Climate risks and adaptation in Asian coastal megacities: A synthesis

© 2010 The International Bank for Reconstructionand Development / THE WORLD BANK1818 H Street, NWWashington, DC 20433, U.S.A.Telephone: 202-473-1000Internet: www.worldbank.org/climatechangeE-mail: feedback@worldbank.orgAll rights reserved.September 2010This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, andconclusions 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 shownon 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 oracceptance of such boundaries.R I G H T S A N D P E R M I S S I O N SThe material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicablelaw. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grantpermission 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/CorbisSmall images: top: Manila, © Francis R. Malasig/Corbis; middle: Bangkok, © I. Saxar/Shutterstock Images, LLC; bottom: Kolkata, © Bruce Burkhardt/CorbisAll dollars are U.S. dollars unless otherwise indicated.

Table of ContentsAcknowledgments...................................................................................................................................viiAbbreviations and Acronyms................................................................................................................ ixExecutive Summary.................................................................................................................................. xi1. Introduction.........................................................................................................................................1Background and Rationale............................................................................................................................. 1Objective........................................................................................................................................................... 2Process of Preparation.................................................................................................................................... 3Overview of Methodology/Approach and Climate Parameters Selected............................................. 3Structure of the Report................................................................................................................................... 42. Methodologies for Downscaling, Hydrological Mapping, and .Assessing Damage Costs ..................................................................................................................5Selection of Emissions Scenarios, Downscaling, and Uncertainties ....................................................... 5Hydrological Modeling for Developing Scenarios of Flood Risk ........................................................... 9Approach to Assessing Damage Costs....................................................................................................... 12Assessment of Damage Costs in the HCMC Study ................................................................................ 17Assumptions about the Future of Cities in Estimating Damage Costs ................................................ 19Conclusion: Methodological Limitations and Uncertainties in Interpreting Resultsof the Study ............................................................................................................................................. 203. Estimating Flood Impacts and Vulnerabilities in Coastal Cities.............................................23Estimating Future Climate-related Impacts in Bangkok......................................................................... 23Main Findings from Hydrological Analysis and GIS Mapping for Bangkok ..................................... 28Estimating Climate-related Impacts in Manila ........................................................................................ 31Findings from the Hydrological Analysis and GIS Mapping for Metro Manila.................................. 35Estimating Climate-related Impacts in Ho Chi Minh City, Vietnam..................................................... 38Main Findings from Hydrological Analysis and GIS Mapping for HCMC......................................... 44Conclusion...................................................................................................................................................... 504. Assessing Damage Costs and Prioritizing Adaptation Options..............................................51Bangkok: Analysis of Damage Costs Related to Flooding in 2008 and 2050........................................ 51Prioritization of Adaptation Options in Bangkok.................................................................................... 56Analysis of Damage Costs Related to Flooding in Metro Manila ......................................................... 60Prioritization of Adaptation Options in Manila....................................................................................... 65Analysis of Damage Costs in HCMC......................................................................................................... 69Analysis of Adaptation in HCMC.............................................................................................................. 72Conclusion...................................................................................................................................................... 73iii

Figure 4.6 Damage Costs Associated with Different Scenarios (PHP) ...................................................... 63Figure 4.7 Damages to Buildings from a 1-in-30-year Flood (2008 PHP)................................................... 63Figure 4.8 Flood Costs as a Percent of 2008 GDP........................................................................................... 65Figure 4.9 Annual Benefits from Adaptation Investments in Metro Manila............................................. 68BoxesBox 2.1 Strengths and Limitations of Different Downscaling Techniques Selected for this Study.......... 7Box 2.2 Downscaling from 16 GCMs................................................................................................................ 8Box 2.3 Some Basic Principles for Hydrological Mapping...........................................................................11Box 3.1 The Bangkok Metropolitan Region (BMR): Some Assumptions about the Future.................... 26Box 3.2 What does Metro Manila Look Like in the Future? ....................................................................... 35Box 3.3 HCMC in 2050 ..................................................................................................................................... 41Box 3.4 Overview of Downscaling and Hydrological Analysis Carried out for HCMC Study............. 42Box 4.1 Examining Building Damages, Income Losses, and Health Costs in Bangkok ......................... 55Box 4.2 Expected Annual Benefits from Adaptation in Bangkok............................................................... 58Box 4.3 Increased Time Costs and Health Risk from Flooding in Manila................................................. 65Box 4.4 Rough Estimate of Viability of Proposed Flood Control Measures.............................................. 72TablesTable 2.1 Climate Change Forecasts for 2050...................................................................................................... 8Table 2.2 Summary of City Case Study Hydrologic Modeling.......................................................................11Table 2.3 Direct and Indirect Costs from Flooding.......................................................................................... 13Table 2.4 Flood Damage Rate by Type of Building in Manila........................................................................ 15Table 3.1 Poverty Line and the Poor in the BMR1........................................................................................... 24Table 3.2 Bangkok Monthly Average Temperature and Precipitation.......................................................... 25Table 3.3 Climate Change and Land Subsidence Parameter Summary for Bangkok................................. 27Table 3.4 Bangkok Inundated Area under Current Conditions and Future Scenarios.............................. 28Table 3.5 Exposure of Bangkok Population to Flooding................................................................................. 30Table 3.6 Manila: Monthly Average Temperature and Precipitation............................................................ 33Table 3.7 Manila Climate Change Parameters.................................................................................................. 35Table 3.8 Manila: Comparison of Inundated Area (km 2 ) with 1-in-100-year flood for2008 and 2050 Climate Change Scenarios with only Existing Infrastructure andwith Completion of 1990 Master Plan............................................................................................... 36Table 3.9 Affected Length of Road by Inundation Depth............................................................................... 38Table 3.10 HCMC District Poverty Rates. 2003.................................................................................................. 39Table 3.11 Ho Chi Minh City: Monthly Average Temperature and Precipitation......................................... 40Table 3.12 Climate Change Parameter Summary for HCMC ......................................................................... 44Table 3.13 Summary of Flooding at Present and in 2050 with Climate Change........................................... 44Table 3.14 District Population Affected by an Extreme Event in 2050............................................................ 46Table 3.15 Districts Affected by Flooding in Base Year and in 2050................................................................ 47Table 3.16 Effects of Flooding on Future Land Use under 2050 A2 Extreme Event ..................................... 49Table 4.1 Summary of Damages Assessed in the Bangkok Study................................................................. 51Table 4.2 Summary of Flood and Storm Damages, Bangkok (million 2008 THB)...................................... 53Table 4.3 Changes in Income Losses to Wage Earners, Commerce, and Industry...................................... 55Table 4.4 Damage Costs in Bangkok and Regional GRDP............................................................................. 56Table 4.5 Investment Costs for Adaptation Projects in Bangkok (million THB)......................................... 57Table 4.6 Flood Damage Costs With and Without a 30-year Return Period Flood ProtectionProject (million THB)........................................................................................................................... 59Table of Contents| v

Table 4.7 Net Present Value of Adaptation Measures to Provide Protection Againsta 1-in-30 and 1-in-10-year Flood (million THB)............................................................................... 59Table 4.8 Flood Damage Costs in Manila (2008 PHP) .................................................................................... 61Table 4.9 Income and Revenue Losses to Individuals and Firms Associated with Floods(2008 PHP)............................................................................................................................................. 64Table 4.10 Damage Costs from 1-in-10, 1-in-30, and 1-in-100-year Floods in Different Scenarios(2008 PHP)............................................................................................................................................. 65Table 4.11 Adaptation Investments Considered for Different Return Periods andClimate Scenarios................................................................................................................................. 67Table 4.12 Investment Costs and Net Present Value of Benefits Associated with DifferentFlood Control Projects in Manila (PHP) using a 15 percent discount rate.................................. 67Table 4.13 Expected Cost of Flooding based on Quadratic Relationship between Durationof Flooding and Land Values in HCMC........................................................................................... 70Table 4.14 Present Value of the Cost of Floods up to 2050 using the GDP Estimation Method.................. 71Table 4.15 Summary of Present Value of Climate Change Costs in HCMC (USD)...................................... 72Table 4.16 Proposed Implementation Arrangements for HCMC.................................................................... 74vi | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

AcknowledgmentsThis synthesis report is a product of a joint programon Climate Adaptation in Asian CoastalMegacities undertaken by the World Bank incollaboration with the Asian Development Bankand the Japan International Cooperation Agency. Itis based on extensive collaboration among the threeagencies, who jointly agreed to undertake severalcity-level studies and prepare a synthesis report.The core team preparing this synthesis reportconsisted of Poonam Pillai (Sr. Environmental Specialistand task team leader), Bradford Ryan Philips(Sr. Civil Engineer, consultant), Priya Shyamsundar(Sr. Environmental Economist, consultant), KaziAhmed (consultant) and Limin Wang (Sr. EnvironmentalEconomist, consultant) and includedextensive collaboration with the different city-levelteams. In particular, we would like to thank JanBojo (World Bank), who led the Bangkok study;Megumi Muto (JICA), who led the Manila studyand was the main focal point from JICA; Jay Roop(ADB), who led the Ho Chi Minh City study and wasthe main focal point from ADB, and Maria Sarrafand Susmita Dasgupta (World Bank), who led theKolkata study. For the Bangkok report, we are alsograteful to the team at Panya consultants; to BangkokMetropolitan Administration professionals;and to Manuel Cocco, Pongtip Puvacharoen, andYabei Zhang. For the HCMC study, we thank theconsulting team at the International Centre for EnvironmentalManagement, including Jeremy Carew-Reid, Anond Snidvongs, Peter-John Meynell, JohnEdmund Sawdon, Nigel Peter Hayball, Tran Thi Ut,Tranh Thanh Cong, Nguyen Thi Nga, Nguyen LeNinh, Nguyen Huu Nhan, and Nguyen Dinh Tho;the Ho Chi Minh City People’s Committee; and theDepartment of Natural Resources and Environment(DoNRE). For the Manila study, we are grateful tothe Metro Manila Development Authority, NationalStatistics Office, Professor Emma Porio and staffat Ateneo de Manila University, CTI EngineeringInternational, and ALMEC Corporation, as wellas the local government officials and communityleaders who provided valuable inputs to the report.We also gratefully acknowledge the contributionsof Professor Akimasa Sumi (University of Tokyo),Professor Nobuo Mimura (Ibaraki Universty),and Dr. Masahiro Sugiyama (Central ResearchInstitute of Electric Power Industry) for providingthe analytic framework for downscaling the IPCCclimate models. We thank the Kolkata team andin particular Subhendu Roy and the INRM teamfor their inputs and collaboration, and to AdrianaDamianova for initially leading the Kolkata study.Suggestions from Ian Noble also helped strengthenthe analysis. We are especially grateful to DanielHoornweg, Anthony Bigio, and Tapas Paul for peerreviewing this report.A special thanks to James Warren Evans (Director,Environment Department, World Bank), MagdaLovei, (Sector Manager, EASER, World Bank), NeerajPrasad (Lead Carbon Finance Specialist, ENVCF),Megumi Muto (Research Fellow, JICA) and Jay Roop(Environmental Specialist, ADB) for initiating thiscollaborative activity and to Kseniya Lvovsky (ProgramManager, Climate Change team, EnvironmentDepartment), Michele De Nevers (Senior Manager,Environment Department, World Bank), GajanandPathmanathan (Manager, SASDO and Acting SectorManager, SASDI), Nessim Ahmad (Director, Environmentand Safeguards, Asian Development Bank), DrKeiichi Tsunekawa (Director, JICA Research Institute)and Mr. Hiroto Arakawa (Senior Special Advisor,JICA) under whose general guidance this reportwas prepared. Thanks to Perpetual Boetang for hervii

Abbreviations and AcronymsADB Asian Development BankAOGCM Atmosphere-ocean general circulationmodelsBMR Bangkok Metropolitan RegionBAU Business as usualCCA Climate change adaptationCoP Conferences of PartiesDIVA Dynamic interactive vulnerabilityassessmentDRR Disaster risk reductionECLAC Economic Commission for LatinAmerica and the CaribbeanGCMs Global climate modelsGDP Gross domestic productGEF Global Environment FacilityGHG Greenhouse gasGPCP Global Precipitation ClimatologyProjectGRDP Gross regional domestic productHCMC Ho Chi Minh City1DD One-degree dailyIPCC Intergovernmental Panel for ClimateChangeIRS3 Integrated research system forsustainability scienceIZ Industrial zonesJICA Japan International CooperationAgencyLGUs Local government unitsMONRE Ministry of Natural Resources andEnvironmentMP Master planNESDB National Economic and SocialDevelopment BoardPC People’s CommitteePCMDI Program for Climate Model Diagnosisand IntercomparisonSRES Special Report on Emissions ScenariosUNDP United Nations Development ProgramUNFCCC United Nations FrameworkConvention on Climate ChangeVOC Vehicle operations costWCRP World Climate Research ProgramWGCM Working Group on Coupled ModelingNote: Unless otherwise noted, all dollars are U.S. dollars.ix

