Manila, these are investments that are average land price and flooding extent and dura- already planned. Thus, climate change adaptation tion were estimated, the relationship between the would require the government to commit to imple- flooding and the decline in the economic value of menting plans that are currently on the books. the flooded land was determined to calculate the cost of flooding due to climate change. The value of land affected by flooding is assessed assuming AnAlySISoFdAMAgeCoSTSIn linear and quadratic relationships between flood hCMC duration and land values. The second approach used by HCMC is to calculate costs on the basis of HCMC has a long history of extreme weather events. expected losses in production (proxied by GDP) due Between 1997 and 2007, almost all of the districts of to climate-change-induced flooding. For each dis- Ho Chi Minh City have been directly affected by trict and for each year (2006 to 2050), the annual cost natural disasters to some extent. The total value of of flooding was calculated and the discounted costs damage to property from natural disasters over the summed for the whole period to find the present last 10 years is estimated at over $12.6 million (202 value of expected lost GDP. In the following pages billion VND). Most impacts have been concentrated and the rest of this report, an average exchange rate in the predominantly rural Can Gio and Nha Be for 2008 of 1 USD = 16302.25 VND is used to convert districts. However, with increased levels of flooding Vietnamese dong to U.S. dollars. and extreme events due to climate change, urban areas are likely to suffer increasing levels of damage. This is likely to increase costs significantly. Macrolevelanalysessuggestthattherewill The HCMC study used a different approach besignificantcostsasaresultofclimate from Bangkok and Manila to estimate the aggregate change costs of climate change impacts to 2050. It does not The study estimates that the losses in the economic take a sectoral approach and estimate in detail the value of land affected by 2050 climate change would costs that will be incurred as a result of climate range from VND 100,358 billion ($6.15 billion) for change. While areas of vulnerability and risks are regular flooding to VND 6,905 billion ($0.42 bil- clear, there has been no attempt to value specific lion) for extreme flooding, assuming a quadratic risks directly. Rather, a first approximation of the relationship between flood duration and land value likely costs of climate change is undertaken based (Table 4.13).56 The HCMC study also reports cli- on macro data. Further, the HCMC study estimates mate change costs assuming a linear relationship the present value of flooding from the current period between flood duration and land values. In this up to 2050. Thus, it presents cost estimates today context, the cost of climate change is estimated to that reflect repeated flooding over a long period of be VND 369,377 billion ($22.7 billion) for regular time. In contrast, the Manila and Bangkok studies floods and VND 111,678 billion ($6.9 billion) for estimate the costs of specific single events of flood- extreme events. The highest costs (using either method) ing in different scenarios. are borne by Binh Chanh district, district 9, Can Gio, As discussed in chapter 2, the HCMC study and Nha Be district. uses two approaches to estimate the costs of climate change: (1) cost estimates based on expected lost 56 In all estimates, the costs of extreme events are much land values; and (2) cost estimates based on ag- smaller than those of regular flooding because (a) flooding gregate GDP loss. To recap briefly, the land value due to extreme events lasts for a smaller number of days method estimates how climate change may affect than that for regular flooding; and (b) the calculation of flooding assumes a return period of 30 years for extreme the value of the land stock in HCMC. The first step flooding. The expected value in any given year is therefore in the analysis was to determine land prices. The 1/30th of the cost of an extreme event. AssessingdamageCostsandprioritizingAdaptationoptions | 69 uncertaintyintheestimationofcosts changes in land values resulting from additional (and as of now unexpected) days of flooding result- These land-based estimates represent a first ap- ing from climate change. proximation of potential losses. There remains Using the GDP loss estimation method (Table considerable uncertainty as to how land values 4.14), the cost of regular flooding in terms of GDP may be impacted by increases in the frequency or loss is estimated to be about VND 806,831 billion duration of flooding. To the extent that the impacts ($49.5 billion) in present value terms. The cost of of existing flood events (climate variability) have extreme flooding is estimated to be approximately already been capitalized in the price of land, the VND 7,978 billion ($0.49 billion). above results should be interpreted as the possible TAble4.13 e xpectedCostofFloodingbasedonQuadraticRelationshipbetween durationofFloodingandlandvaluesinhCMC Flooded area (ha) Average duration (days) Land value Expected cost (billion VND) District Regular Extreme Regular Extreme (1,000VND/sq.m) Regular Extreme 1 34 249 81 12 22,410 380 60 2 3,036 4,115 127 22 1,556 5,719 233 3 0 105 0 0 17,407 0 0 4 78 348 21 6 7,369 19 7 5 12 243 95 10 12,946 103 24 6 45 594 4 2 8,508 0 2 7 1,004 2,451 52 9 4,070 830 61 8 655 1,768 12 4 4,318 31 9 9 5,877 6,696 140 28 1,621 14,014 639 10 30 198 88 11 10,759 189 19 11 14 99 73 11 7,245 40 7 12 2,241 2,465 140 29 2,043 6,735 318 Go Vap 184 455 65 12 4,072 237 20 Binh Thanh 685 1,383 62 10 10,127 2,001 105 Phu Nhuan 0 9 0 0 9,234 0 0 Thu Duc 1,593 1,913 95 21 3,995 4,312 253 Cu Chi 7,234 10,875 117 26 717 5,330 396 Hoc Mon 3,538 5,040 155 35 1,237 7,892 573 Nha Be 7,948 8,421 95 20 1,778 9,572 450 Can Gio 46,435 48,486 88 20 423 11,417 616 Tan Phu 0 45 0 0 4,796 0 0 Tan Binh 0 58 0 1 4,934 0 0 Binh Tan 917 2,300 29 12 1,932 112 48 Binh Chanh 18,890 22,057 83 24 3,217 31,423 3,068 Total 100,450 120,372 VND 100,358 6,905 Total (USD) 6.156083 0.4235612 Source: ADB (2010). Some differences due to exchange rate variations. 1 USD=16302.25 70 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport TAble4.14 p resentvalueoftheCostofFloodsupto2050usingthegdp estimationMethod Regular Flooding Extreme Flooding Total District GDP Loss (VND) GDP Loss (VND) GDP Loss (VND) District 1 4,262,678,022,700 118,930,920,400 4,381,608,943,100 District 2 91,872,408,828,200 684,250,653,100 92,556,659,481,300 District 3 -- 1,045,570,300 1,045,570,300 District 4 2,111,546,363,400 79,830,492,300 2,191,376,855,700 District 5 2,036,113,916,300 63,099,964,700 2,099,213,881,000 District 6 102,479,383,000 23,505,283,400 125,984,666,400 District 7 23,901,066,736,700 319,603,147,100 24,220,669,883,800 District 8 3,089,545,724,600 162,117,638,100 3,251,663,362,700 District 9 203,476,562,493,300 1,573,188,396,600 205,049,750,889,900 District 10 5,608,083,746,600 104,469,939,500 5,712,553,686,100 District 11 1,631,296,447,400 36,684,956,800 1,667,981,404,200 District 12 107,560,392,251,700 816,151,762,100 108,376,544,013,800 District Go Vap 13,980,278,430,700 30,422,891,000 14,010,701,321,700 District Tan Binh -- 1,135,984,100 1,135,984,100 District Binh Thanh 43,871,270,790,100 414,883,343,800 44,286,154,133,900 District Phu Nhuan -- 2,004,000 2,004,000 District Thu Duc 59,101,002,606,400 533,992,286,200 59,634,994,892,600 District Cu Chi 28,955,319,716,000 365,741,723,400 29,321,061,439,400 District Hoc Mon 44,762,782,087,100 565,835,862,900 45,328,617,950,000 District Binh Chanh 97,469,697,076,900 1,276,837,444,400 98,746,534,521,300 District Nha Be 42,592,220,679,200 331,433,560,700 42,923,654,239,900 District Can Gio 24,169,804,222,200 206,645,566,200 24,376,449,788,400 District Tan Phu -- 23,680,600 23,680,600 District Binh Tan 6,276,999,023,500 268,297,958,200 6,545,296,981,700 Total 806,831,548,546,000 7,978,131,029,900 814,809,679,575,900 USD 49,492,036,286 489,388,338 49,981,424,624 Source: ADB (2010). Some differences due to exchange rate variations. 1 USD=16302.25 Table 4.15 presents a summary of the cost the land value methodology. This may result from estimates from HCMC based on different method- two different factors. First, it is well known that ad- ologies. In summary, a first approximation of the ministratively determined land prices (as opposed costs of climate change for regular flooding events is between $6.15 billion and $49.5 billion in present value terms, and between $0.42 billion and $6.9 57 The HCMC numbers are much higher than the dam- billion for extreme flooding.57 age costs associated with flooding in Manila and Bangkok. The results of the two damage valuations show However, as previously noted, these numbers reflect the sum of a series of annual damages that are expected to oc- considerable divergence, with GDP figures sug- cur from now up to 2050 and are not directly comparable gesting double the damage costs achieved using with the damage costs of one event in Bangkok or Manila. AssessingdamageCostsandprioritizingAdaptationoptions | 71 TAble4.15 S ummaryofpresentvalue AnAlySISoFAdApTATIonIn ofClimateChangeCosts hCMC inhCMC(uSd) Unlike the Bangkok and Manila studies, the HCMC Regular Flooding Extreme Flooding Total study did not undertake a detailed cost benefit Land (linear) 22,658,038,001 6,850,465,427 29,508,503,427 analysis to prioritize adaptation options. While Land 6,156,082,749 423,561,165 6,579,643,914 it did provide an estimate of the viability of the (quadratic) government's flood protection project (Box 4.4), the GDP 49,492,036,285 489,388,337 49,981,424,623 main focus was to identify institutional mandates with respect to urban planning, flood protection, Source: ADB (2010). Some differences due to exchange rate variations. 1 USD=16302.25 and adaptation and propose a range of adaptation options that need to be undertaken in coordination with different sectors. Broadly, adaptation measures in the context of to market prices for land) undervalue land in the managing floods can be categorized as those that city. On this basis alone, had market prices for land involve (a) protection against predicted climate values been used instead of administratively deter- change, (b) accommodation to improve resilience, mined land prices (which were effectively used), the (c) retreat to reduce exposure, and (d) improved estimated cost of climate change using land values management (see Annex C).58 For instance, infra- would have been higher than presented earlier. structural measures such as construction of flood Second, GDP per capita figures at the district level embankments, polders, sea walls, and pumped were unavailable. This may overestimate the GDP drainage are common engineering solutions and loss across the city as rural districts in the south are being implemented in all three cities discussed such as Can Gio and Nha Be are responsible for in this report. "Accommodation" measures that very little GDP production, and yet which may be seek to minimize vulnerability include measures more vulnerable to climate change. Further detailed such as raising houses on stilts, adjusting cropping investigations and data collection would be required patterns, and revising building codes for housing to refine these figures. and industry. In some cases, where the risks of loss of life or assets is severe, "retreat" as a planning op- tion can reduce exposure to extreme events. Used in conjunction with restoring natural ecosystems that provide flood protection benefits, it can be a box4.4 Roughestimateof useful planning tool. Finally, flooding needs to be viabilityofproposedFloodControl considered in the context of overall water basin Measures management and the institutional capacity to man- age the resource. HCMC has already proposed to build a system of flood defenses that will significantly alter the pattern of flooding and the hydrol- ogy of the city. This project has an estimated capital cost of $750 developingsectorspecificadaptation million and is expected to be completed by 2025. The effects of the optionswithfocusonthepoor proposed flood control measures on regular flooding would seem to be significant in reducing the population exposed to flooding in In the HCMC study, a combination of these mea- 2050 from an estimated 10.2 million to 6.7 million, a reduction of 35 sures is proposed. Specifically, the study argues for percent. Based on the GDP method of estimation, the project would proportionately decrease damage costs by 35 percent, making this development of sector specific adaptation options an economically viable project. However, further economic analyses with a focus on the poor. For instance, in terms of this project needs to be carefully undertaken. Source: ADB (2010). 58 Drawn from IPCC, AR4, chapter 6, Coastal systems and low-lying areas, Figure 6.11 Evolution of planned coastal adaptation measures, pg. 342, 2007. 72 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport of reducing vulnerability of the poor, the study Strengtheninginstitutionalcapacityto recommends livelihood protection and livelihood adapttoclimaterelatedrisks diversification schemes, improved early warning systems, improved building construction require- The HCMC study argues that comprehensive ad- ments, land use planning and use of open space aptation planning is needed to provide the overall for flood management, zoning controls to ensure framework and direction for adaptation at the city that low income housing is located outside of level, in line with national level goals and targets. flood prone zones. HCMC already has a planned A key institutional recommendation arising from program of resettlement away from rivers and the study is the development of an HCMC Climate drains and this may require expansion to include Change Adaptation Plan. The HCMC Peoples Com- vulnerable areas (Can Gio and Nha Be districts) mittee (PC), the study argues, can take a proactive identified by the HCMC analysis. Priority adapta- leadership role in this context and in driving climate tion actions in the transport sector include review change adaptation in the city. Further, there are sev- and revision of design standards for roads, bridges, eral key agencies that shape overall land use, spatial and embankments so they are consistent with ex- zoning, environmental quality, and natural disaster pected flooding and climate conditions. Further, response management in the city, each of which in light of the findings of the study, new transport can pursue a range of actions within their sectoral infrastructure needs to be reassessed. Given that domain (Table 4.16). Since one sector or area's plan most industrial zones and clusters in HCMC will has implications on activities carried out by other be at direct risk of flooding with or without the sectors, coordination between different adaptation proposed flood control plan, a number of adapta- plans and planning processes is critical. tion measures--such as locating industrial zones outside of vulnerable areas and retrofitting existing infrastructure--are proposed. ConCluSIon Given the magnitude of climate change costs, adap- Combininginfrastructurebasedsolutions tation to climate change clearly needs to be a serious witheco-systembasedadaptationmeasures consideration. As discussed, adaptation invest- An important recommendation of the HCMC study ments are already under way in all three cities. Each is to combine infrastructure-based solutions with of the cities will be better protected against climate the use of ecosystem-based adaptation measures, change with the flood control measures that are which provide a buffer against climate risks. The either already planned or are slowly being imple- mangrove forests to the south of HCMC provide mented. However, additional policy, institutional, significant protection against storm surges, but are and ecosystem-based measures also need to be put under severe pressure due to land use change and in place and prioritized by the city governments. encroachment. Natural systems of the Dong Nai The analysis undertaken in the three case River basin provide a range of ecosystem services studies needs to be viewed as an initial attempt at such as regulation of hydrological flows, freshwater estimating the impacts and damage costs related storage, erosion control, and water purification. to climate change. As discussed in chapter 2, there These too are under threat due to land use change are a number of uncertainties associated with each and urban development. Adaptation approaches level of analysis. The climate downscaling and the suggested by the study include reforestation of hydrological models provide results that are by no the Dong Nai River basin watershed, restoration means certain. The economic analysis overlays a of wetlands, rehabilitation of canals and rivers, large number of assumptions related to prices, GDP, strengthening zoning regulations to protect ecosys- population distribution, growth, the structure of tem resilience, and planting of buffer zones along cities and so on, over and above these uncertainties. dykes and riverbanks, including dykes proposed Therefore, they should be viewed as a preliminary by the planned flood control system. attempt to be followed by improved studies. AssessingdamageCostsandprioritizingAdaptationoptions | 73 TAble4.16 p roposedImplementationArrangementsforhCMC Authority Priority actions HCMC DONRE (i) Revise the HCMC land use strategy and action plan to incorporate climate change issues and adapta- tion measures (ii) Prepare assessment guidelines for reviewing sector and spatial plans adaptation requirements and consistency with the city adaptation plan (iii) Prepare assessment guidelines for integrating adaptation in SEA and EIA when applied to devel- opment plans and project proposals HCMC Environment Protection Agency (HEPA) Prepare adaptation monitoring and audit guidelines to keep track of adaptation performance Department of Planning and Architecture Revise the HCMC urban strategy and plan to set out the spatial land uses, controls, and safeguards and its Institute for Urban Planning for adaptation Department of Construction with MOC Revise and pilot the Building Code in the city so that it responds to climate change Line departments and institutes responsible (i) Audit existing infrastructure and development plans and orientations for (a) transport, (b) power supply, (c) water (ii) Retrofit adaptation measures in existing infrastructure management and supply, (d) water quality (iii) Define sector-specific adaptation options and sanitation, (e) industry, (f) agriculture and fisheries, and (g) public health (iv)Upgrade sector design standards (v) Prepare strategies and plans for the next development period so they address climate change (vi)Introduce monitoring, auditing, and reporting on adaptation performance HCMC Steering Committee for Flood and (i) Support each commune and district in reviewing and revising their specific contingency plans to Storm Control protect and cope with more extreme flooding and storm events, and identifying the key assets and residential areas that need to be protected, up to and including evacuation of residents if necessary (ii) Improve early warning systems for floods, storms, tidal conditions, and drought (iii) Support ports, airports, and rail authorities in developing contingency plans in the event of major flood events (iv) Develop an early warning system for traffic and alternative transport routes in the event of floods Source: ADB (2010). 