Executive SummaryIntroduction and RationaleCoastal areas in both developing and more industrializedeconomies face a range of risks related to climatechange and variability (IPCC 2007a). Potentialrisks include accelerated sea level rise, increase in seasurface temperatures, intensification of tropical andextra tropical cyclones, extreme waves and stormsurges, altered precipitation and runoff, and oceanacidification (Nicholls et al. 2007). The IntergovernmentalPanel for Climate Change Fourth AssessmentReport (IPCC 2007a) points to a range of outcomesunder different scenarios. It identifies a number ofhotspots—including heavily urbanized areas situatedin the low-lying deltas of Asia and Africa—asespecially vulnerable to climate-related impacts.The number of major cities located near coastlines,rivers, and deltas provides an indication of thepopulation and assets at risk. Thirteen of the world’s20 largest cities are located on the coast, and morethan a third of the world’s people live within 100miles of a shoreline. Low-lying coastal areas represent2 percent of the world’s land area, but contain13 percent of the urban population (McGranahan etal. 2007). A recent study of 136 port cities showedthat much of the increase in exposure of populationand assets to coastal flooding is likely to be in citiesin developing countries, especially in East and SouthAsia (Nicholls et al. 2008).In terms of population exposed to coastal flooding,for example, in 2005 five of the ten most populouscities included Mumbai, Guangzhou, Shanghai,Ho Chi Minh City, and Kolkata (formerly Calcutta).By 2070, nine of the top ten cities in terms of populationexposure are expected to be in Asian developingcountries (Nicholls et al. 2008). The vulnerability ofthe East Asia region is also highlighted by the globalstudy on the economics of adaptation to climatechange, which estimates that the cost of adaptationto climate change is likely to be the highest in thisregion (World Bank 2010). In flood-prone cities suchas Ho Chi Minh City, Kolkata, Dhaka, and Manila,potential sea level rise and increased frequency andintensity of extreme weather events poses enormousadaptation challenges. The urban poor—often livingin riskier urban environments such as floodplains orunstable slopes, working in the informal economy,and with fewer assets—are most at risk from exposureto hazards (Satterthwaite et al. 2007).Despite its importance, few developing countrycities have attempted to address climate change systematicallyas part of their decision-making process.Given the risks faced by coastal cities and the importanceof cities more broadly as drivers of regionaleconomic growth, adaptation must become a coreelement of long-term urban planning. The Mayor’sSummit in Copenhagen in December 2009—andfollow-on efforts to institutionalize a Mayor’s TaskForce on Urban Poverty and Climate Change—signifymuch-needed attention to this issue.In response to client demand and recognizingthe importance of addressing urban adaptationand major vulnerabilities of Asian coastal cities, theAsian Development Bank (ADB), the Japan InternationalCooperation Agency (JICA), and the WorldBank agreed to undertake an analysis in severalcoastal megacities to address climate adaptationand prepare a synthesis report based on the citylevelfindings. The selected cities included Manila(led by JICA), Ho Chi Minh City (led by ADB), andBangkok (led by the World Bank). 11Kolkata is also one of the selected cities but is not includedin the synthesis report as it was ongoing at the time of the preparationof this report. A brief overview is included in Annex A.xi

Why these three cities? The three developingcountry cities selected for this study are all coastalmegacities with populations (official and unofficial)ranging from 8 to 15 million people. Two are capitalcities and all three are centers of national and regionaleconomic growth contributing substantially to theGDP of the respective countries. However, being lowlyingcoastal cities situated in the deltas of major riversystems in the East Asia region, all three are highlyvulnerable to climate-related risks and rank high inrecent rankings of exposure and vulnerability. HoChi Minh City and Bangkok are among the top 10cities in terms of population likely to be exposed tocoastal flooding due to climate-related risks in 2070,according to the first global assessment of port cities(Nicholls et al. 2008). Further, Manila has been identifiedas particularly vulnerable to typhoon damage,and HCMC ranks fifth by population exposed tothe effects of climate change (Nicholls et al. 2008). Arecent study also identifies Manila, Ho Chi Minh City,and Bangkok among the top eleven Asian megacitiesthat are most vulnerable to climate change (Yusufand Francisco 2009). 2 Devastating floods in Manilain 2009 only confirm the vulnerability of this city toextreme weather events. For instance, flooding inManila from tropical storm Ketsana in Septemberwas the heaviest in almost 40 years, with flood watersreaching nearly 7 meters. More than 80 percent ofthe city was underwater, causing immense damageto housing and infrastructure and displacing around280,000–300,000 people. 3 All of this highlights theneed to better understand and prepare for such climaterisks and incorporate appropriate adaptationmeasures into urban planning.While there is a growing literature on cities andclimate change, as yet there is limited research onsystematically assessing climate-related risks at thecity level. This report aims to fill this gap. Further,it aims to provide evidence-based information tosupport urban policy and planning as these issuesare debated at the local, national, and global levels.ObjectiveThe main objective of this report is to strengthenour understanding of climate-related risks and impactsin coastal megacities in developing countriesusing case studies of three cities that are differentin their climate, hydrological, and socioeconomiccharacteristics. Specifically, it draws on an in-depthanalysis of climate risks and impacts in Bangkok,Manila, and Ho Chi Minh City to highlight to nationaland municipal decision makers (a) the scaleof climate-related impacts and vulnerabilities atthe city level, (b) estimates of associated damagecosts, and (c) potential adaptation options. Whilethe report focuses on three cities in East Asia, thepolicy implications resulting from the comparativeanalysis of these cities has broader relevance forassessing climate risks and identifying adaptationoptions in other coastal areas.Approach and MethodologyThe approach to assessing climate risks and impactsconsists of the following sequential steps: (1)determining climate variables at the level of thecity/watershed through downscaling techniques;(2) estimating impacts and vulnerability throughhydrometeorological modeling, scenario analysis,and GIS mapping; and (3) preparing a damage/loss assessment and identification/prioritizationof adaptation options.As a first step, each of the city-level studiesconsidered two IPCC scenarios, a high- and a lowemissionsscenario, 4 and estimated climate risksto 2050. The 2050 time horizon for the study is appropriategiven city-level planning horizons andthe typical time frame for major flood protectionmeasures. The downscaling analysis allowed estimationof changes in temperature and precipitationin 2050. These parameters were used as inputs to thehydrological modeling. In addition to this, assumptionsand estimates were also made about changesin sea level rise and storm surge in 2050 based onpast historical data and available estimates.2Vulnerability in the scorecard was understood interms of exposure, sensitivity, and adaptive capacity ofthe cities. See also http://www.idrc.ca/uploads/user-S/12324196651Mapping_Report.pdf.3http://edition.cnn.com/2009/WORLD/asiapcf/09/27/philippines.floods/index.html.4Different scenarios were considered to assess impact dueto the uncertainties in projecting future climate conditions.xii | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

For each city, complex hydrometeorologicalmodels were then developed using a whole hostof local information. These included (a) climatevariables such as changes in temperature, precipitation,sea level rise, and storm surge; (b) socioeconomicand developmental factors such as landsubsidence, land use, and population increases;and (c) local topographical and hydrologicalinformation. Flooding in the metropolitan areaswas chosen as the key variable to assess impact.The hydrological analysis allowed determinationof the area, depth, and duration of flooding underdifferent scenarios. This information was used toidentify the scale of risks and vulnerability of sectors,local populations, and districts (representedin GIS maps), as well as estimate damage costs.Two of the three studies undertook cost-benefitanalysis to prioritize adaptation options, whilethe third approached the issue of adaptation morequalitatively. To understand the impact of climatechange in 2050 in each city, an important assumptionmade by all teams was that without climatechange, the climate in 2050 would be similar to the2008/ base-year climate. Various climate scenariosare overlaid on this assumption.Process of PreparationThe analysis was carried out over a period of oneand-a-halfyears. The synthesis team and the citylevelteams met periodically and worked closely todevelop common terms of reference to guide thecity-level studies, as well as share methodologicalissues and ongoing findings. These discussions andthe analysis undertaken for each city have formedthe basis of this report. Further, each city-levelteam worked with their respective country/urbancounterparts to build ownership and capacity forthe analysis. For instance, the main counterparts inHCMC were the HCMC People’s Committee andthe Department of Natural Resources and Environment(DoNRE). The study sought to inform thepreparation of HCMC’s citywide adaptation plan.In Bangkok, the main counterpart was the BangkokMunicipal Authority. In Metro Manila, the maincounterpart was Metro Manila Development Authority(MMDA). At the global level, a preliminaryversion of the study has been presented at severalinternational forums.Uncertainties, Limitations,and Interpreting theFindings of this StudyAny study forecasting conditions four decadeshence will be faced with large uncertainties andthese need to be borne in mind in interpreting theresults of this study. One uncertainty concernsthe pathway of GHG emissions. To address thatissue, the city case studies examined both a highand a low GHG emissions scenario to bracket thelikely future conditions. In the climate changedownscaling methodologies, there are uncertaintiesin forecasting the increase in extreme andseasonal precipitation under the different scenarios.The techniques applied in the statisticaldownscaling examined the results from sixteenatmosphere-ocean general circulation models(AOGCM). Robust relationships were identifiedfor temperature (with a ~ 10 percent internalerror) and precipitable water increases (with a ~10–20 percent error) (Sugiyama 2008). Hydrologicmodels can simulate flood events with relativelysmall errors (

Key FindingsFrequency of extreme events likely to increaseAll three cities are likely to witness increases intemperature and precipitation linked with climatechange and variability. In Bangkok, temperatureincreases of 1.9 ° C and 1.2 ° C for the high and lowemissions scenarios respectively are estimated for2050 and are linked with a 3 percent and 2 percentincrease in mean seasonal precipitation respectively.In Manila, the mean seasonal precipitation is expectedto increase by 4 percent and 2.6 percent forthe high and low emissions scenarios. In HCMC,future projections suggest greater seasonal variabilityin rainfall and increasing frequency of extremerainfall related to storms.Increase in flood-prone area due to climatechange in all three citiesIn all three megacities, in 2050, there is an increase inthe area likely to be flooded under different climatescenarios compared to a situation without climatechange. In Bangkok, for instance, under the conditionsthat currently generate a 1-in-30-year flood,but with the added precipitation projected for a highemissions scenario, there will be approximately a 30percent increase in the flood-prone area. In Manila,even if current flood infrastructure plans are implemented,the area flooded in 2050 will increase by 42percent in the event of a 1-in-100-year flood underthe high emission scenario compared to a situationwithout climate change. In HCMC, for regularevents in 2050, the area inundated increases from 54percent in a situation without climate change to 61percent with climate risks considered under the highemission scenario. For extreme (1-in-30 year) events,in 2050, the area inundated increases from 68 percent(without climate change) to 71 percent (with climaterisks considered) under the high emission scenario.Further, there is a significant increase in both depthand duration for both regular and extreme floodsover current levels in 2050 in HCMC. The analysisalso highlights areas that will be at greater risk offlooding in each metropolitan area. In Metro Manila,for instance, areas of high population density suchas Manila City, Quezon City, Pasig City, MarikinaCity, and San Juan Mandaluyong City are likely toface serious risks of flooding.Increase in population exposed to floodingIn all three cities, there is likely to be an increase inthe number of persons exposed to flooding in 2050under different climate scenarios compared to a situationwithout climate change. For instance, in Bangkokin 2050, the number of persons affected (floodedfor more than 30 days) by a 1-in-30-year event willrise sharply for both the low and high emission scenarios—by47 percent and 75 percent respectively—compared to those affected by floods in a situationwithout climate change. In Manila, for a 1-in-100-yearflood in 2050, under the high emission scenario morethan 2.5 million people are likely to be affected (assumingthat the infrastructure in 2050 is the same asin the base year), and about 1.3 million people if the1990 master plan is implemented. In HCMC, currently,about 26 percent of the population would beaffected by a 1-in-30-year event. However, by 2050,it is estimated that approximately 62 percent of thepopulation will be affected under the high emissionscenario without implementation of the proposedflood control measures. Even with the implementationof these flood control measures, more than half ofthe projected 2050 population is still likely to be at riskfrom flooding during extreme events. How to plan forsuch large percentages of population being exposedto future flooding needs to be seriously considered.Costs of damage likely to be substantial andcan range from 2 to 6 percent of regionalGDPIn Bangkok, the increased costs associated withclimate change (in a high emission scenario) froma 1-in-30-year flood is THB 49 billion ($1.5 billion),or approximately 2 percent of GRDP. These are theadditional costs associated with climate change. Theactual costs of a 1-in-30-year flood—including costsresulting from both climate change and land subsidence—areclose to $4.6 billion in 2050. In Manila, asimilar 1-in-30-year flood can lead to costs of floodingranging from PHP 40 billion ($0.9 billion)—givencurrent flood control infrastructure and climate conditions—toPHP 70 billion ($1.5 billion) with similarxiv | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