74 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Conclusions and 5 Policy Implications T he studies undertaken in the three cities show scenarios respectively are estimated for 2050 and the challenges faced by coastal megacities are linked with a 3 percent and 2 percent increase with respect to regional and global climate in mean seasonal precipitation for the high and low change and variability. These studies highlight the emissions scenarios. A flood that currently occurs importance of achieving sustainability in urban areas once in fifty years may occur as frequently as once through a better understanding of the relationship in 15 years by 2050, highlighting the potential for between urbanization, local hydrological systems, an increase in the frequency of extreme events. In and climate. All of the cities considered in this report Manila, the mean seasonal precipitation is expected are megacities with (registered and unregistered) to increase by 4 percent and 2.6 percent for the populations close to or greater than 10 million. Each high and low emissions scenarios. In HCMC, the is a driver of national and/or regional economic average annual temperature is expected to rise by growth and all depend on the capacity of local hy- 1.4°C in 2050 (with only slight differences between drological systems--such as rivers and streams, the high and low emissions scenarios). Further, a watersheds, and wetlands--to provide a range of 20 percent and 30 percent increase in precipitation services that are critical to the sustainability of these is expected for a 1-in-30-year and a 1-in-100-year cities and its residents. Given that these are all coastal event respectively under the high emission (A2) megacities, they face increased climate-related risks scenario. While extreme rainfall linked with storms such as sea level rise and increased frequency of is expected to become more common, preliminary extreme events. In this chapter, we conclude by analysis suggests that the frequency of dry season highlighting key findings and policy implications drought in HCMC is also likely to increase by 10 that would be useful for urban policy makers to percent under the low emissions scenario. These consider with respect to adaptation at the city level. risks need to be recognized and better understood by urban planners in the context of specific coastal cities, particularly in developing countries. KeyFIndIngSAndleSSonSFoR polICyMAKeRS Increaseinflood-proneareasduetoclimate changeinallthreecities Majorrisksareposedbyincreased temperatureandprecipitationincoastal In all three megacities in 2050, there is an increase in megacities the area likely to be flooded under different climate scenarios compared to a situation without climate All three cities are likely to witness increases in change. In Bangkok, for instance, under the condi- temperature and precipitation linked with climate tions that currently generate a 1-in-30-year flood but change and variability. In Bangkok, temperature with the added precipitation projected for a high increases of 1.9°C and 1.2°C for the A1FI and B1 emissions scenario, there will be approximately 75 a 30 percent increase in the flood-prone area. In still likely to be at risk from flooding during extreme Manila, even if current flood infrastructure plans events. How to plan for such large percentages of are implemented, the area flooded in 2050 will in- population being exposed to future flooding needs crease by 42 percent in the event of a 1-in-100-year to be seriously considered. flood under the high emission scenario compared to a situation without climate change. In HCMC, Costsofdamagelikelytobesubstantial the area inundated increases (from 54 percent in inandcanrangefrom2to6percent a situation without climate change to 61 percent ofregionalgdp in 2050 with climate risks considered) for regular events and for extreme (1-in-30 year) events from In Bangkok, the increased costs associated with 68 percent (without climate change) to 71 percent climate change (in an A1FI scenario) from a 1-in- (with climate risks considered) in 2050 under the 30-year flood is THB 49 billion ($1.5 billion) or high emission scenario. Further, there is a significant approximately 2 percent of GRDP. These are the ad- increase in both depth and duration for both regular ditional costs associated with climate change. The and extreme floods over current levels in 2050. The actual costs of a 1-in-30-year flood would include analysis also highlights areas that will be at greater costs resulting from land subsidence and would be risk of flooding in each metropolitan area. In Metro even higher. In Manila, a similar 1-in-30-year flood Manila, for instance, areas of high population den- can lead to costs of flooding ranging from PHP 40 sity such as Manila City, Quezon City, Pasig City, billion ($0.9 billion) given current flood control Marikina City, and San Juan Mandaluyong City are infrastructure and climate conditions to PHP 70 likely to face serious risks of flooding. billion ($1.5 billion) with similar infrastructure but future A1FI climate. Thus, the additional costs of Increaseinpopulationexposedtoflooding climate change from a 1-in-30-year flood would duetoclimatechange be approximately PHP 30 billion ($0.65 billion) or 6 percent of GRDP. The HCMC study adopts a dif- In all three cities, there is likely to be an increase in ferent methodology to analyze costs and its results the number of persons exposed to flooding in 2050 cannot directly be compared to the costs of Manila under different climate scenarios compared to a and Bangkok. The HCMC study uses a macro ap- situation without climate change. For instance, in proach and estimates a series of annual costs up to Bangkok, in 2050, the number of persons affected 2050. The flood costs to HCMC, in present value (flooded for more than 30 days) by a 1-in-30-year terms, range from $6.5 to $50 billion. The "annual- event will rise sharply for both the low and high ized" costs of flooding would likely be comparable emission scenarios by 47 percent and 75 percent to the costs of Bangkok and Manila. respectively compared to those affected by floods in a situation without climate change. In Manila, for a damagestobuildingsisanimportant 1-in-100-year flood in 2050, under the high emission componentofflood-relatedcosts scenario more than 2.5 million people are likely to be affected assuming that the infrastructure in 2050 Damage to buildings emerges as a dominant com- is the same as in the base year, and about 1.3 million ponent of flood-related costs, at least in Bangkok people if the 1990 master plan is implemented. In and Manila. In these cities, over 70 percent of HCMC, currently, about 26 percent of the population flood-related costs in all scenarios are a result of would be affected by a 1-in-30-year event. However, damage to buildings. Cities are, almost by defini- by 2050, it is estimated that approximately 62 percent tion, built-up areas full of concrete structures. It of the population will be affected under the high is, therefore, not surprising that the main impacts emission scenario without implementation of the of floods is on these structures and the assets they proposed flood control measures. Even with the carry. In HCMC, 61 percent of urban land use, 67 implementation of these flood control measures, percent of industrial land use, and 77 percent of more than half of the projected 2050 population is 76 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport open land use is expected to be flooded in 2050 ered for the infrastructure to provide the expected in an extreme event if the proposed flood control level of protection from future climate risks. In measures are not implemented. Future flooding HCMC for example, which has a long history of re- also has major sector implications. For instance, in sponding to natural disasters, a number of measures the transportation sector in HCMC, even with the such as dykes and early warning systems (which are implementation of the proposed flood protection also important aspects of a climate change adapta- system, a significant percentage (76 percent of axis tion strategy) have been developed. However, the roads, 58 percent of ring roads) of the existing and likelihood of increasing frequency and intensity of planned roads will be exposed to increased flood- extreme events has not been built into these disaster ing. Thus, as cities develop over the next 40 years, management plans. Consideration of climate-related it would be important to design their commercial, risks has also not been sufficiently integrated into residential, and industrial assets and zones to mini- urban planning and in various sectoral plans in the mize these costs. cities. As the HCMC case shows, in addition to the urban master plan, there are a number of city-level Impactonthepoorandvulnerablewill sectoral plans with little coordination between them. besubstantial;however,evenbetter-off The main policy implication is that climate change communitieswillbeaffectedbyflooding adaptation measures need to be well-integrated into urban, sectoral, and flood protection plans and In Bangkok and Samut Prakarn, the study es- the coordination between these plans needs to be timates that about 1 million inhabitants will be strengthened. affected by the A1FI climate change condition in 2050. One out of eight of the affected inhabitants Reductionoflandsubsidenceshouldbe will be those living in condensed housing areas consideredanimportantpartofurban where the population primarily lives below the adaptation poverty line. Of the total affected population, ap- proximately one-third may have to encounter more One of the main findings of this study is that non- than a half-meter inundation for at least one week, climate-related factors such as land subsidence are marking a two-fold increase in the vulnerable important and in some cases even more important population. People living in the Bang Khun Thian than climate risks in contributing to urban flood- district of Bangkok and the Phra Samut Chedi dis- ing. In Bangkok, for instance, there is nearly a trict of Samut Prakarn will be especially affected. two-fold increase in damage costs between 2008 In HCMC, in general, poorer areas in HCMC are and 2050 due to land subsidence. Further, almost more vulnerable to flooding. However, in some of 70 percent of the increase in flooding costs in 2050 the areas, the poor and non-poor are both at risk. in the city is due to land subsidence. While data The Manila study shows that while low-income for land subsidence was not available for Manila residents have devised many coping strategies to and HCMC and this issue was not considered in extreme events, state and non-state actors need to the hydrological modeling for these two cities, play a more proactive role in addressing vulner- available literature suggests that it is an important ability of the urban poor. factor in all three cities and should be considered in follow-up studies. Even though the megacities urbanplansandfloodprotection have already undertaken a number of measures infrastructureneedtotakeclimaterisks to slow down land subsidence, further regulatory intoaccount and market incentives are clearly required to stem groundwater losses. City governments need to Flood protection plans are already in place in all better assess factors contributing to land subsid- three mega-cities considered. However, in all three ence and consider options to reduce it. Based on cities, these have been prepared without considering qualitative information provided in the city studies future climate-related risks, which need to be consid- --the impact of other non climate factors such as ConclusionsandpolicyImplications | 77 dumping of solid waste into the city's canals and that serve urban residents. These options need to waterways, poor dredging of canals, siltation of be considered in developing a comprehensive ap- drains, deforestation and soil erosion in the upper proach to urban adaptation. Dykes themselves are watershed, and land use patterns--are also critical susceptible to erosion by rainfall and storms. As and should be part of a broader and multisectoral such, it is also important to consider protection of approach to addressing urban adaptation. dykes through planting. Importanttoconsidervariationinclimate Importanceoflocallyinducedclimate risksandfactorscontributingtoflooding changefactors An important finding of this study is that while Even though this report does not take into account a combination of climate-related factors can con- locally induced climate change factors, these are tribute to urban flooding, some factors are much very significant (Grimm et al. 2008). For example, more important than others in different cities. For the rate of increase of temperature in HCMC is al- instance, in HCMC, storm surges and sea level most double that of the surrounding Mekong Delta rise are important factors contributing to flood- region. Between 1997 to 2006, the average tempera- ing. However, in Bangkok these factors seem to ture of HCMC increased .34° C compared to .16° C be relatively less important. In part this has to for the Mekong Delta region. This increase in tem- do with the location, elevation, and topography perature for HCMC has coincided with increasing of the city. The policy implication is that adapta- urbanization in the city. In HCMC, the heat island tion measures need to be designed based on the effect is changing the climate and urbanization is specific hydrological and climate characteristics having a significant effect on increases in tempera- of each city. ture, rainfall, and flooding in and around HCMC. Further analysis is needed to better understand needforgreateremphasisonecosystem- this issue. basedapproaches In all of the cities considered, an important fo- leSSonSonMeThodology cus of flood management and control is on hard FoRFollow-upSTudIeS infrastructure solutions. However, the studies (Bangkok and HCMC) show that ecosystem-based Even though analysis similar to that undertaken approaches are also important and can be usefully in this report has been carried out for cities in de- combined with infrastructure interventions. The veloped countries (Rosenzweig et al. 2007), it has HCMC study, for instance, argues for combin- so far not been done in developing country cities. ing construction of dykes (as approached by the Moreover, preparation of this report has spurred planned flood control measure) with management similar studies in other cities in the Africa and and rehabilitation of the mangrove systems in Can Middle East and North Africa regions. It is impor- Gio, reforestation of the Dong Nai River Basin up- tant therefore to conclude with a few lessons for per watersheds, river and canal bank protection, follow-up studies. and implementation of basin-wide flow manage- ment strategies. Urban wetlands provide a range Approachestodownscaling of services, including flood resilience, allowing groundwater recharge and infiltration, and provid- In this study two different approaches to down- ing a buffer against fluctuations of sea level and scaling were used, namely statistical scaling used storm surges. Rehabilitation of urban wetlands is by JICA and use of a dynamic regional model as therefore critical. Upstream protection of water- done by the HCMC study. Each method has its sheds through reforestation is also important in strengths and limitations. Dynamic regional models managing the release of runoff into the reservoirs can simulate daily information on a broad range of 78 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport hydrometeorological data, yielding information on change adaptation studies. These synergies need to floods and droughts and on the joint occurrence of be recognized and strengthened. variables, like the pattern of precipitation over a large watershed. Dynamic models that are nested lessonswithrespecttodamagecost in AOGCMs are necessarily bound to the parent analysisandprioritizingadaptationoptions model's intrinsic biases. There is, for example, a wide variation in the results of the AOGCMs in The three studies were based on available data. This forecasting extreme precipitation, and use of one meant that various assumptions had to be made model alone can yield unreliable results. Statistical regarding the number of households affected, the models yield single parameters and shed little light "damage rate" or extent of damage to various assets on the joint probability of occurrence. Neverthe- as a result of floods of different intensity, and the less, because the results are based on a correlation costs associated with these damages. The underly- with the results of many models, the downscaled ing data and valuation would be much improved if parameters have less intrinsic bias than those gener- there was scope for survey-based data collection. A ated by a single model. These constraints need to be rather important gap in the studies is the analysis considered in undertaking similar city-level studies. of health impacts. Floods do have serious health consequences and while the studies have attempted needtoforgestrongerlinksbetween to estimate these, the full importance of these im- methodologiesdevelopedbydisasterrisk pacts has not been captured. To carefully examine reductioncommunityandclimatechange the health impacts of floods would require first an adaptation estimation of disease prevalence in the future (tak- ing into account climate and economic conditions) There is a strong global consensus on the increasingly and second, application of robust dose-response urgent need to address the underlying causes of cli- relationships to establish the extent of flood-related mate change and its impacts through mitigation of illnesses. This is an impact issue that requires more greenhouse gas emissions and development of effec- careful consideration. In terms of adaptation to tive adaptation strategies. This has been underscored climate change, the Bangkok and Manila studies at the recent United Nations Framework Convention are unable to incorporate "soft options" such as on Climate Change (UNFCCC) Conferences of Par- changes in legal and institutional issues into their ties (COP) in Copenhagen (COP-15), which again cost-benefit analyses. Thus, it is possible that there highlighted the need to foster better linkages between are less costly policy changes that might bring climate change adaptation (CCA) and disaster risk about the same type of results as the engineering reduction (DRR). The Hyogo Framework of Action, solutions identified. However, identifying the full signed by 168 countries in 2005, is the guiding docu- implications of different policy reforms and then ment for the DRR initiatives to mainstream DRR into incorporating these into a cost-benefit framework the national, sectoral, and local-level development is difficult. Future studies would do well to look planning process. The DRR community has well- at behavioral change options, zoning possibilities, established procedures and terms of reference for risk legal and institutional reforms, and conservation assessment, including probabilistic risk assessment, possibilities carefully and rank them relative to that allow assessment of risks based on an analysis engineering solutions. of hazards, exposure, vulnerability, and capacity (ECLAC 2003). The use of simulation models to de- needfordetailedinstitutionalanalysisin velop event-loss relationships are likewise commonly futurestudies applied in the DRR work (e.g., hydrometeorological, earthquake, storm surge, tsunami, and landslide In undertaking similar studies in the future, a more models). In addition, the DRR community has the detailed analysis of institutional and organizational local networks and knowledge management sys- capacity for urban adaptation is suggested. While tems that will provide significant benefit for climate all three city studies considered in this report do ConclusionsandpolicyImplications | 79 consider broad institutional issues, a more detailed forecasting the increase in extreme and seasonal capacity analysis could provide a more nuanced precipitation under the different scenarios. The understanding of policy and regulatory gaps, exist- techniques applied in the statistical downscaling ing planning and decision-making processes, coor- examined the results from numerous AOGCMs. dination between key municipal organizations, and Robust relationships were identified for tempera- entry points for strengthening state, private sector ture and precipitable water increases, where the and civil society relations with respect to addressing internal error was estimated at ~ 10 percent error urban adaptation. and ~ 10­20 percent error, respectively (Sugiyama 2008). Hydrologic models can simulate flood events uncertainties,limitations,andinterpreting with relatively small errors (<10 percent) if sufficient thefindingsofthisstudy data is available for good calibration. For future forecasts, however, land use changes in the water- Any study forecasting conditions four decades sheds and drainage areas can dramatically affect hence will be faced with large uncertainties. One flood patterns and can be further examined in future issue facing the analysis was what would be the sensitivity analyses. In addition, each of the studies pathway of GHG emissions. To address that issue, makes numerous assumptions regarding what each the city case studies examined both a high and a city will look like (for example in terms of growth, low GHG emissions scenario to bracket the likely population increase) in 2050. These uncertainties future conditions. In the climate change down- and errors need to be considered in interpreting the scaling methodologies, there are uncertainties in results of this study. 