infrastructure but a high emission climate scenario.Thus, the additional costs of climate change froma 1-in-30-year flood would be approximately PHP30 billion ($0.65 billion) or 6 percent of GRDP. TheHCMC study adopts a different methodology to analyzecosts and its results cannot directly be comparedto the costs of Manila and Bangkok. The HCMC studyuses a macro approach and estimates a series of annualcosts up to 2050. The flood costs to HCMC, inpresent value terms, range from $6.5 to $50 billion. 5The “annualized” costs of flooding would likely becomparable to the costs of Bangkok and Manila.Damage to buildings is an importantcomponent of flood-related costsDamage to buildings is a dominant component offlood-related costs, at least in Bangkok and Manila.In these cities, over 70 percent of flood-related costsin all scenarios are a result of damages to buildings.Cities are, almost by definition, built-up areas fullof concrete structures, so it is not surprising thatthe main impact of floods is on these structuresand the assets they carry. In HCMC, 61 percent ofurban land use and 67 percent of industrial landuse are expected to be flooded in 2050 in an extremeevent if the proposed flood control measures arenot implemented. Potential flooding in HCMC alsohas major implications for planning in key sectorssuch as transportation and waste management. Forinstance, the city’s existing and planned transportationnetwork, wastewater treatment plants andlandfill sites are likely to be exposed to increasedflooding under the high emission scenario even withthe implementation of the proposed flood protectionsystem, raising important issues for plannerssuch as managing the environmental consequencesof flooding. Thus, as cities develop over the next 40years, it will be important to consider climate risksin designing their commercial, residential, and industrialassets and zones.Impact on the poor and vulnerable will besubstantial, but even better-off communitieswill be affected by floodingIn Bangkok, the study estimates that about 1 millioninhabitants will be affected by flooding undera high emission scenario in 2050. One out of eightof the affected inhabitants will be those living incondensed housing areas where the populationprimarily lives below the poverty line. Of the totalaffected population, approximately one-third mayhave to encounter inundation of more than a halfmeterfor at least one week, marking a two-foldincrease in the vulnerable population. People livingin the Bang Khun Thian district of Bangkok and thePhra Samut Chedi district of Samut Prakarn will beespecially affected. In HCMC, in some of the areas,both the poor and non-poor are at risk. However, ingeneral, poorer areas are more vulnerable to flooding.Thus city planners need to devise strategiesthat focus on the poorer sections of the city throughimproved access to housing, infrastructure anddrainage, devising appropriate land use policiesand improving the level of preparedness amongthe more disadvantaged social groups.Land subsidence is a major problem andcan account for a greater share of thedamage cost from flooding compared toclimate-related factorsOne of the main findings of this study is that nonclimate-relatedfactors such as land subsidence areimportant and in some cases even more importantthan climate risks in contributing to urban flooding.In Bangkok for instance, there is nearly a two-foldincrease in damage costs between 2008 and 2050 dueto land subsidence. Further, almost 70 percent of theincrease in flooding costs in 2050 in the city is dueto land subsidence. While data for land subsidencewere not available for Manila and HCMC and thisissue was not considered in the hydrological modelingfor these two cities, available literature suggeststhat it is an important factor in all three cities andshould be considered in follow-up studies. Eventhough the megacities have already undertaken anumber of measures to slow down land subsidence,further regulatory and market incentives are clearlyrequired to stem groundwater losses. City governmentsneed to better assess factors contributing toland subsidence and consider options to reduce it.5The exchange rates used were the average exchange ratesin 2008: 1 USD = THB 33.31, PHP 44.47 and VND 16,302.25.Executive Summary| xv

RecommendationsCoastal cities in developing countries face enormouschallenges linked with current patternsof population and economic growth, associatedenvironmental externalities, urban expansion andexisting climate variability. Climate change willpose additional risks beyond those currently facingcoastal megacities. As the study shows, these riskswill also be associated with significant costs to localpopulations and infrastructure. Strong political willis thus needed to strengthen the capacity to addressboth existing climate variability and additional risksposed by climate change. Three main lessons standout from the study.Better management of urban environmentand infrastructure will help manage potentialclimate-related impactsAnalysis carried out in the city case studies showthat sound urban environmental management isalso good for climate adaptation. As the Bangkokstudy shows, land subsidence, if not arrested, wouldcontribute a greater share of damage costs fromfloods than a projected change in climate conditions.Thus, addressing land subsidence and factors contributingto it is important from the perspective ofurban adaptation. While the HCMC study has notestimated the damage costs due to other environment-developmentfactors—such as the presence ofsolid waste in the city’s drains and waterways, poordredging of canals, siltation of drains, deforestationin the upper watershed—it provides extensivequalitative evidence to demonstrate the role thesefactors play in contributing to urban flooding.Collectively, the studies highlight the importanceof addressing existing environment-developmentfactors as a critical part of urban adaptation. Theyalso show that given the high risks of continuingto urbanize according to current patterns, muchmore effort should be given to considering theenvironmental implications of urban growth andexpansion in the context of managing current andfuture climate risks.Climate-related risks should be considered asan integral part of city and regional planningWhile improved urban environmental managementis important, the studies also show that given theadditional costs linked with climate change, citiesneed to make a proactive effort to consider climaterelatedrisks as an integral part of urban planningand to do so now. First, city planners need to developstrategic urban adaptation frameworks formanaging climate risks involving a range of toolssuch as policy and regulatory reforms, investments,and capacity building. Such a strategy can providean overarching framework for actions taken withineach sector at the regional, delta, and city levels.Second, much more emphasis needs to be given toimproving the knowledge base regarding climaterisks and related socioeconomic and developmentfactors. Developing and updating scenarios andplanning for a range of potential outcomes will becritical for urban planners. This can be accomplishedby strengthening the collaboration between planningand sector agencies and research institutions,thus giving municipal agencies the tools to makedecisions regarding risk management over the longterm (Rosenzweig et al. 2007) Third, it is importantto strengthen the capacity of local urban governmentalinstitutions to adapt to climate change.Among other things, this involves strengthening thecapacity to prioritize different adaptation options,improving coordination between various urbansector agencies and sector plans, and incorporatingclimate change considerations into the earlieststages of decision making.Targeted, city-specific solutions combininginfrastructure investments, zoning, andecosystem-based strategies are requiredGiven that cities are characterized by distinct climatic,hydrological, and socioeconomic features—but also that the urban poor in general are morevulnerable to increased flooding due to climatechange—targeted, city-specific, and cutting edgeapproaches to urban adaptation are needed. First,xvi | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

these include strategies that focus on the morevulnerable areas of the city and the urban poor.Second, as the studies show, hard infrastructureinterventions can also be usefully combined withecosystem-based solutions. For instance, constructionof dykes can be matched with managementand rehabilitation of mangrove systems, reforestationof upper watersheds, river and canal bankprotection, and implementation of basin-wide flowmanagement strategies. Urban wetlands providea range of services, including flood resilience, allowinggroundwater recharge and infiltration, andproviding a buffer against fluctuations in sea leveland storm surges. Thus, rehabilitation of urbanwetlands is critical. Third, as the city case studiesshow, while a combination of climate-related factorscan contribute to urban flooding, some factors aremuch more important than others in different citiesdepending on location, elevation, and topographyof the city. For instance, in HCMC, storm surgesand sea level rise are important factors contributingto flooding. However, in Bangkok these factors arerelatively less important. The policy implication isthat adaptation measures need to be designed basedon the specific hydrological and climate characteristicsof each city. Fourth, damages to buildingsemerge as a dominant component of flood-relatedcosts, at least in Bangkok and Manila. Vulnerabilitymapping, land use planning and zoning could beused to restrict future development in hazardouslocations, ultimately retiring key infrastructureand vulnerable buildings in these areas. Similarly,building codes aimed at flood-proofing buildings(including the lowest habitable elevation in vulnerableareas) could dramatically reduce damage costs.Such targeted measures could go a long way inhelping coastal megacities to adapt to current andfuture climate risks.Executive Summary| xvii

1IntroductionBackground and RationaleAs recent weather events have illustrated, coastalareas in both developing and more industrializedeconomies face a range of risks related to climatechange (IPCC 2007a). Anticipated risks include anaccelerated rise in sea level of up to 0.6 meters ormore by 2100, a further rise in sea surface temperaturesby up to 3° C, an intensification of tropicaland extra tropical cyclones, larger extreme wavesand storm surges, altered precipitation and runoff,and ocean acidification (Nicholls et al. 2007).The Intergovernmental Panel for Climate ChangeFourth Assessment Report (IPCC 2007a) points toa range of outcomes under different scenarios andidentifies a number of hotspots—including heavilyurbanized areas situated in the large low-lyingdeltas of Asia and Africa—as especially vulnerableto climate-related impacts. For instance, by 2080,the report points out, many millions more peoplemay experience floods annually due to sea levelrise (IPCC 2007a). More frequent flooding and inundationof coastal areas can also result in variousindirect effects, such as water resource constraintsdue to increased salinization of groundwater supplies.Human-induced pressures on coastal regionscan further compound these effects.The location of many of the world’s major cities—suchas Mumbai, Shanghai, Jakarta, Lagos,and Kolkata—around coastlines, rivers, and deltasprovides an indication of the population and assetsat risk. Thirteen of the world’s 20 largest citiesare located on the coast and more than a third ofthe world’s population lives within 100 miles ofa shoreline. Low-lying coastal areas—defined asareas along the coast that are less than 10 metersabove sea level—represent 2 percent of the world’sland area, but contain 13 percent of the urban population(McGranahan et al. 2007). A recent study of136 port cities showed that the population exposedto flooding linked with a 1-in-100-year event islikely to rise dramatically, from 40 million currentlyto 150 million by 2070 (Nicholls et al. 2008).Similarly, the value of assets exposed to floodingis estimated to rise to $35 trillion, up from $3 trilliontoday. The study also shows that significant,increasing exposure is expected for the populationsand economic assets in Asia’s coastal cities.In flood-prone cities such as Manila, potentialsea level rise and increased frequency and intensityof extreme weather events poses enormouschallenges on urban local bodies’ ability to adapt.Apart from their location, the scale of risk is alsoinfluenced by the quality of housing and infrastructure,institutional capacity with respect toemergency services, and the city’s preparednessto respond. The urban poor are most at risk fromexposure to hazards in coastal cities, as they tendto live in riskier urban environments (such asfloodplains, unstable slopes), tend to work in theinformal economy, have fewer assets, and receiverelatively less protection from government institutions(Satterthwaite et al. 2007).Despite its importance, few developing countrycities have initiated efforts to integrate climatechange issues as part of their decision-makingprocess. Given the risks faced by coastal cities andthe importance of cities more broadly as drivers of1

policy implications have broader relevance forassessing climate risks and identifying adaptationoptions in other coastal areas.Process of PreparationThe analysis was carried out over a period of oneand-a-halfyears. The synthesis team and the citylevelteams worked closely to develop commonterms of reference to guide the city-level studies.Further, while the synthesis team helped coordinatethe process, each city-level team worked independentlywith their respective country counterparts tobuild ownership and capacity for the analysis. Thecity teams were comprised of members with a rangeof skills, including climate modeling, hydrologicalanalysis, GIS mapping, economic analysis, andurban planning. The city teams and the synthesisteam preparing this report also met periodically toshare methodological issues and ongoing findingsand research. These discussions and the analysisundertaken for each city have formed the basis ofthe preparation of this synthesis report. At the levelof each city, the teams have undertaken stakeholderconsultations with city officials and governmentagencies at different levels, nongovernmentalorganizations, the private sector, and other constituencies.For instance, the main counterparts inHCMC were the HCMC People’s Committee andthe Ministry of Natural Resources and Environment(MONRE); the study sought to inform preparationof HCMC’s city-wide adaptation plan. In Bangkok,the main counterpart was the Bangkok Municipalauthority. In Metro Manila, it was the Metro ManilaDevelopment Authority (MMDA). At the globallevel, preliminary findings have already been presentedat several international forums to reachurban planners, municipal decision makers, andresearchers.Overview of Methodology/Approach and ClimateParameters SelectedThe city-level studies considered two IPCC emissionsscenarios, 9 A1FI and B1 (with the exception ofHCMC, which considered the A2 and B2 scenarios),and estimated climate risks to 2050. The 2050 timehorizon for the study is appropriate, given planninghorizons in most cities and given that the typicaltime frame for major flood protection planning isabout 30 years. Moreover, the uncertainty in climateprojections expands rapidly past roughly themid-21st century, providing additional justificationfor limiting the time horizon to 2050. Climate variablesconsidered included changes in temperature,changes in precipitation, estimated sea level rise,and estimated storm surge. In addition, non-climatefactors—such as land subsidence, land use changes,salinity intrusion, and population increases—werealso considered. Flooding in the metropolitan areaswas chosen as the key climate variable to beexamined. The approach consisted of the followingsequential steps: (1) downscaling climate variablesto the level of the city/watershed; (2) hydrometeorologicalmodeling and scenario analysis, presentedin GIS maps; and (3) damage/loss assessment andidentification/prioritization of adaptation options.These steps are discussed in more detail in chapter 2.To support this analysis, each city team collectedextensive historical and city-specific data relatedto past climate events such as storms and flooding,socioeconomic data, information about local topographyand hydrology, information on land use, andso forth. Data limitations were a major challenge,but each team worked with existing data from publicsources, as well as data made available by citygovernments and institutions. There are numerousuncertainties at each step of the analysis.While the main focus of this report is on assessingfuture climate risks at the city level, it buildson the recognition of strong links between climateadaptation and ongoing efforts toward disasterrisk management. Despite the institutional differencesin terms of how these efforts have emerged,and differences in how climate change/variabilityand disasters manifest themselves, they both sharecommon ground in striving toward strengtheningadaptive capacity of vulnerable communities, buildingresilience, and reducing the impact of extreme9Different scenarios were considered to assess impact dueto the uncertainties in projecting future climate conditions.Introduction | 3

events. The analysis undertaken in this report usesseveral methodologies that have long been usedin the context of disaster risk management—suchas damage cost assessment and probabilistic riskanalysis—illustrating the opportunities for crossfertilizationin both areas.Structure of the ReportChapter 2 presents methodologies used to determineclimate change risks at the city/river-basinlevel through downscaling techniques, flood riskassessment through hydrometeorological models,and damage cost analysis. Chapter 3 presentsthe main findings from climate downscaling,hydrological modeling analyses, and use of GISmapping. 10 Chapter 4 presents the analysis andfindings relating to damage cost assessment, aswell as an analysis of adaptation options. Finally,Chapter 5 draws broad policy lessons and presentsconclusions.10For a broader set of GIS maps, please refer to city-specificreports.4 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

2Methodologies forDownscaling, HydrologicalMapping, and AssessingDamage CostsIn order to assess the impact of climate changein terms of increased flooding in 2050 in eachof the coastal cities, three main methodologicalsteps were taken. These include (1) determiningclimate-related impacts at the city/river-basin levelthrough downscaling; (2) developing flood riskassessment hydrometeorological models for eachcity to estimate flooding in 2050 under different scenarios;and (3) assessing damage costs. This chapterprovides a summary of these methodologies. Ithighlights the climate change scenarios selected, approachesto downscaling, assumptions underlyinghydrological analysis, and the approach to damagecost assessment. Some of the methodologies usedhere—such as damage cost analysis and probabilisticrisk assessment—are also used in disaster riskmanagement. 11 Uncertainties and errors involvedin different steps of the analysis are also discussed.Selection of EmissionsScenarios, Downscaling,and UncertaintiesTo measure the impact of climate change on thecities in 2050, it was necessary to assume emissionsscenarios and as a first step, “downscale” climatechange forecasts to local levels so that the meteorologicalparameters—such as changes in temperatureand precipitation—could be applied as inputs to thehydrometeorological models.Range of emissions scenarios consideredThe potential impact of climate change can varygreatly depending on the development pathwaythat is assumed. Beginning in 1992, the IPCC hasprovided various scenarios for the emissions ofgreenhouse gases based on assumptions of differentdevelopment pathways—namely, complex anddynamic interactions among future demographicchanges, economic growth, and technological and environmentalchanges. These emissions scenarios areprojections of what the future may look like and area tool to model climate change impacts and relateduncertainties. As described in the IPCC Special Report11A probabilistic risk assessment provides an estimateof the probability of loss due to hazards. It is commonlyused in disaster risk management planning and provides aquantitative baseline for measuring the benefits (or lossesavoided) of disaster management alternatives. In climatechange impact and adaptation studies, it also provides abaseline for assessing the change in risks due to the increasinghydrometeorological hazards associated withclimate change. See, for instance, Earthquake VulnerabilityReduction Program in Colombia, A Probabilistic Cost-benefitAnalysis (World Bank Policy Research Working Paper3939, June 2006) for an example of a probabilistic risk assessmentused in disaster risk management planning. Theprocess involves the development of several interconnectedmodules, which calculate in turn the hazard probability,exposure, vulnerability (or sensitivity to damage),damages, and losses. While the approaches used in the developmentof the modules and the calculation of the lossesvaried, each city case study did, however, follow a similaranalytical process.5

Box 2.1 ■ Strengths and Limitations of Different Downscaling TechniquesSelected for this StudyThe city studies utilized two different downscaling techniques—pattern scaling which is a kind of statistical downscaling, and dynamic downscalingtechniques. 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 statisticalrelationships are then applied to large-scale climate variables from AOGCM simulations to estimate corresponding local/regional characteristics (liketemperature and precipitation). Pattern scaling, one of the techniques used in this work, is a form of statistical downscaling that requires minimalregional meteorological data and is computationally not very complicated. In pattern scaling the change in the local variable of interest—liketemperature or precipitation—is represented in terms of the GCM’s change for the variable per unit change in global temperature (this is the scalingfactor). 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 mainadvantage of statistical downscaling is that it is computationally not very expensive and can provide information at point locations. However, amongseveral limitations, one issue is that the statistical relationships may not remain the same in a future climate world and the method does not provideinformation 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 physicalsystems that influence the climate of a region. They are rapidly becoming the most widely applied downscaling technique. New systems like PRECIS—usedby 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 relationsfor events in 2050 that can include both floods and droughts. Boundary conditions need to be derived from coupled GCMs. Nevertheless, RCMs also havea number of potential pitfalls that need to be understood. They require large amounts of boundary data linked with the parent GCM, acquire the errors ofthe 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 oneAOGCM. Before applying the RCM, the user needs to understand whether it can model the types of events (e.g., typhoons) that are of particular interestin the study. Typically, for studies that involve downscaling analysis such as disaster risk assessments, when trying to incorporate climate change riskfactors 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 temperatureincrease) that are generated are applied in estimatingflood impacts at the city level. Trying to quantifythese uncertainties—and given a confidence intervalfor the outcomes—is extremely difficult, and someuncertainties—like future emission scenarios—cannotbe assessed. However, despite these uncertainties,the IPCC AR4 recognizes that the AOGCMclimate change models do allow global forecasts,and that increasingly the downscaling techniquesare providing information on the likely scale of differentclimate impacts at local levels (Giorgi 2008).Estimates for changes in temperature andprecipitation in 2050 based on statisticaldownscaling approachThe statistical downscaling technique used in thestudy (see Sugiyama 2008) 16 provided estimates of(a) the expected temperature increases, (b) extreme24-hour precipitation increase factors, and (c) seasonalmean precipitation increases for 2050 for thetwo climate change scenarios A1FI and B1. Thesefactors were derived from16 AOGCMs models forall four cities and are shown in Table 2.1.The robustness of the relationships was examinedstatistically by the IR3S team by examiningthe scatter among the data points generated by thedownscaling of the 16 AOGCMs. They found robustrelations for global and local temperature andprecipitable water (used as a proxy for extreme precipitation)relationships, and viable but less strongrelationships for the temperature/mean seasonalprecipitation increases. How this was handled isexplained in Box 2.2.16This part of the analysis was carried out by IntegratedResearch System for Sustainability Science (IR3S) at theUniversity of Tokyo. The results are presented in Sugiyama(2008).Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 7

Table 2.1 ■ Climate Change Forecasts for 2050City Manila Bangkok Ho Chi Minh City KolkataApprox. longitude 121° E 100° E 106° E 88° EApprox. latitude 14° N 13° N 10° N 23° N∆T local/ ∆T global[unit less] 0.88 0.94 0.89 0.90∆T local(A1FI) [K] 1.8 1.9 1.8 1.8∆T local(B1) [K] 1.1 1.2 1.2 1.2present∆P mean/ P mean2.0 1.5 2.2 5.1/ ∆T global[% / K]present∆P mean/ P mean(A1FI) [%] (June-July-August) 4.0 3.0 4.4 10present∆P mean/ P mean(B1) [%] (June-July-August) 2.6 2.0 2.9 6.6∆P extreme/ P extreme/ ∆T global[% / K]present8 8 8 9present∆P extreme/ P extreme(A1FI) [%] 14 15 14 16present∆P extreme/ P extreme(B1) [%] 9.2 9.8 9.3 11Source: Sugiyama (2008).Box 2.2 ■ Downscaling from 16 GCMsTo develop the estimates, the study first downscaled global mean temperature and precipitation. The pattern scaling technique adopted ignores thedifferences 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 meanprecipitation increases. They found significant variability among them in the simulation. For example, the factor increase for extreme precipitationin 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 increaseamong 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 precipitation),for which there was a more robust relationship within the models, which researchers have indicated would serve as a good scaling factor forextreme precipitation (Allen and Ingram 2002) as a proxy for the increases in extreme precipitation. Such relationships have been established in theliterature (see the report for the references). For mean precipitation, the team found a stronger correlation in monthly mean precipitation estimatesand used these data sets to make the mean increase estimates.Source: Sugiyama, 2008Use of regional climate modeling in HCMCstudyIn addition to the above, the HCMC study also utilizedthe PRECIS model nested in the low resolution(~2o) ECHAM Version 4 AOGCM model (EuropeanCentre for Medium Range Weather Forecasts, Universityof Hamburg) to generate daily hydrometeorologicaldata for the study. 17 For this modeling,A2 and B2 scenarios were used, which for the year2050 do not deviate dramatically from the A1FI andB1 scenarios respectively. For the parent AOGCMused, the A1FI and B1 forecast simulations werenot available. Although HCMC utilized a dynamicdownscaling technique, they also compared theirprecipitation forecasts with those derived from thestatistical method and found the results broadlyconsistent with their forecasts.17The Southeast Asia (SEA) Regional Center (RC) for theSystems for Analysis, Research and Training (START) undertookthe work.8 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Uncertainties in the assignment of futureprobabilitiesHydrometeorological probability assessments (asused in this study) are based on the analysis ofhistorical data of extreme events and assume thatthe underlying population’s statistics (e.g., mean,standard deviation, skew, etc.) do not change overtime. In this regard, the discussion of probabilitiesassociated with future extreme events presentsa conundrum, which is approached differentlydepending on whether a statistical or nested RCMdownscaling technique has been used. For the statisticalmethodology used in this study, a factor increasehas been computed for extreme 24-hour andmean seasonal precipitation events. These factorswere applied to current events with the comparableprobability. For the RCM methodology, simulationsover a 40-year time slice around the study date(2050) allowed calculation of the probability of theextreme events. These approaches have some inherentlimitations. For the statistical methodology,it assumes that the parent population’s inherentvariability remains the same, which could easilybe wrong. For the RCM, decadal variations, whichalso affect the population sets’ characteristics, maynot be reflected. The statistical downscaling “factorincrease” or RCM forecasts can have significanterrors even though they both reflect best practicesin climate science. The specific question of theinfluence of these errors on downstream impactassessments has not been studied in any detail. Inaddition to estimating changes in temperature andprecipitation under different scenarios in 2050, eachcity team also made estimates for changes in stormsurge and sea level rise. Further, where data wasavailable, estimates for land subsidence were alsoconsidered in estimating flooding under differentscenarios. How each city study approached theseissues is discussed in more detail in chapter 3.Hydrological Modelingfor Developing Scenariosof Flood RiskThe second sequential step of the analysis involvedhydrological analysis to estimate the extent offlooding in each coastal city under different scenariosin 2050. Outputs from the downscalinganalysis—including estimates and assumptionsfor a range of climate variables (such as sea levelrise, storm surge) and non-climate variables (suchas land subsidence, land use, population changes,economic growth, existing and planned infrastructure)—wereused as inputs into the hydrologicalanalysis and GIS mapping. This part of the analysisis discussed below.Estimating flooding in urban areas underdifferent scenariosThe risk of loss from a hazard such as floodingis based on exposure and, consequently, is sitespecific. Thus, to estimate the extent of flood riskin each city, hydrometeorologic models specific toeach city (Box 2.3) were developed. These modelswere based on a host of historical and city-specificinformation—such as existing drainage and seweragesystems, soil characteristics, river flow,canals, dams, land subsidence, siltation, existingflood protection infrastructure, and so forth—toestimate future flooding under different scenarios.Floods can occur from a combination of factors,including upstream storm runoff, storm rainfallon the city, or sea level rise and storm surge actingalone or in conjunction with other storm events.These factors were incorporated into the scenarioanalysis.The extent of flooding was estimated in termsof the depth, duration and area flooded and overlaidon GIS maps with land-use mapping of thepopulation (disaggregated by poverty levels andother vulnerability criteria), assets, infrastructure,utilities, and environmental and cultural resources.This was done both for the base case and for theyear 2050 under different climate scenarios. For thepurposes of this study, the city studies assumedthat without climate change, the climate in 2050would be similar to the climate in 2008/base year.This helped assess the scale of physical damagethat can be caused to infrastructure, income, land,populations, etc, in the cities with and withoutclimate change.Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 9