80 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Bibliography Allen, M. R., and W. J. Ingram. 2002. "Constraints Giorgi, Filippo. 2008. "Regionalization of climate on future changes in climate and the hydrological change information for impact assessment and cycle." Nature 419: 224­232. adaptation." WMO Bulletin 57 (2): 86­92. ALMEC Corporation. 2008. Final Report on the Grimm, N. B. et al. 2008. "Global Change and the Study on Climate Change Impact over Asian Mega Ecology of Cities." Science 319 (5864): 756­760. Cities (Phase 2): Metro Manila. Final Report to JICA. Manila: ALMEC. Gupta, A. et al. 2009. @ 110 Km/hr. Down to Earth. June 16­30, 2009. American Society of Civil Engineers. 1996. Hydrology Handbook. ASCE Press: New York. Ibaraki University, Institute for Global Change Ad- aptation Science. 2008. Climate Change Impact over Asian Development Bank (ADB). 2010. Ho Chi Minh Asian Mega Cities: Advisory for Coastal Engineering and City. Adaptation to Climate Change. In collaboration Adaptation. Paper Commissioned by JICA. Ibaraki: with the HCMC People's Committee and DONRE. Ibaraki University. Prepared by ICEM - International Center for Envi- ronmental Management. Intergovernmental Panel on Climate Change (IPCC). 2000. Special Report on Emissions Scenarios. Asian Institute of Technology. 2004. An Assessment http://www.ipcc.ch/pdf/assessment-report/ar4/ of the Socioeconomic Impacts of Floods in Large Coastal syr/ar4_syr.pdf Areas. Final Report for APN CAPaBLE Project: 2004-CB01NSY-Dutta. Intergovernmental Panel on Climate Change (IPCC). 2007a. Climate Change 2007: Synthesis Report. CTI Engineering International Co. Ltd. 2008. The http://www.ipcc.ch/pdf/assessment-report/ar4/ Study on Climate Impact Adaptation and Mitigation syr/ar4_syr.pdf in Asian Coastal Mega Cities -Case of Metro Manila. Final Report to JICA. Tokyo: CTI Engineering In- Intergovernmental Panel on Climate Change ternational Co. Ltd. (unpublished). (IPCC). 2007b. Climate Change 2007. Impacts, Ad- aptation and Vulnerability. Contribution of Work- ECLAC (Economic Commission for Latin America ing Group II to the Fourth Assessment Report of and the Caribbean). 2003. Handbook for Estimating the Intergovernmental Panel on Climate Change Socioeconomic and Environmental Effects of Disasters. (M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van Mexico: ECLAC. der Linden and C.E. Hanson, eds.) Cambridge, UK: Cambridge University Press. http://www. General Statistics Office (GSO) Vietnam. 2006. ipcc.ch/publications_and_data/ar4/wg2/en/ Vietnam Household Living Standard Surveys, 2002 and spm.html 2004: Basic Information, Hanoi, Vietnam. 81 IMF. 2009. World Economic Outlook Database-October Metro Manila. JICA Manila Report forthcoming as 2009. Washington, DC: International Monetary JICA Research Institute Working Paper. Fund. National Economic and Social Development Board Jonkman, S. N, M. Brinkhuis-Jak, and M. Kok. 2004. (NESDB). 2007. "Poverty Line." Bangkok. http:// "Cost benefit analysis and flood damage mitigation www.nesdb.go.th/Default.aspx?tabid=402 in the Netherlands." Heron 49 (1): 95­111. Nicholls, R.J. et al. 2007. "Coastal Systems and Jones, R et al. 2004. Generating High Resolution Low Lying Areas." In M.L. Parry, O.F. Canziani, Climate Change Scenarios using Precis. National J.P. Palutikof, P.J. van der Linden, and C.E. Hanson, Communications Support Unit Handbook. UNDP, eds. Climate Change 2007. Impacts, Adaptation and GEF and the Hadley Center. Exeter, UK: Met Office Vulnerability. Contribution of Working Group II to Hadley Center. the Fourth Assessment Report of the Intergovern- mental Panel for Climate Change. Cambridge, UK: McGranahan, G., D. Balk, and B. Anderson. 2007. Cambridge University Press. Available at: zones." Environment and Urbanization 19 (1): 17­37. Nicholls, R.J. et al. (2008). "Ranking Port Cities Meehl, G. A., T. F. Stocker, W.D. Collins, P. Friedling- with High Exposure and Vulnerability to Climate stein, A.T. Gaye, J.M. Gregory, A. Kitoh, R. Knutti, Change: Exposure Estimates." Environment Work- J.M. Murphy, A. Noda, S.C.B. Raper, I.G. Watterson, ing Paper No. 1. ENV/WKP (2007)1. Paris: OECD. A.J. Weaver, and Z. C. Zhao. 2007. "Global Climate Projections." In S. Solomon, D. Qin, M. Manning, Panya Consultants Co. Ltd. 2009. Climate Change Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and Impact and Adaptation Study for the Bangkok Metro- H.L. Miller, eds. Climate Change 2007: The Physical politan Region. Final Report Commissioned by the Science Basis. Contribution of Working Group I to the World Bank. Bangkok. Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New Rosenzweig, C., D.C. Major, K. Demong, C. Stanton, York: Cambridge University Press. R. Horton, and M. Stults. 2007. "Managing climate change risks in New York City's water system: As- Mehrotra, S. et al. 2009. "Framework for City Cli- sessment and adaptation planning." Mitig. Adapt. mate Risk Assessment: Buenos Aires, Delhi, Lagos Strategies Global Change 12: 1391­1409. and New York." World Bank Commissioned Re- search. Fifth Urban Research Symposium, Cities Satterthwaite, D. et al. 2007. "Building Climate and Climate Change, Marseille, France. Change Resilience in Urban Areas and Among Urban Populations in Low and Middle Income Messner, F., E.C. Penning-Rowsell, C. Green, V. Nations." Paper presented at Building for Climate Meyer, S.M. Tunstall, and A. Van der Veen. 2007. Resilience: Center for Sustainable Urban Develop- "Evaluating Flood Damages ­ Guidance and Rec- ment, New York, July 8­13. ommendations on Principles and Methods." Report No T09-06-01. Floodsite. http://www.floodsite. Smith, K, and Ward. R.1998. Floods: Physical Process net/html/partner_area/project_docs/T09_06_01_ and Human Impacts. Chichester: John Wiley and Flood_damage_guidelines_D9_1_v2_2_p44.pdf Sons. Muto, M., K. Morishita, and L. Syson. 2010. Impacts Solecki, W., and C. Rosenzweig. 2008. "Climate of Climate Change on Asian Coastal Areas: The case of Change Adaptation Strategies and Disaster Risk 82 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Reduction: Inventory and Prospect." GFDRR White World Bank. 2010. The Cost to Developing Countries Paper. (Draft.) of Adapting to Climate Change: New Methods and Es- timates. Washington, DC: World Bank. Stern, N. 2007. The Economics of Climate Change ­ the Stern Review. Cambridge, UK: Cambridge Univer- WWF. 2009. Mega Stress for Mega Cities: A Climate sity Press. Vulnerability Ranking of Major Coastal Cities in Asia. Gland, Switzerland: WWF International. Sugiyama, M. 2008. Study on Climate Change Adapta- tion and Mitigation in Asian Coastal Mega-cities. Inte- Yohe, G., J. Neumann, and H. Ameden. 1995. "As- grated Research System for Sustainability Science sessing the Economic Cost of Greenhouse-Induced (IR3S), University of Tokyo. Report Commissioned Sea Level Rise: Methods and Application in Sup- by JICA. port of a National Survey." Journal of Environmental Economics and Management 29:S-78­S-97. Sugiyama, M. 2009. "Precipitation extreme changes exceeding moisture content increases in MIROC Yusuf, A.A., and H.A. Francisco. 2009. "Climate and IPCC climate models." Proceeding of the Na- Change Vulnerability Mapping for South Asia." Sin- tional Academy of Sciences of the United States 107 gapore: EEPSEA. Available at: United Nations. 