Return periods considered 18In this study, for each of the cities, flood risk wasestimated for events of different intensities or returnperiods, namely a 1-in-10-year, 1-in-30-year, and1-in-100-year floods in the base year and in 2050. 19This was based on a statistical evaluation of historicstorm events that allowed the assignment of probabilitiesto extreme precipitation events 20 associatedwith the floods.Hydrological modeling undertaken toestimate flooding under different scenariosThe hydrological models simulate the movementof water on land after it falls as precipitation. 21 Akey initiating input parameter for such modelingis precipitation (derived for each city from thedownscaling analysis). The manner and timing ofthe precipitation’s arrival in the rivers is driven bythe precipitation intensity and duration, the terrainslopes and land use cover (which can drive infiltrationand evapotranspiration), and the channelcharacteristics. Computer models, in which the relevantinformation is digitized, allow the otherwisedaunting calculation of the movement of waterthrough the watershed and channel. In additionto the input parameter of precipitation and spatialdata affecting the flow of the runoff, the boundary(or tail-water) conditions where the flood exits asystem into the sea can create a backwater effect,raising the water level profile of the river upstreamfrom the sea and increasing the overbank spillageand flood inundation levels. Storm surge can alsocause the movement of the water upstream from thesea into the coastal cities. The hydrometeorologicalmodels were designed to model these interactionsand are based on general principles outlined below(Box 2.3).Manila and Bangkok applied commerciallyavailable software to develop the hydrologic modelsand HCMC utilized hydrologic models that hadalready been developed and calibrated by the governmentfor their catchment area. Outputs from theHCMC hydrologic models can also be linked to commercialor open-sourced software models to refineinundation information. Typically a combinationof models was used in the city studies (Table 2.2).Large watersheds with many sub-basins, reservoirs,diversions, etc. have necessarily complexmodels describing the water transfer within thebasin. For the Chao Phraya River Basin, for example,a complex combination of models neededto be interlinked to accurately simulate flooding inBangkok. The schematic shown in Figure 2.1 belowshows the interconnectedness of rainfall-runoff (onthe upper watersheds), reservoir and river routing,and flood routing (2D). The complexity has an impacton the time and depth of analysis. For example,a simulation run for Bangkok with one set of inputparameters took 24 hours of computer time, and thestudy required over 30 simulation runs not countingthe runs required for calibration. HCMC used asimulation model that had been previously developedand calibrated by the government. Manila’shydrologic models were far simpler and the periodof flooding modeled was only a few days, so thelength of the simulation runs in terms of computertime was far shorter than for Bangkok.Calibration of hydrometeorological modelsAs noted above, to ensure the accuracy of the models,rigorous calibration is necessary. The process requiressignificant trial and error in getting the modelto accurately reflect historic movements of floods asrunoff, through river channels and overland floodevents. In Figure 2.2 below for Manila, the dottedline shows the observed discharge and the blue line18This refers to the recurrence interval or the return periodof a flood and is an estimate of the time interval betweentwo flood events of certain intensity. It is a statistical measureof the average recurrence interval over a long period oftime and is the inverse of the probability that the event willbe exceeded in any one year. For example, a 10-year floodhas a 1 / 10 = 0.1 or 10 percent chance of being exceededin any one year and a 50-year flood has a 0.02 or 2 percentchance of being exceeded in any one year. Thus, a 100-yearflood has a 1 percent chance of occurring in any given year.However, as more data becomes available, the level of the1/100 year flood will change.19HCMC restricted its analyses to 1-in-100 and 1-in-30-yearfloods.20An extreme event that causes flooding could be fromextreme 24-hour rainfall on a small watershed, a high seasonalrainfall on a large watershed, or a typhoon or cycloneand associated storm surge.21Storm surge and SLR have been incorporated into themodels as boundary tailwater conditions for the hydrologicmodels.10 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Box 2.3 ■ Some Basic Principles for Hydrological MappingHydrologic models use a mass balance approachin developing the empirical relationships, whereinthe volume of inflow (or precipitation) minusthe losses equals the volume of outflow. For therainfall-runoff relationship, the rainfall excess,which appears as runoff in the river channels, isequal to the precipitation minus the evaporation,infiltration, depression storage, and interceptionas depicted in the figure below. The empiricalmodels must be calibrated against observedevents in order to provide useful estimates.The empirical relationships generallycapture physical information (e.g., land useinformation, soil types, antecedent rainfall,area, and gradients for rainfall-runoff models,or river gradient, cross-sections, and roughnessRATE, DEPTH PER UNIT TIMEestimates for river-channel routing models) that are used with parameters to generate the input-output relationships. The parameters are adjustedin 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).INFILTRATIONDEPRESSION STORAGE+ INTERCEPTIONCONSTANT INTENSITY RAINFALLTIMERAINFALL EXCESSEVAPORATIONTable 2.2 ■ Summary of City CaseStudy Hydrologic ModelingCity StudyBangkokHCMCManilaHydrologic models usedRainfall-runoff modeling of the upper Chao Phrayawatershed entailed dividing the basin into 15 subbasinsand using NAM model; the MIKE FLOOD softwaremodel was used for the lower Chao Phraya coupling theMIKE 11 and MIKE 21 for 1D and 2D flows.*Rainfall in the upper watershed of the Saigon andDong Nai rivers are captured by the Dau Tieng and TriAn reservoirs; outflows from these were modeled bythe Southern Institute of Water Resource and Planningbased on historical (2000) flooding. The HYDROGISmodel was used for the modeling of HCMC. It has beenspecifically designed and calibrated by Vietnam’sMinistry of Natural Resources and Environment’sInstitute of Meteorology, Hydrology and Environmentand models both 1D and 2D flows.The NAM software was used to model the rainfall-runoffrelations linked to the MIKE FLOOD; MIKE 11 and 21were used to model floods in Manila.shows the simulated discharge using the modeland the actual input precipitation data, indicatingthe robustness of the model. This type of “groundtruthing”was done for the hydrological modelingin the Bangkok and HCMC studies as well.Figure 2.1 ■ HydrometeorologicalModel Schematic forChao Phraya WatershedNote: * When the water is flowing in a channel, the model is frequently referred to asone-dimensional (1-D) and overland flow models are referred to as two-dimensional(2-D) models.Source: Panya Consultants.Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 11

Figure 2.2 ■ Manila Rainfall-Runoff Calibration HydrographsSource: Muto et al. (2010).Infrastructure scenarios and additionalclimate inputs into hydrological modelingSince the extent and impact of flooding in 2050 indifferent cities will depend not only on climate andweather variables but also on the infrastructureavailable in the cities to drain water and preventflooding, the hydrological models also assumedcertain infrastructure and flood protection scenariosto estimate flood occurrences. Further, the studiesalso made estimates and made assumptions regardingstorm surge and sea level rise as inputs to thehydrological modeling. These assumptions andestimates are discussed in more detail in chapter 3.Approach to AssessingDamage CostsThe main impact of climate change on the four citiesin 2050 is assumed to be in the form of increasedflooding. Flooding has direct and indirect effects.The direct impacts have to do with immediatephysical harm caused to “humans, property andthe environment” (Messner et al. 2007) and involvecosts from loss in the stock of infrastructure, tangibleassets and inventory, agricultural and environmentalgoods, and injuries and life loss. 22 The indirectimpacts of flooding are a result of a loss in the flowof goods and services to the economy (ECLAC 2003;Messner et al. 2007). 23 These costs are in terms oftraffic delays, production losses, or emergency expenditures.Some delays may be immediate, whileothers may be more medium term because of thesecondary effects of flood-related transportationbottlenecks and re-directed traffic, for example.Table 2.3 summarizes the different direct and indirectand tangible and in-tangible costs associatedwith flooding.The Bangkok and Manila case studies follow theapproach outlined above and estimate direct andindirect costs for four areas: (1) buildings, industry,and commerce; (2) transportation and relatedinfrastructure; (3) public utilities such as energyand water supply and sanitation services; and (4)people, income, and health. The HCMC study alsodiscusses the impacts of extreme events on naturalecosystems. However, the HCMC study does notprovide any detailed monetization of impacts. Themore macro approach it takes to damage cost assessmentis discussed in detail later in this chapter. The22One of most difficult of the direct costs to value is loss oflife. Manila and HCMC have both witnessed loss of livesduring major storms in the past but do not attempt to valuesuch losses from future storms.23It is important to be careful not to double count damagessince the value of the sum for flows from an asset wouldequal the value of the asset.12 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 2.3 ■ Direct and Indirect Costs from FloodingTangibleDirect Costs Repair, replacement, and cleaning costs associated with physical damage to assetsPublic infrastructureCommercial and residential buildingsInventoryVehiclesLoss of productive land and livestockCrop loss 1IndirectCostsLoss of industrial production or revenuesIncreased operational costs for commercial entitiesLost earnings or wagesTime costs from traffic disruptionsPost-flood flood proofing investments by the private sectorEmergency flood management costs to the public sectorIntangibleLoss of human lifeLoss of ecological goodsLoss of archaeological resourcesLong-term health costs from pollution or floodinjuriesPost-flood recovery inconvenience andvulnerability1Income 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 toinclude. 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 todevelop cost estimates are discussed below. Figure2.3 summarizes the main steps taken in estimatingcosts and then estimating the net benefits frominvestments to reduce flooding.Damages to buildings, industry, commerce,and residents in Bangkok and ManilaA key impact of floods is on existing buildings.Floods are assumed to cause damage but not destroyFigure 2.3 ■ Estimating Impacts—A Flow ChartCurrent and 2050Floods (all cities)1 Area flooded2 Duration of flood3 Depth of floodbased on different climatechanges, land subsidence,and assumptionsabout flood control infrastructureImpacts in floodedareas on:1 Industry and Commerce2 Land, agriculture andecosystems3 Transportation4 Energy, water andsanitation5 Income, population,and healthbased on flood durationand depthValuation of damages(Bangkok and Manila):1 Repairs of buildings2 Losses in household assets andcommercial inventory and assets3 Transport delay and infrastructurecosts4 Revenue losses to the public sector(electricity utilities, hospitals,transportation authorities, waterand sanitation utilities)5 Wage and income lossesVulnerability assessment (HCMC)and valuation of:1 Infrastructure at risk2 People at risk3 Land at risk4 Likely GDP losses5 Ecosystems at risk (no valuation)AdaptationIdentification of adaptationstrategies (all cities)Estimation of adaptationinfrastructure costs andannualized benefits(Bangkok and Manila)Estimation of discountednet benefits (Bangkok andManila)Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 13

Figure 2.3 ■ Estimation of Damage to Buildings, Assets, and Inventories in theBangkok and Manila Cases1Identify buildings inflooded areas usingbuilding /infrastructuresurveys.2Classify buildings intocommercial, industrial,residential or othersub-classifications.3Identify assets andinventories in buildingsbased on surveysor on the type ofbuilding.4Establish average valuesof buildings andassets/inventoriesbased on tax information,industrial surveys,householddurable wealth dataetc.5Apply a damage rateto different buildingsand assets/inventoriesto establish actualdamage at differentlevels of flooding. Thedamage rate isobtained from previousflood informationor from secondarysources and varies withthe level of floods.buildings. The steps used in estimating damagecosts in the case of Bangkok and Manila are identifiedin Figure 2.4.Identifying, classifying, and valuing damageto buildingsThe Bangkok and Manila case studies used informationfrom building censuses and other infrastructure-relatedstudies to classify buildings first intoresidential, commercial, and industrial buildings.Based on national statistical organization classifications,these categories of buildings were furthersubdivided. For example, commercial buildingsalone were subcategorized into nine groups in thecase of Bangkok.Once buildings were classified into differentsubgroups, they were valued based on availablegovernment data. The Bangkok case study identifiedthe book values of buildings of different typesbased on construction costs when the buildingswere legally registered. These values were thendepreciated to reflect aging. In Manila, the baseunit construction cost of residential and commercialbuildings was obtained from the governmentassessor’s office and the median value of differenttypes of buildings estimated. Then, based on theestimated depth of floods (output from the hydrologicalanalysis), the value of floor areas of floodaffectedbuildings was identified.In order to estimate the actual cost of damageto buildings, the case studies used different damagerates depending on the level of the flood. 24 The damagerate or the percent of the value lost from floodsis dependent on the depth of inundation. A differentrate is applied to assess damage to buildings anddamage to assets. The Bangkok study used a 1997survey to establish damage rates; the rate was usedfor all types of buildings and housing (residential,commercial, and industrial). The Manila study useddamage rates, which vary across buildings and assets.Assets such as machinery, office furniture,and inventory damaged by floods also need to bevalued. The loss of assets from damaged buildingswas evaluated based on “representative assets”in different types of buildings. For example, inthe case of the Bangkok case study, average assetvalues obtained from census data for buildings areTHB 328889 ($9,873) for Bangkok and THB 220180($6,609) for Samut Prakarn. In the Manila study,the value of stocks, assets, and inventories of commercialand industrial enterprises was obtainedfrom the National Statistical Organizations’ data of24The flood damage rates used in the case studies were obtainedfrom either local surveys or from secondary sources.14 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