2007. World Urbanization Prospects: Zhang, Yabei, and Jan Bojö. 2009. "Note on the Cost 2007 Database. http://www.un.org/esa/popula- Benefit Analysis for the Flood Protection Improve- tion/publications/WUP2005/2005WUP_DataTa- ment Project." Prepared as a background note to the bles11.pdf Bangkok Study. (Draft.) bibliography | 83 Vulnerability of Kolkata Metropolitan Area Annex to increased Precipitation A in a Changing Climate globAlAndRegIonAl Studyobjectives ConTexT This study sought to: Low-lying coastal areas are highly vulnerable to Compile a database with past weather-related flooding. A number of recent studies have shown information and damage caused by extreme that coastal areas are vulnerable to a range of weather-related episodes. risks related to climate change, including coastal Determine scenarios most appropriate for as- flooding. Among these coastal areas, the Intergov- sessing the impact of climate change. ernmental Panel on Climate Change specifically Develop hydrological, hydraulic, and storm identifies as hotspots the heavily urbanized mega- drainage models to identify vulnerable areas cities in the low-lying deltas of Asia (IPCC 2007 b). and determine physical damage estimates re- Among Asian countries, India-- with its 7,500 km sulting from climate change effects. long predominantly low-lying and densely popu- Assess monetary, social, and environmental im- lated coastline--is particularly vulnerable. A recent pacts resulting from such climate change events. global survey identified Kolkata and Mumbai as Formulate adaptation proposals to cope with among the top ten cities with high exposure to damage arising from climate change effects. flooding under the current climate change fore- Strengthen local capabilities so that Kolkata's casts (Nicholls et al. 2008). The study also shows planning process can account for climate-related that exposure will increase in the future; by 2070, damage effects in analyzing new projects. Kolkata is expected to lead the top 10 list in terms of population exposure. In this study, precipitation events in Kolkata In response, the World Bank--in collaboration based on available historical rainfall data for 25 with the Asian Development Bank (ADB) and the years were used as a baseline (without climate Japan Japan International Cooperation Agency change) scenario. For modeling climate change, (JICA)--launched four country-level studies for predictions for temperature and precipitation Manila (led by JICA), Ho Chi Minh City (led by changes in Kolkata in 2050 for A1FI and B1 emission ADB), and Bangkok and Kolkata (led by the World scenarios--from an analysis of 16 GCMs used for the Bank). IPCC Fourth Assessment Report--were provided by 85 a paper commissioned by JICA (Sugiyama 2008). A use in slums make these areas highly vulnerable to sea level rise of 0.27 m by 2050 was also added to extreme weather-related events, especially flooding. the storm surge for the A1FI and B1 climate change scenarios based on current estimates. All these sce- narios (without and with climate change effects) MeThodology were then modeled to assess the impact in terms The study modeled the impact of climate change of the extent, magnitude, and duration of flooding. on increased flooding in KMA. The main causes of flooding in KMA are intense precipitation, overtop- ping of the Hooghly River due to water inflow from geogRAphICAnd local precipitation as well as that from the catchment SoCIoeConoMICConTexTS area, and storm surge effects. Land subsidence-- The geographic area covered in the study is the which occurs in a few pockets--was not included Kolkata Metropolitan Area (KMA),. a continuous in the study. The study covered three main sources urban area stretching along the east and west bank that aggravate.flooding in KMA: of the Hooghly River surrounded by some rural areas lying as a ring around the conurbation and Natural factors. Natural factors include flat acting as a protective green belt. KMA has an area topography and low relief that cause riverine of 1,851 square kilometers and consists of a complex flooding and problems with drainage. set of administrative entities comprising 3 municipal Developmental factors. Developmental factors in- corporations (including Kolkata Municipal Corpo- clude unplanned and unregulated urbanization;. ration, or KMC), 38 other municipalities, 77 non- low capacity drainage; sewerage infrastructure municipal urban towns, 16 out growths, and 445 that has not kept pace with the growth of the rural areas. KMC, the core of the city, lies along the city or demand for services; siltation in available tidal reaches of the Hooghly and was once mostly channels; obstructions caused by uncontrolled a wetland area. The elevation of KMA ranges from construction in the natural flow of storm wa- 1.5 to 11 meters above sea level (masl). The eleva- ter; and reclamation of natural drainage areas tion of KMC area ranges from 1.5 to 9 masl with an (marshlands). average of 6 masl. Climate change factors. Climate change factors With a population of about 14.7 million (includ- include an increase in the intensity of rainfall, ing 4.6 million in KMC), KMA is one of the 30 largest sea level rise, and an increase in storm surge megacities in the world.(United Nations 2007). The caused by climate change effects. average population density in KMA is 7,950 people per km2; in KMC, it is 23,149 per km2. The average Flooding from intense precipitation was mod- per-capita income in KMA in 2001­02 was $341 (at eled for three scenarios--30-year, 50-year, and 1993­94 prices). 100-year return period flood events--assuming no climate change effects. The climate change effects were added to the 100-year flood event using the TheSTudyAReA A1FI and the B1 scenarios. Assumptions about the impacts of climate A special characteristic of KMA is its large slum change in Kolkata in 2050 (Sugiyama 2008) in- population, comprising more than a third of the total cluded (a) a temperature increase of 1.8°C for the population. These slums not only lack basic infra- A1FI scenario and 1.2°C for the B1 scenario; (b) a structure and services, but are also the hub of many fractional increase in the precipitation extremes of informal manufacturing activities, some of which in- about 16 percent for the A1FI and 11 percent for the volve highly toxic industries. Little oversight of such B1 scenarios; and (c) sea level rise of 0.27 m by 2050, activities is carried out by government agencies. This which was added to the storm surge for the A1FI mixed residential and commercial/industrial land and B1 climate change scenarios based on current 86 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport estimates. All these scenarios (with and without because of data limitations. The vulnerability analy- climate change effects) were then modeled to assess sis was based on three separate indices: (1) a flood the impact in terms of the extent, magnitude, and vulnerability index based on the depth and duration duration of flooding. of flooding, (2) a land-use vulnerability index based Three separate models were used to capture on the nature of land use, and (3) a social vulnerabil- the overall effect of natural, developmental, and ity index based on the existing infrastructure and climate change factors that lead to flooding in KMA. the socioeconomic characteristics of the population. A hydrological model (SWAT model) was used Finally, the three separate indices were combined to to develop the flow series for the whole Hooghly form a composite vulnerability index. catchment. The generated data was then fed into a Vulnerability was assessed at both the ward and hydraulic model (HECRAS model) to analyze the subward levels. To evaluate the impact of flooding in implication of the flood passing through the river KMC, the analysis identified the 9 most vulnerable stretch. Finally, a storm drainage model (SWMM wards that may need specific attention in design- model) was used to determine the flooding that will ing adaptation strategies. To assess vulnerability result once the river flooding is combined with local within each ward in greater detail, the analysis was precipitation and drainage capability of the urban extended to the sub-ward level using spatial data. area under an extreme flood situation. dAMAgeASSeSSMenT vulneRAbIlITyAnAlySIS The vulnerability analysis was followed by a The models generated the increase in depth, dura- separate economic damage assessment. Due to tion, and extent of flooding in Kolkata due to climate data limitations, this was also restricted to the KMC change effects. A separate vulnerability analysis area. Damage to stocks measured primarily physical was done to assess the impact of flooding; this part damage arising out of water submersion (sectors of the analysis was restricted to only the KMC area included residential buildings and property, com- IBRD 37779 West Bengal Kolkata Metropolitan Area INDIA SEPTEMBER 2010 Area: 88752 sq. km. Area: 1851.41 sq. km. WEST BENGAL STATE Population: 80 million Pop. Density: 904/ Population: 14.7 million Pop. Density: 7950/ AND THE CITY OF KOLKATA sq. km. sq. km. River Kolkota y ghl Hoo West INDIA Bengal Kolkata Arabian Kolkata Municipal Corporation Sea Area: 187 sq. km. Bay of Population: 4.57 million Bengal Pop. Density: 23149/sq. km. r ve H o o g h ly Ri Garden Reach Tiljala This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any INDIAN OCEAN endorsement or acceptance of such boundaries. vulnerabilityofKolkataMetropolitanAreatoincreasedprecipitationinaChangingClimate | 87 mercial and industrial establishments, and major 74, 80, and 108. Six wards (14, 57, 58, 66, 67, and public infrastructure). Flow damage included loss of 108) in the eastern part of the city and ward 80 are income and increased morbidity, which are primar- vulnerable because of inadequate infrastructure, ily linked to the duration of a flood. The damage unplanned land use, and poor socioeconomic and estimates were based on data extrapolated to the environmental conditions. Infrastructural problems KMC population in 2050. The analysis assumed no are getting worse with increased building activity, additional investments in flood protection measures as these areas have become attractive to developers that may be implemented in the future to lower after becoming part of KMC. flood damage. Inflation also was not considered, The other two wards--63 and 74--are highly and all estimates used 2009 prices. Damage assess- vulnerable due to their topography. In these wards ments were estimated for the 100 year flood return the capacity of the sewerage system has not kept period and the A1FI climate change scenario to pace with changes in population. These have been determine the additional damage caused by climate further aggravated by inadequate maintenance, change effects. as well as the siltation of the existing trunk sewer systems, which have considerably reduced their car- rying capacity. While the sewer networks in KMC AdApTATIonAnAlySIS under such partially silted condition still provide reasonable hydraulic capacity for carrying the dry Finally, a separate analysis was done to examine weather flow, they are inadequate for carrying storm adaptation measures in KMC that can alleviate weather flow, even with normal precipitation during some of the problems posed by flooding. The the rainy season. analysis mainly focused on gains from the complete Additional losses