establishments. A damage rate (Table 2.4) was appliedto assets and inventories to determine losses.The Bangkok study used a similar approach.To estimate losses to residential assets, the valueof household goods first needs to be established. Inthe Manila case, the value of household effects wasestimated to be 35 percent of the value of residentialconstruction costs of flooded buildings. Similarly, finishingswere estimated to be 25 percent of constructioncosts. In the Bangkok study, average householdassets or durables were obtained from the Populationand Housing Census of 2000. The value of theseassets was based on market prices. Once the assetvalues were obtained, damage rates were used toestimate the losses from different levels of flooding.Assessing impacts on roads andtransportation networksFlooding affects road and train networks in all thecities. There are direct and indirect costs associatedwith loss of transportation infrastructure. Directcosts are (a) construction or repair costs; and (b)any losses of vehicles or damaged trains and so on.These costs are estimated where applicable, basedon average repair costs. Indirect costs associatedwith damage to transportation infrastructure includeincreased vehicle operations costs (VOC) resultingfrom damages to roads from flooding or because ofchanges in traffic pattern (increased petrol costs), ortime costs from losses in productivity (ECLAC 2003).More than 90 percent of inhabitants of Bangkokand Samut Prakarn travel by roads and highways.However, the road network in Bangkok, SamutPrakarn, and the BMR are set at an elevation of1.5–2.0 meters above ground, and most of themare covered with reinforced concrete pavement or7–10 cm of asphaltic pavement. Therefore, they arenot expected to incur flood damages. Rail lines aregenerally set at about 1 meter above the surroundingground level. The Bangkok rapid transit system,both elevated and underground, has been designedto be protected from overflow flood water. Thus,the costs associated with transport networks fromflooding in Bangkok are expected to be minimal.The Manila study undertook a detailed analysisof damages to the transport sector. First, based onthe transport master plan, future road constructionTable 2.4 ■ Flood Damage Rate by Typeof Building in ManilaType of BuildingAffected AssetsFlood Level100–200 cm 200–300 cmResidential Finishings 0.119 0.58Household Effects 0.326 0.928Business Entities Assets 0.453 0.966Stocks 0.267 0.897Source: Adapted from Manual on Economic Study of Floods, JapanNote: Damage rates are available for less than 50cm and more than 300 cm as wellwas identified. The length of major and minor currentand future roads likely to be affected by floodswas then established. Based on unit maintenancecost data from the Department of Public Worksand Highways, the cost of maintaining these floodaffectedroads was estimated. The Manila study alsoestimated indirect costs from flooding related totime delays from traffic disruptions and problemswith vehicle operations. In the Manila case, thevehicle operations cost (VOC) is obtained from theDepartment of Public Works and Highways and includesfixed costs of operating vehicles and runningcosts. The costs of floods or the change in VOC perkm was estimated as the difference between VOCon flooded roads and on roads in good condition.This difference in VOC is multiplied by the numberof affected roads in 2008 and 2050 to obtain damagecosts in these two periods. To estimate time-delaycosts, data on flood related transport delays andthe number of road commuters were obtained fromsecondary sources and transportation surveys. 25The number of commuters delayed by floodingwas then multiplied by average hourly income asa proxy for time costs.Assessing direct and indirect costs inenergy, water supply and sanitationEnergy, like the transportation sector, may sustaindirect and indirect damages from floods. Directdamages to electricity generation plants, transmis-25The Manila study draws heavily on the government’sTransport Master Plan, which lays out potential transportationchanges up to 2015, and on the Metro Manila UrbanTransportation Integration Study (MMUTIS).Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 15

sion and distribution lines, and energy distributioncenters depends on whether these facilities are inthe flood-affected area of each city. If they are, thenthe extent of damage is estimated and replacementand repair costs are used to value these damagesbased on cost information available with governmentdepartments. In most cases, electric utilitieshave strict rules for cutting electricity during certaintypes of floods. This leads to revenue losses, whichwere estimated in the Bangkok case study. In termsof direct costs, future power generation plants forBangkok and Samut Prakarn are planned outsidethe cities and there will be no direct impact. TheManila study did not estimate energy-related costs.As with the case of the energy sector, costs towater and sanitation systems are evaluated only ifit is established that there is a likelihood of flooddamage to specific subcomponents such as waterintake facilities, pumping stations, water treatmentplants, main lines to storage tanks, storage tanks,distribution networks, and so on. In the case ofBangkok, it is assumed that water supply will becomedysfunctional if water floods reach 2 metersabove the ground surface. This will disrupt waterservices to the area and result in revenue losses tothe utilities as well as other costs, including healthcosts. The calculation of sales loss for the water supplysystem due to flood damages in Bangkok wasestimated as follows:Water supply sales loss = No. affected usersx water demand per user x water sales ratex flood durationThe data for this estimation came from the five-yearrecord of the Metropolitan Waterworks Authority,which indicates water sales for a residential unit are0.48 m 3 per day per household and water revenue is2.06 baht per m 3 . Water sales for a nonresidential unitare 3.71 m 3 per day per customer and water revenueis 2.83 baht per m 3 . In the Bangkok study, direct damageto the water supply and sanitary infrastructureis not assessed since they are protected from theworst possible flood in the future. The calculationof income losses to sanitation service providersbecause of the shutdown of sanitation services wassimilarly undertaken. The Manila study did not findmajor costs associated with the water delivery sectorbecause of the position of the pipes and extentof pressure in them.Estimating income lossesAs previously noted, each city estimates the populationlikely to be affected by 2050 floods. Wherepossible, survey data and other secondary sourceswere used to identify the characteristics of the affectedpopulation, particularly to examine whetherfloods will affect the poorest communities.Income losses to the commercial sectorIn order to estimate income losses from floods, theManila and Bangkok case studies examined lossesto firms and individuals. During the duration of theflood, income losses are borne by the private commercialsector because of a halt in their activities.Since it is difficult to directly estimate these costswithout detailed production information, whereavailable, national statistics on income per day associatedwith different kinds of commercial structureswere used.The Bangkok study identified the number ofbuildings in flood-affected areas and then, basedon the duration of the flood, estimated the rate ofdamage to these buildings. To obtain business losses,the Bangkok case study then identified the averageincome per day from commercial establishmentsbased on business surveys. The loss in value addedper day (net income) was adjusted due to savings inbusiness expenditure when an operation is closeddown. The net commercial income was estimatedto be THB 4930 ($149) per establishment per day.A similar accounting was used to estimate averageincome to industrial establishments. These averagevalues are multiplied by the number of days of floodingand the number of buildings affected by floods toestimate income losses from commerce and industry.The Manila case study took a similar approach.The analysis was based on a 2008 survey on businessincome losses to flood-affected firms. The studyused data on sales losses as a substitute for incomeor profit data, which were unavailable. Buildingsthat were affected by floods were categorized accordingto different commercial activities; each floorwas assumed to host one firm. By multiplying the16 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

sales loss data (categorized by type of economicactivity) by the number of affected buildings in eacheconomic category, the study established the valueof income losses to the commercial sector.Estimation of income losses for poor andnon-poor householdsHouseholds who work in flood-affected areas maylose income during the duration of the flood. Ageneral assumption made in the Bangkok case isthat salaried workers will not see income lossesfrom floods. However, the informal sector needsfurther accounting. In Bangkok, the low-incomehousing developments, called condensed housing,in flooded areas are identified and—based onestimates of household size—the number of dailywage earners in this area is estimated. Then basedon the daily wage rate, the income loss to poor andnon-poor households is calculated.The Manila case study examines income lossesto workers in the formal and informal sectors. Becauseof lack of data on number of households, thecase study, like Bangkok, relied on the number ofbuildings affected by floods. Non-poor residentswere assumed to occupy residential buildings at therate of 1.5 households per building. This allowed theManila study to estimate the total number of nonpoorresident households in flood-affected “formal”residential buildings. The total number of affectedhouseholds was multiplied by the average incomeper household (based on government statistics) toobtain income losses to the formal sector.To establish losses to the poor, the Manila studyfirst estimated the number of informal structuresaffected by floods. It was assumed that at least twohouseholds live in each structure. Thus, the totalnumber of affected households was estimated.Again, using government statistics on the povertylevel per day of PHP 266 per household, the totallosses to the informal sector was then established.Limited analysis of health impacts in thecity studiesAn important question concerns the impact of floodson public health. Each city is different and has healthimpact data based on previous experience with floodsand expert advice. The main diseases associated withfloods are diarrhea, conjunctivitis, athlete’s foot,malaria, cholera, and typhoid. Ideally, we wouldrequire information on different outbreaks and costper episode per person for specific diseases in orderto value health impacts. However, this is difficultto obtain and some approximations are made. TheBangkok study makes an attempt to examine healthcosts associated with floods. In this case study, thenumber of individuals affected by floods is estimatedand multiplied by the average per person health costsassociated with hospital admissions. This is clearly afirst approximation for what might be the real costson health from flooding. However, lack of data, anddifficulties in establishing dose-response informationrelated to different flood-related diseases, leadto this more simplistic analyses. The Manila studydoes not estimate the health costs from floods eventhough it makes an attempt to estimate some of thehealth impacts of flooding. A more detailed analysisof health impacts needs to be carried out as a followupto this study.Assessment of Damage Costsin the HCMC StudyThe HCMC study used a macro-approach to assessdamage costs. Two different methods were used tocalculate the cost of climate change at an aggregatemacro-level: (1) cost estimates using land values; and(2) cost estimates based on aggregate GDP loss. Themethodology applied in each of these approachesand the results obtained are described below.The land value approachEconomically speaking, the value of any asset isthe sum of the value it is expected to generate overtime, discounted to reflect people’s preferences forconsumption in the present. The same is true forland. In principle, the current value of the land stockin HCMC is the value of future production that canbe expected to be generated with this asset. Landvalue can be a particularly good guide to the cost ofclimate change as land values capitalize most valuesthat need to be captured in any assessment of cost(such as roads, railways, water supply systems,Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 17

Figure 2.5 ■ Possible Relationships between Flood Duration and Land Value LossLoss of land values10.90.80.70.60.50.40.30.20.1Linear relationship betweenpercentage of time the area isflooded and % loss of land valueQuadratic relationship betweenpercentage of time the area isflooded and % loss of land value0 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 flood/365)Source: ADB (2010).drainage systems, energy infrastructure, hospitaland schools). An estimation of how the impacts ofclimate change may affect the value of the land stockin HCMC thus provides a first-order approximationof the economic costs of climate change for the city.The first step in the analysis was to determineland prices. Land price data was aggregated bydistrict and an average land price determined foreach district. The administratively determinedprice for land is published by HCMC People’sCommittee annually, as required under the 2004Land Law. These prices are used to determine thelevel of compensation for state appropriations ofland and for taxation purposes. In principle, theseadministrative prices are based upon the potentialproductive value of the land, rather than marketprices (although following the 2004 Land Law, theadministrative price for land has moved closer toland market values). These are current land prices,not 2050. These are also not market prices.The area subject to flooding both in extremeevents and regular flooding was then determinedunder future scenarios using the HydroGIS modeling.Once the average land price and floodingextent and duration were estimated, the relationshipbetween the flooding and the decline in theeconomic value of the flooded land was determinedto allow the calculation of the cost of flooding dueto climate change. The study assumes that there issome positive relationship between the duration ofthe flooding and the loss in land value: that is, thelonger the period of flooding, the greater the loss inland value. However, the precise nature of that relationshipbetween the economic value of land andthe number of days in any given year that an area isflooded is not known, and must be assumed.In the HCMC report, two types of relationshiphave been assumed: The loss of land value is directlyproportional to the proportion of time the land isinundated; this is the linear relationship in Figure 2.5.For example, if the land is flooded for 182 days peryear, (about half of the time), this assumption yieldsan estimated loss of 50 percent of the land value. Aproportional relationship would overestimate theloss in land value. Indeed, if land is flooded only afew days a year, we might suppose that this wouldhave a limited impact on the value of the land, if any.On the other hand, the proportion of the loss of landvalue may increase at an increasing rate as the lengthof time the land is flooded increases, reaching 100percent of land value lost when the land is floodedfor 100 percent of the time (the quadratic relationshipis shown in Figure 2.5). The quadratic relationshipis considered to represent a more realistic relationshipbetween flood duration and land value loss assmall periods of flooding (e.g. for one or two days)are unlikely to influence land value much, whereaslonger term flooding (e.g. for 300 days a year) islikely to reduce the value to near zero.The GDP approachThe second approach to estimating the costs ofclimate change is to calculate costs on the basis ofexpected losses in production (proxied by GDP) dueto climate-change-induced flooding. Like land, GDP18 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

measures capture a number of important values (levelof economic development, extent of industrial activities,transportation system, water supply, drainage,etc). To calculate costs using this method, a numberof important assumptions were integrated into theanalysis: Areas subject to flooding lose all theirproductivity for the duration of the flood. The GDPproduced in a particular area is directly proportionateto population (i.e. if an area accounts for 1 percent ofthe population it will also account for 1 percent of theGDP). GDP is produced evenly over time (i.e. with1/365th of annual GDP being produced everyday).Based on GSO statistics, GDP growth will average11.5 percent between 2006–10; 8.7 percent between2011–25; and 8 percent between 2026–50. Populationgrowth will take place in line with the high estimatediscussed in chapter 4. Future values are discountedat an annual rate of 10 percent. 26Given these assumptions and the informationon flooding derived from the HydroGIS models, anestimation of the costs due to flooding is relativelystraightforward. For each district and for each yearuntil 2050, the following calculation was performed:Annual costof flooding=Number of days the district is flooded xnumber of people affected x GDP/capita/dayThe annual cost of flooding was calculatedfor each year between 2006 and 2050, and the discountedcosts summed for the whole period. TheHCMC study presents results from both the landvalues and GDP measures. The differences in theresults and approaches are discussed in chapter 4.Assumptions aboutthe Future of Cities inEstimating Damage CostsCities in 2050 are likely to be vastly different fromtoday’s cities. Understanding how different is ahuge task and there is no attempt to model economicgrowth and link it to urban development. Instead,existing data and official plans and projections areused to understand and develop GIS-based maps.Further, in assessing damage costs, each city studymade assumptions about population size and distribution,poverty, economic growth and city-specificgross domestic product, urbanization, flood andtransportation infrastructure, and spontaneousadaptation and migration. Where possible, theybuilt on existing plans and studies undertaken bycity governments related to urban growth, publicutilities and transportation networks, and so forth.Population in 2050 is estimated based on the bestavailable local projections for each city. The populationdistribution within each city in 2050 buildson available estimates of population density anddistribution and some understanding of futurepopulation dynamics. Economic growth and povertyprojected in 2050 varies across the case studies.These assumptions are made explicit in the followingchapters with respect to each city.Human spontaneous adaptation to the possibilityof increased flooding is considered but is notfully incorporated into the studies. This is partlybecause of the difficulties in predicting adaptationbehavior and partly because populations, whilelikely to respond to incremental change, may nothave a measured response to the type of extremeevents considered in this study. 27 The case studiesdo take into account human migration in identifyingpopulation distribution in 2050. In general, populationsexposed to flooding are likely to continue tobe exposed to flooding except in cases where thereis improved flood infrastructure put into place.Assumptions about pricesGiven difficulties in estimating future prices, allthe three case studies largely keep the real value ofgoods and services unchanged. The Bangkok studyestimates the costs of damage in 2008 using current26Because in the case of land, the land value already representsthe discounted future income stream capitalized inland value, so there is no need for discounting when usingland values to estimate the cost of climate change in the previousanalysis.27Yohe et al. (1995), for example, argue that if permanentinundation is expected due to SLR, it might be appropriatein cost-benefit analyses to treat the value of inundatedbuildings as zero or very low. If there is full information andefficient adaptation, economic agents can be expected to letthese buildings fully depreciate before they are inundated.In our study, we do not assume full information. Nonetheless,rational households who can afford to move are likelyto move away from flood-prone areas, which may be thenpopulated by poorer communities.Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 19

prices and the same prices are applied to 2050 damages.The Manila study assumes a 5 percent annualgrowth in the value of goods and services lost fromfloods; that is, damages grow annually at 5 percentup to 2050. However, for comparison, the Manilastudy deflates its 2050 values and presents the 2008values of all damages whether they occur in 2008 or2050. The HCMC study uses a macro approach toestimating costs and examines current land valuesas well as annual income loss based on annuallyprojected GDP. 28The interest rate is an important price variablefor assessing adaption investments. The case studiesuse discount rates that are normally used to discountpublic investments in each city. The Bangkokstudy presents results using 8 percent, 10 percent,and 12 percent rates. The Manila study uses a 15percent discount rate and the HCMC study uses a10 percent rate.Conclusion:Methodological Limitationsand Uncertainties inInterpreting Resultsof the StudyDifficulties with 2050 projectionsWhile the studies attempt to establish some understandingof how the cities may look in 40 or so yearsfrom now as is discussed in the following chapters,this is really very difficult to do. Hydrologicalmodels in all three studies incorporate flood controlinfrastructure and the type of land use that may existin the future. Future transportation networks areincorporated into damage assessments undertakenby Manila and Bangkok, and official populationprojections are used. Yet there are numerous assumptionsunderlying each of these projections andthe future may look different from what is assumed.The road networks may not be built, new land useplans may be put in place, and exogenous shockscould change the population distribution. We canimagine a situation where flood-prone areas aremade into wetlands or parks and set aside for someform of recreation. In these circumstances, the costsof flooding may be very different.Impacts on populations working in floodedareas are another area of concern. It is hard to establishwhat categories of people may actually inhabitflood-prone areas in 2050. While the studies incorporatetrends in population distribution, it is notclear that these trends will remain unchanged in thenext forty years. It is rational to consider a scenariowhere increased flooding will lead to commercialestablishments moving away from flood-proneareas. Thus, these areas could become recreationalareas as previously discussed, or they could becomea refuge for the poor and vulnerable because of lowland values. These complex issues cannot be teasedout in the damage costs framework that is used.There are two confounding issues to considerin examining health impacts: (a) climate change islikely to change the incidence and geographic rangeof different diseases such as dengue or malaria, forexample, and it is difficult to predict changes in thegeneral prevalence of these diseases in the three citiesin 2050; and (b) economic growth and increasededucation will potentially reduce the prevalenceof certain diseases such as diarrhea. Each city hasmade its own assumptions about health impactsand provides qualitative and some quantitativeinformation on costs, but this information is subjectto significant uncertainty.Estimating physical damages requires avariety of assumptionsEstablishing exactly how and what will be affectedby natural disasters is not easy. The Bangkok andManila studies examine a variety of impacts onbuildings and income, but a series of assumptionsunderlie this analysis. In order to estimate incomelosses, for example, both case studies have to makeassumptions about the number of households andfirms residing in flood-affected buildings. Without acareful survey of each flood-affected area, it would28In order to examine adaptation-related investments, someadditional assumptions are made. The Bangkok study estimatesnet benefits under two scenarios: (a) prices stay thesame, and (b) net benefits grow at an annual rate of 3 percent .The Manila study uses a 5 percent annual rate of growth.20 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

e very difficult to know how accurate these assumptionsare.Valuation considerationsThere are numerous challenges to valuing floodimpacts that have to do with current and future dataavailability. Flooding damages to building stocksand machinery, for instance, are best valued at theirdepreciated replacement cost. In practical terms,the valuation exercise was carried out in differentways in each study depending on data availability.For example, in Bangkok the value of an entire assetor building was assessed and a percentage ofthe value was treated as the damage cost. Further,in Manila and Bangkok, buildings are evaluated atbook-values, or the cost of the buildings when theywere initially established. In HCMC, land is evaluatedat government-established prices, but in somecases the market value of this land may be higher.Also, the cost of repairing flooded components of abuilding may be very different from the value obtainedby applying a damage rate to the entire valueof the building. Another issue is how losses relatedto electricity and water supply have been evaluatedbecause of lack of appropriate information. The costsof supply disruptions should be assessed by estimatingthe economic value of public water/electricitysupply, or, for example, by estimating the coststo water consumers of securing alternative waterwhen service is interrupted. In Bangkok, water andenergy-related flood costs, however, are estimatedby examining revenue losses to the government.Another important issue is the role of prices.It is very difficult to estimate the value of assetsand incomes damaged by flooding in 2050. Givendifficulties in predicting prices, most prices areheld constant or in some cases some reasonablegrowth rates are assumed. However, urban realestate prices and the prices of a variety of goodsand services damaged by floods can change significantlyand this would affect the evaluation in 2050.There are also difficulties in estimating income orproductivity losses as a result of floods. The Manilastudy and Bangkok (for low income households)estimate the number of households affected byfloods and multiply this by an average value ofhousehold income based on the type of housesthese families live in. However, average valuesused may not apply for flood-prone areas. Further,the families may have a way of earning incomeeven under flood circumstances if they can leavethe area for work. In the HCMC study, a macroapproach is applied and GDP per capita is usedas the value of productivity lost in flooded areas.However, if these areas may not be producing the“average GDP per capita,” this could lead to anoverestimation of impacts. These uncertainties andlimitations need to be borne in mind in interpretingthe results of this study.Methodologies for Downscaling, Hydrological Mapping, and Assessing Damage Costs | 21

3Estimating Flood Impactsand Vulnerabilitiesin Coastal CitiesIn this chapter, the main findings of the hydrologicalanalysis—in terms of potential increasein flooding in 2050 under different scenarios—are presented. For each city, the analysis presentsthe main drivers of flooding, how climate changeis likely to influence flooding, and the potentialphysical consequences on the city and its residentsin 2050. The findings from the hydrological modelingwere overlaid on GIS maps, thus relating flooddepth/duration and area flooded (estimated fromthe hydrologic modeling) with maps of futureland use; location of residential, commercial, andindustrial areas; and essential infrastructure andenvironmental resources. Use of GIS enables avisual mapping of the area inundated and districtsand sectors likely to be at risk from flooding underdifferent scenarios in 2050. 29 For the purposes of thisreport, only a selection of GIS maps is included; amore extensive set of maps can be obtained fromthe city-level studies.Estimating Future ClimaterelatedImpacts in BangkokThe Bangkok Municipal Region covers an area ofabout 1,569 sq kms and is located in the delta of theChao Phraya River basin—the largest river basin inThailand, covering an area of 159,000 sq kms, or 35percent of the total land area of the country (Figure3.1). The basin area is flat at an average elevationof 1–2m from mean sea level. Some areas are at seaFigure 3.1 ■ Location of Bangkok in theChao Praya River BasinTHAILANDCHAO PHRAYA RIVER18°M YA N M A RAndamanSea0 100KILOMETERSPROJECT RIVERSELECTED CITIESNATIONAL CAPITALINTERNATIONALBOUNDARYThis map was produced by the Map Design Unit of The World Bank.The boundaries, colors, denominations and any other informationshown on this map do not imply, on the part of The World BankGroup, any judgment on the legal status of any territory, or anyendorsement or acceptance of such boundaries.MYANMARNAYPYIDAW15°NArea ofMapAndamanSea15°N15°NLAOP.D.R.VIENTIANETHAILANDBANGKOKGulf ofThailandPHNOMPENH100°E MALAYSIA 105°EHANOIVIETNAMCAMBODIASource: Thailand map IBRD 37477.14°Chiang MaiTakPathum ThaniNonthaburiBangkokSamut PrakanRatchaburi SamutSakhonGulf ofThailand100°level due to land subsidence. Bangkok straddlesthe Chao Phraya River approximately 33 kilometersabove the Gulf of Thailand.29For 2050, projected land use changes—where available—were incorporated into the GIS to assess the future impacts.YomKamphaengPhetPing100°Nakhon SawanUthai ThaniSupham BuriTha ChinPhraeUttaraditPhichitNanChainatChao PhrayaNanT H A I L A N DSing BuriThahanbok Lop BuriAng ThongL A OP. D . R .Phra Nakhon Si AyutthayaIBRD 37477102°102°20°18°16°14°JANUARY 201023