likely to occur due to climate de-siltation of trunk sewers by modeling flooding change. Damage from a 100-year flood will increase under a completely de-silted trunk sewer scenario. by about $800 million--to more than $6.8 billion in The study also examined other proposals to build 2050--due to climate change (A1FI scenario). The new sewers and upgrade sewers in vulnerable areas, impacts by sector (in Indian rupees at 2009 prices) as well as institutional changes that can help cope are shown in the chart below. Local currency was with future flood damage. converted to dollars using the purchasing power parity index for India of 2.88 (IMF 2009). The larg- est damage components--under both the 100-year MAInFIndIngS return period flood and the A1FI climate change The most vulnerable wards to climate change. The scenario--are for residential property and buildings study identified the most vulnerable wards to cli- and health care. Commerce, industry, and other mate change--namely, wards14, 57, 58, 63, 66, 67, infrastructure like roads and transport services also TotallossesinmajorsectorsinKMC(Rsmillionin2050) 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 Residential Residential Residential Commerce Industry Healthcare Roads Transport Electricity building property income loss Current climate 100 year ood A1F1 climate 100 year ood 88 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport sustain significant damage. Due to data constraints, The strategy should protect major urban servic- some impacts could not be quantified in this analy- es, including roads, traffic, water supply, electric- sis, so the estimates provided are likely to understate ity, and telecommunications. It should recognize the overall impact of climate change. the importance of open space and green areas as Investing in de-silting of trunk sewers will an integral part of city development. reduce the area and population affected by floods. The strategy should spell out the climate risks The study looked at the impact of investing in and mitigating factors needed in operational de-silting trunk sewers both in the town and plans for key relevant agencies. suburban systems of KMC. The business as usual scenario considers an average 30 percent silting in Investing in soft infrastructure. To ensure long- trunk sewers. The adaptation scenario considers term financial, institutional, and environmental investing in de-silting and reducing it to zero. The sustainability, the adaptation strategy should also findings indicate that this simple investment can include: reduce the area affected by a flood by 4 per cent and the population affected by floods by at least Strengthening disaster management and 5 percent. preparedness for both pre- and post-disaster situations. Enforcing land use and building codes to reduce AdApTATIonMeASuReS obstruction and encroachment of floodplains and environmentally sensitive areas such as The current adaptation deficit..Climate change is canal banks and wetlands and to prevent con- likely to intensify urban flooding through a combi- version of green spaces and natural areas that nation of more intense local precipitation, riverine can act as retaining zones during flooding. flooding in the Hooghly, and coastal storm surges Introducing sustainable financing--emphasiz- A major cause of such periodic flooding during the ing both cost reduction and cost recovery--for rainy season is the city's current adaptation deficit, infrastructure investment and maintenance. including deficiencies in physical infrastructure, Increasing the budget for sewerage and problems with land use, and socioeconomic and drainage maintenance and the allocation of environmental factors. money for silt removal and mechanical sewer Adaptation strategy. The city needs a com- cleaning. prehensive and effective strategy that invests in Adopting flood insurance that incorporates suit- both soft and hard infrastructure to tackle flooding able incentives for adaptation and minimizes problems in Kolkata. The goal of the strategy is to (a) flood damage. reduce the percentage of people affected by flood- Strengthening the regulatory and enforcement ing and sewage-related diseases in KMC, and (b) process, including improving institutional man- target the most vulnerable areas. The strategy should agement and accountability. include preparedness both before and during the Enforcing pollution management frame- event, as well as post-event rehabilitation strategies. works, including introduction of incentives Investing in hard infrastructure. Investing in and disincentives to ensure compliance with hard infrastructure should take into account the regulations. following: The government of West Bengal has already The strategy needs to follow a comprehensive started investing in adaptation. Among the sug- approach to planning that recognizes drainage gested adaptation measures, a number of projects system complexity and interconnectivity of its el- are either currently under way or are planned ements such as storm water drainage, water sup- for future implementation in KMA under the ply, wastewater, water pollution control, water Jawaharlal Nehru National Urban Renewal Mis- reuse, soil erosion, and solid waste management. sion (JNNURM) and the KEIP scheme funded by vulnerabilityofKolkataMetropolitanAreatoincreasedprecipitationinaChangingClimate | 89 ADB. The selection and prioritization of projects climate change in such cost benefit analysis may for adaptation should be based on cost benefit render many projects--which previously did not analysis using the net present value (NPV) ap- show an adequate return on investment--eco- proach. Factoring in the additional impact due to nomically viable. 90 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport Scenarios Applied in the Annex Hydrodynamic Modeling in B the HCMC study INPUTS Seasonal Extreme Storm Upstream SCENARIOS Rainfall Rainfall Surge Mean Sea Level Inflow Landuse PURPOSES I.BL.0 Baseline_Max No No Baseline Baseline Baseline Extent and duration of the "regular" sea- sonal monsoon floods with current land use I.B2.0 B2_Max No No B2 2050+24cm Baseline Baseline I.A2.0 A2_Max No No A2 2050+26cm Baseline Baseline I.BL.1 Baseline_Max No No No Baseline 2050 Extent and duration of the "regular" sea- I.B2.1 B2_Max No No B2 2050+24cm Baseline 2050 sonal monsoon floods with planned future land use and flood control system I.A2.1 A2_Max No No A2 2050+26cm Baseline 2050 I.BL.0_1 Baseline_Max Baseline_30y No No 100y 2050 Extent of the floods caused by 30y storm I.B2.0_1 B2_Max B2_30y No B2 2050+24cm 100y 2050 rainfall (but without storm surge) and 100y extreme upstream inflow; with planned I.A2.0_1 A2_Max A2_30y No A2 2050+26cm 100y 2050 future land use and flood control system II.BL.0 Baseline_Max Baseline_30y Linda No 100y Baseline Extent of the "extreme" floods driven by II.B2.0 B2_Max B2_30y Linda B2 2050+24cm 100y Baseline combination of tropical storms (typhoons) surge, mean sea level rise, 30y rainfall and II.A2.0 A2_Max A2_30y Linda A2 2050+26cm 100y Baseline 100y upstream inflow; with current land use II.BL.1 Baseline_Max Baseline_30y Linda No 100y 2050 Extent of the "extreme" floods driven by II.B2.1 B2_Max B2_30y Linda B2 2050+24cm 100y 2050 combination of storms surge, mean sea level rise, 30y rainfall and 100y upstream II.A2.1 A2_Max A2_30y Linda A2 2050 +26cm 100y 2050 inflow; with planned future land use and flood control system II.BL.0_a Baseline_Max No Linda No Baseline Baseline Extent of floods and saltwater intrusion II.B2.0_a B2_Max No Linda B2 2050 +24cm Baseline Baseline caused by storm surge and sea level rise only in current land use II.A2.0_a A2_Max No Linda A2 2050 +26cm Baseline Baseline II.BL.0_b Baseline_Max Baseline_30y No No 100y Baseline Extent of the floods caused by 100 y II.B2.0_b B2_Max B2_30y No B2 2050 +24cm 100y Baseline extreme upstream inflow and 30y storm rainfall (without storm surge) in current II.A2.0_b A2_Max A2_30y No A2 2050 +26cm 100y Baseline land use III.bl.0 Baseline_Min No No No Baseline Baseline Extent of combination effect of drought and III.B2.0 B2_Min No No B2 2050 +24cm Baseline Baseline sea level rise on water quality especially risk of salinization under current land use III.A2.0 A2_Min No No A2 2050 +26cm Baseline Baseline I.BL.1_1 Baseline_Max Baseline_30y No No 100y 2050 Extent of floods driven by 30y rainfall, 100y I.B2.1_1 B2_Max B2_30y No B2 2050 +24cm 100y 2050 upstream inflow and mean sea level rise; with planned future land use and flood I.A2.1_1 A2_Max A2_30y No A2 2050 +26cm 100y 2050 control system Source: ADB (2010). 91 Annex Adaptation to Increased C Flooding: Brief Overview I n response to the hazard of floods increasing The design discharge for the conveyance systems from climate change, the Asian coastal cities in are normally based on storms with a return period the study have a range of possible adaptation between 1-in-5-years to 1-in-10-years, as compared approaches drawn from existing flood management to design standards of 1-in-30 to 1-in-100-year re- techniques. Each adaptation approach attempts to turn periods commonly used for flood protection reduce risk by either reducing exposure or vulner- embankments. In an urban setting, roads, build- ability, or by increasing capacity. The adaptation ings, and paved surfaces offer little opportunity for measures can be broadly grouped as:59 infiltration of storm rainfall, which means that most of the precipitation appears as storm runoff. In this Protect regard, the predicted increase in storm precipitation Accommodate (ranging from 9 to 15 percent depending on the city Retreat and the scenario) will generate a similar increase in Improved management the design discharge for the storm drainage systems. To provide the same level of protection against protection--reducingexposurewith storm-induced urban flooding, the cities will need structuralinterventions to adapt by either increasing the discharge capac- ity of their storm drainage systems or identifying Flood protection measures seek to minimize the ways to delay or reduce precipitation runoff. For exposure of urban areas to inundation from floods. example, adaptation measures might include storm Common engineering interventions include flood attenuate ponds (using inner city parks) or making embankments, polders, and sea walls, frequently streets, alleys, and other paved areas more porous combined with pumped drainage. To allow the to rainfall to facilitate infiltration. transfer of a flood wave past a city without damage, Flood protection measures always carry a re- other protection techniques include (a) increasing sidual risk of failure, which can be catastrophic as in the hydraulic efficiency60 of the flood channel with the case of New Orleans. An important outcome of dredging, widening, and removal of obstructions; the study is that even for cities with well-designed (b) diverting the flood flows around the city through flood protection plans (e.g., Bangkok and Manila), diversion channels; and (c) attenuating61 the flood the flood protection measures as currently planned flows upstream with reservoirs or through the will not offer the expected level of protection. managed flooding of the agricultural and wetland areas. With the exception of planned flooding of agricultural and wetlands to attenuate flood peaks, 59 Drawn from IPCC, AR4, Chapter 6, Coastal systems and the measures outlined above would be classified low-lying areas, Figure 6.11 Evolution of planned coastal primarily as structural adaptation measures. adaptation measures, pg. 342, 2007. 60 Improving the hydraulic efficiency lowers the flood profile. Within a city, storm drainage systems are de- 61 Attenuation upstream lowers the flood profile down- signed to remove storm runoff without flooding. stream. 93 ACCoMModATe--ATRAdITIon ReTReAT--whenRISKSARe AppRoAChInMonSoonAlASIA ToohIgh,ReTReATReduCeS ToReduCIngvulneRAbIlITy expoSuRe Accommodation acknowledges that people and Retreat as planning option to reduce exposure can be assets are exposed to floods, but seeks to minimize applied in urban areas through urban land use plans vulnerability. In rural Asia, houses located in flood and zoning codes. The social, environmental, and eco- plains and exposed to frequent flooding are usually nomic implications need to be carefully studied. In ur- built on stilts, which reduces the vulnerability of the ban areas with high property values, it is usually only people and the household assets to flooding. Crop- applicable where the risks of loss of life or assets are ping calendars designed around annual monsoon high enough to warrant the action. In rural areas, re- floods and floating varieties of rice are other ways treat is common practice vis-à-vis flood embankments in which rural Asia accommodates floods. In cities as rivers modify courses and erode embankments. where ankle-deep flooding occurs regularly, wooden Forced retreats are occurring in Bangkok's coastal area planks on blocks are common sights to allow passage due to coastal erosion and land subsidence. and access of pedestrians. In periurban areas, landfill Used in conjunction with restoring natural is frequently used to elevate new construction sites ecosystems that provide flood protection benefits, along major roads above the expected flood levels. retreat can be a useful adaptation planning tool. In Accommodation, however, can be applied on coastal areas, land that becomes untenable due to a broader scale. Flood losses in urban areas arise SLR could be converted to natural systems (e.g., primarily from damage to buildings, houses, and mangroves) that can help capture sediment and commercial/industrial properties. In this regard, protect against SLR and storm surges. In urban risks could be dramatically reduced if building areas, reclaiming land along rivers and converting codes specified minimum elevations for the first these to linear parks that will inundate during floods occupied level for housing, commerce, or industry (increasing hydraulic capacity and lowering the (based on predicted inundation levels for the zones flood profile) help beautify the city while provid- in the city). These elevations could be reviewed and ing a range of benefits. Similarly, agricultural lands increased as warranted with long-term (5 to 10 year) that may become unviable due to flooding could advance projections to developers. Given the time be purchased and converted to wetlands. Retreat frame of the study (40 years), this gradual approach sometimes occurs following major disasters when gives urban planners a powerful tool to reduce risks the perceived risks outweigh the desire to recon- with minimal economic impact. struct among the private sector. Planners should Accommodation also relates to measures to seize such opportunities and develop the lands for "flood-proof" structures and building exteriors that their ecological and flood protection potential. may be exposed to floods. Flood-proofing dramati- In addition to the retreat from section of a city cally reduces damages from floods. These standards or coast line, facilities that are critical in disaster should be included in urban building codes as part situations or at particular risk (like schools) can be of an adaptation program. The study has demon- retired and moved as part of a longer term adapta- strated that traditional accommodation methods tion program. may not reduce vulnerability in 2050. In HCMC, for example, areas with frequent ankle-deep flooding will be facing knee-to-waist-deep floods, which are IMpRovedMAnAgeMenT-- not easily managed by pedestrians. This will reduce buIldIngCApACITyToReduCe access and increase losses. Similarly, many houses RISKS and buildings that have been built with main floors above the current 1-in-30-year flood levels will be Flooding needs to be considered in the context of inundated by 1-in-30-year floods in 2050. overall water basin management and the capacity 94 | ClimateRisksandAdaptationinAsianCoastalMegacities:ASynthesisReport of people and institutions to manage the resource. With the hazard of flooding increasing, adap- Experience has shown that piecemeal or ad hoc ap- tation measures aimed at improved management proaches are at risk of causing significant adverse can focus on (a) reassessing operational rule curves impacts elsewhere. For example, flood embank- for reservoirs, given the increasing precipitation ments protect one area by aggravating the flood in variability; (b) increasing awareness and commu- another area. Water uses and hazards need to be nity-based adaptation initiatives (e.g., shelters and addressed holistically for the water basin to maxi- early warning systems); and (c) strengthening the mize benefits. planning and risk-sharing mechanisms related to controlled flooding of agriculture lands to attenuate flood peaks when protecting cities. AdaptationtoIncreasedFlooding:briefoverview | 95 Annex Comparison of Costs D across Cities A n interesting question to ask is how climate ComparingdamageCostsinbangkok change costs compare across the different cit- andManila ies. The figure below provides a comparison 80% 72% between the Manila and Bangkok case studies in 70% 60% the context of a 1-in-30-year flood. HCMC is not 49% 50% included in this analysis because the methods for 40% assessing damage costs were very different. The 30% 20% 15% figure below shows that in Bangkok, for instance, 9% 10% 3% 5% 6% the damages from a 1-in-30-year flood would 2% 0% increase by almost 50 percent going from current Percent increase Costs of 1/30 Costs of 1/30 Additional costs of in costs of 1/30 ood as a % of ood as a % of climate change as climate to an A1FI climate change scenario, while a ood due to GRDP (without GRDP (with A1FI a % of GRDP similar flood in Manila will have a larger effect and climate change climate change) climate change) increase costs by some 79 percent. However, while Manila 2008 Bangkok 2008 the Bangkok damage costs account for land subsid- ence, the Manila costs do not. Thus, accounting for land subsidence may make the increase in costs in the two cities rather similar. whether they are in the same order of magnitude, The costs of climate change are accurately we annualize these costs. We find that annual measured as the costs in similar scenarios with and equivalent of the damage costs to HCMC falls without climate change. Thus for Manila, the costs between $0.67 and $5 billion,62 which on the lower of climate change are represented by the difference end is comparable to the costs borne by Bangkok between the 30-A1FI-EX and 30-SQ-EX scenarios. and Manila. The costs for Manila from a 1-in-30-year flood are $0.67 billion, which amounts to 6 percent of its 2008 GRDP. Similarly, the costs for Bangkok--the difference between 2050-LS-SR-SS-A1FI and 2050- LS scenarios--amount to $1.47 billion, or 2 percent of Bangkok's 2008 GRDP. Given that Bangkok is a much wealthier city than Manila, these numbers are credible. How does HCMC stand in relation to the two 62 These numbers are the equal annual equivalents (EAE) cities? As previously noted, the costs to HCMC, (using a 10 percent discount rate and time period of 42 because they are in present value terms and be- years) of the present value of total costs from regular and cause of the different methodology used, are not extreme flooding. The range reflects the minimum and maximum present values of the two land and GDP-based comparable to the estimates from Bangkok and estimations of climate change costs. EAE = NPV [(i(1 + i) Manila. However, to get a rough understanding of n] / [(1 + i)n - 1)] where i=interest rate and n=time period. 97 THE WORLD BANK 1818 H Street, NW Washington, DC 20344 Telephone: 202.473.1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org Contacts: Poonam Pillai, ppillai@worldbank.org Jan Bojo, jbojo@worldbank.org Maria Sarraf and Susmita Dasgupta, msarraf@worldbank.org, sdasgupta@worldbank.org Asian Development Bank Contact: Jay Roop, jroop@adb.org Japan International Cooperation Agency Contact: Megumi Muto, Muto.Megumi@jica.go.jp