Importance of BMR to the regionaleconomyThe Bangkok Metropolitan Region (BMR) is theeconomic center of Thailand. It is the headquartersfor all of Thailand’s large commercial banks andfinancial institutions. The area to the east of Bangkokand Samut Prakarn is also an important industrialzone. In 2006, the gross domestic product (GDP) ofthe Bangkok Municipal Region was 3,352 billionbaht, or 43 percent of the country’s GDP (7,830 billionbaht). The annual average growth rate was 7.04 percent,and per capita GDP was 311,225 baht (Bangkokcase study report). The current official population ofBangkok is estimated to be about 10 million people(based on 2007 estimates) with an estimated growthrate of .64 percent between 2003–07. 30Nature of urban povertyAs stated in the Bangkok study, in 2007, 0.6 percentor 88,361 people in the BMR were poor (Table 3.1).The poverty line for the Bangkok municipal regionwas 1,638 baht ($49) per person per month in 2007. 31The number of poor is an official estimate and doesnot include unregistered people. Most of the poorlive in condensed housing and are unregistered.Statistics from the Office of National Economic andSocial Development Board (NESDB 2007) show768,220 people living in 133,317 housing units ofthe condensed housing area. Therefore in the assessmentof flood impacts on the poor, this housingunit and population was used as a base for the assessmentof income loss of the poor.Climate and precipitation in the ChaoPhraya River BasinBangkok has a tropical monsoon climate. The averageannual rainfall over the basin is 1,130 mm, and ishigher in the northeastern region of the basin (Table3.2). About 85 percent of the average annual rainfalloccurs between May and October (Panya Consultants2009). Tropical cyclones occur between Septemberand October. In this case, rainfall continuesfor a long period of time in a relatively wide area.The peak river discharge is registered in October atthe end of the rainy season. Severe flood damagemay arise with high tide in this period (ibid.). BMA’smaximum temperature is in April and its minimumtemperature is in December. The mean temperatureranges from 26 ° C to 31 ° C.Main climate-related drivers of floodingSevere flooding in Bangkok is associated withheavier than normal rainfall occurring over several30This figure does not include unregistered migrants.31Poverty incidence is measured at the household level bycomparing per capita household income against the povertyline, which is the income level that is sufficient for anindividual to enjoy society’s minimum standards of living.If an individual’s income falls below the poverty line, he orshe is classified as poor.Table 3.1 ■ Poverty Line and the Poor in the BMR 1Poverty Line (Baht/person/month) Proportion of the Poor (%) No. of the Poor (person)Province2006 2007 2006 2007 2006 2007BMA 2,020 2,065 0.51 1.14 28,692 64,422Samut Prakarn 1,647 1,712 — 0.78 — 9,961Samut Sakhon 1,511 1,564 0.76 0.42 4,313 2,436Nonthaburi 1,529 1,561 0.30 0.06 4,124 845Pathum Thani 1,409 1,458 0.56 0.20 5,376 1,939Nakhon Pathom 1,434 1,466 0.45 0.98 3,918 8,758BMR 1,592 1,638 0.43 0.60 46,422 88,361Source: The Office of National Economic and Social Development Board (NESDB), 2007 as cited in the Bangkok city study.1Most 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 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 3.2 ■ Bangkok Monthly Average Temperature and PrecipitationMonth Temperature Average Low ( o C) Temperature Average High ( o C) Average precipitation (mm)January 23° 32° 7.2February 24° 33° 13.7March 26° 34° 32.5April 27° 35° 58.3May 27° 34° 170.8June 26° 33° 112.5July 26° 33° 118.1August 26° 33° 160.0September 25° 33° 262.5October 25° 32° 207.3November 24° 32° 32.4December 22° 32° 3.3Total 1,171.4Source: http://weather.uk.msn.com/monthly_averages.aspx?wealocations=wc:THXX0002months during the monsoon period over the ChaoPhraya watershed. The flooding can last severalweeks to over a month. Intense short-duration rainfallover the city can cause localized flooding thatnormally lasts for less than 24 hours. Bangkok isnot subject to direct hits from tropical typhoons orcyclones. Large floods have occurred in 1942, 1978,1980, 1983, 1995, 1996, 2002, and 2006. A numberof dams in the upper watershed—including theBhumibol Dam (1964) and Sirikit Dam (1971)—havehelped reduce flooding in Bangkok. In 1995 a seriousflood occurred due to a series of tropical stormsreaching the upper watershed from the end of Julyto September. The resultant runoff exceeded the storagecapacity of the Sirikit Dam, which had to release2,900 million m 3 of storm runoff. Despite efforts toattenuate the flood wave by allowing flooding ofagriculture lands upstream of Bangkok, the floodwave crest reached Bangkok at the same time as thehigh spring tide, flooding 65 percent of BMA withinundation depths up to 2 meters and left some areasflooded into December. The flood event of 1995 isestimated to have a return frequency of 1-in-30 years.Non-climate drivers of floodingIn addition to climate-related drivers, non-climatedrivers such as land subsidence, deforestation,urbanization, and removal of natural attenuationbasins (like wetlands) also contribute to flooding.Land subsidence is a critical factor in flooding andis discussed later in the section.Existing flood protection in Bangkok—diversion canals and flood dikes withpumped drainageIn response to the flooding, Bangkok has implementeda large-scale flood protection system thatincludes a flood embankment surrounding thecity and a pumped drainage system. In 2006 therewas another significant flood event in the upperChao Phraya watershed. Flood peaks were againattenuated by allowing the agricultural lands northof Bangkok to flood, and flood flows were also divertedto channels to the east of Bangkok. The floodembankment system and pumped drainage systemprotected the city. Nevertheless, areas surroundingBangkok continue to experience inundation duringpeak flood flow. The city and national plannershave plans to add further flood protection, whichhas been planned and budgeted through 2014. Thelong-term land use plan until 2057 for the greaterBangkok region provides for development areasand environmental zones that will be used to allowflood flows and drainage. Most of the flood controlEstimating Flood Impacts and Vulnerabilities in Coastal Cities | 25

activities that are being considered in the Royal IrrigationDepartment master plan are infrastructureactivities (such as construction and improvement ofcanals, dikes, pumping stations, etc.). Flooding ofagricultural lands upstream of Bangkok is utilizedduring major floods to help attenuate the flood peak,lowering the flood profile in Bangkok.Current and planned land use and urbanplanning in Bangkok—Bangkok in the futureThe Bangkok study projected what the city wouldlook like in the future based on a wide range ofavailable data about population and growth and anumber of sectoral and urban development plans(Box 3.1). Modern urban land use planning inBangkok started in the 1950s. Bangkok’s inner citycontains the Grand Palace, government offices, majoruniversities and educational establishments, and2-to-4-story row houses that are used as commercialand residential units. The inner city is a nationalhistoric conservation area where construction ofhigh-rise buildings is prohibited. Urban growth inBangkok, however, has progressed in a sometimesad hoc manner and without a unified plan linkingit to the surrounding areas. To remedy this problem,the government has developed a 50-year regionalspatial plan (finished in 2007) that foresees growthof the surrounding areas. This plan also providesfor significant environmental protection and wetlandareas.Climate change parameters—statisticaldownscalingFor the 2050 simulations, the Bangkok City casestudy applied the statistical downscaling factorsestimated by the University of Tokyo’s IntegratedResearch System for Sustainability Science (IRS3)discussed in chapter 2. Given the size of the ChaoPhraya watershed and the long time it takes upperwatershed rainfall to reach Bangkok as stream flow,the team applied the mean increase in seasonalprecipitation (3 percent and 2 percent for A1FI andB1 respectively) to the three-month rainfall eventsfor the corresponding 1-in-10, 1-in-30 and 1-in-100-year precipitation events for the watershed. Thesetotal precipitation levels were distributed over thewatershed using spatial and temporal patterns fromhistoric flood years. The forecasted temperatureincreases of 1.9 ° C and 1.2 ° C for the A1FI and B1scenarios respectively for 2050 were used in the hydrologicalmodeling. The SLR estimates used were0.29 m and 0.19 m for the A1FI and B1 scenariosrespectively. This was based on IPCC AR4 (WorkingGroup 1, Chapter 10, Table 10.7) estimates. It wasdecided to use half of the 2099 increase for the 2050estimates; namely, 0.29 m and 0.19 m for the A1FIand B1 scenarios respectively.Box 3.1 ■ The Bangkok MetropolitanRegion (BMR): SomeAssumptions about theFutureBangkok is changing rapidly and some of these changes havebeen taken into account in modeling the future. Bangkok isincreasingly “suburbanizing.” Between 1998 and 2003, BMR sawa 74 percent increase in urban development. The city’s housingstock has doubled in the last decade. This trend is expected tocontinue. The Government’s Land Use Plan for 2057 is used toproject the number of commercial, residential, and industrialbuildings in Bangkok in 2050. Available government plans aretaken into account in establishing future road and transportationnetworks. Plans suggest tightened urban areas within a networkof expressways bounded by large environmental protection areasto divert any floods from the centers.Based on existing trends, the city of Bangkok is not expectedto grow a lot, while neighboring suburbs will continue to grow.For instance, suburban Nonthaburi increased its population by47 percent during 2002–07. Bangkok City is expected to have apopulation of 10.55 million (4.65 million households) and BMR apopulation of nearly 16 million by 2050. To forecast the populationin 2050 for BMR, information on population projections for Thailand2003–30 prepared by the government (NESDB) was used as a base.The NESDB report projects population to the year 2030 for the wholekingdom and to 2025 and 2020 for Bangkok and other provincesrespectively. In projecting the population to 2050, a regressionfunction was applied.In terms of the BMR economy, regional GDP is projected for twoareas (Bangkok and Samut Prakarn) of BMR most likely to be affectedby flooding. GDP for Bangkok and Samut Prakarn in 2050 is expectedto be six-fold higher than the current GDP. In terms of poverty inthe cities, in Bangkok in 2007 only 88,361 people (0.6 percent ofthe total) in the BMR were officially considered poor. However, thisnumber excludes some 768,220 people living in 133,317 condensedhousing units. In the flood impact assessment, the higher, moreaccurate number was used.Source: Adapted from Panya Consultants (2009).26 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Estimation of storm surge in the case ofBangkokFor storm surge, the team examined historic stormsurges in the Gulf of Thailand. There have beenthree events since 1962 along the coast of southernThailand. No recorded typhoon or cyclone has had atrack that approached the mouth of the Chao Phrayaat the north of the Gulf of Thailand. Nevertheless,one typhoon created wind patterns that drove 2 to 3meter waves near the mouth and drove up sea levels.As a result, the team used a storm surge of 0.61 m forthe study. It needs to be noted that the time frame forthe storm surge would be on the order of 24 hours,and only one event has been recorded in 30 years.The hydrometeorological processes driving theseevents have little or no correlation. To have a jointimpact, the storm surge effect would have to occurnear the peak storm discharge for the Chao Phrayawatershed. The joint probability of occurrence, althoughnot calculated in detail, is less than 1/1000.Table 3.3 summarizes the climate change parametersused to model the 2050 impacts in Bangkok.Land subsidence – a significant contributorto future floodingAlthough land subsidence is not driven by climatechange, its impact on flooding in Bangkok needs tobe considered for the 2050 simulations. The teamundertook a comprehensive assessment of historicand forecasted land subsidence for Bangkok, andheld technical consultations with the concernedagencies. Historic land subsidence in Bangkok hasreduced from highs of 10 cm/year to 1 to 2 cm/yearover the period of 1978 to 2007, and from 2002 to2007 the rate had declined to 0.97 cm/year. The teampredicted that due to government efforts to controlgroundwater pumping, the average subsidence ratewould continue to decline by 10 percent/year, assumingthat current efforts to reduce groundwaterextraction continue. As a consequence, the team’sestimate of land subsidence by 2050 varied from 5 to30 cm depending on the location. Figure 3.2 showsland elevations in 2002 versus projected land elevationsin 2050 as a consequence of land subsidence.Simulation events modeledThe study ran simulations for 1-in-10, 1-in-30, and1-in-100-year flood events for 2008 as the baselinefor the inundation hazard. The study then ran simulationsapplying only the land subsidence expectedby 2050 without applying the climate change factorsfor precipitation or SLR. This was done to allow differentiationof the contribution of land subsidenceto the flood hazards and risks. Then simulation runswere made for 1-in-10, 1-in-30, and 1-in-100-yearflood events in 2050, applying either the A1FI or B1forecasted variables (precipitation and SLR). Finally,simulation runs were made for the impact of stormsurge on the 1-in-30-year A1FI and 1-in-100 A1FIand B1 simulations.Infrastructure scenarios were assumed as follows.For the 2008 base year calculations, it wasassumed that the existing and nearly completedflood protection infrastructure was in place. For thefuture 2050 calculations, it was assumed that theplanned flood protection infrastructure will havebeen implemented (Panya Consultants 2009).Conservative values were used in forecastsThe use of the mean increase in seasonal precipitationfor the Chao Phraya watershed is appropriate, as itmatches the time to concentration for the floodingin Bangkok. Mean seasonal increases were appliedto the 1-in-10-year, 1-in-30-year, and 1-in-100-yearprecipitation events and distributed across the watershedspatially and temporally using historical rainfallTable 3.3 ■ Climate Change and Land Subsidence Parameter Summary for BangkokIPCC ScenarioTemperature increase( o C)Mean SeasonalPrecipitation Increase (%) Sea Level Rise (m) Storm Surge (m) Land subsidence (m)B1 1.2 2 0.19 0.61 0.05 to 0.3A1FI 1.9 3 0.29 0.61 0.05 to 0.3Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 27

2.1.2 Land Subsidence 2.1.2 Land SubsidenceFrom observed data, From the observed average land data, subsidence the average rate land in subsidence the Bangkok rate Metropolitan in the Bangkok Region Metropolitan (BMR) Region (BMR)has gradually reduced has gradually from 10 cm reduced per year from to 101-2 cm cm per per year year to from 1-2 cm 1978 per to year 2007. from Furthermore, 1978 to 2007. Furthermore,during the last 5 years during (2002-2007), the last 5 years the average (2002-2007), land subsidence the average rate land has subsidence reduced to rate 0.97 has cm reduced per to 0.97 cm peryear. Therefore, it year. is expected Therefore, that it the is expected land subsidence that the rate land would subsidence reduce rate by 10% would per reduce year. by Thus, 10% per year. Thus,the accumulated land the accumulated subsidence during land subsidence 2002-2050 during (48 years) 2002-2050 would (48 spatially years) vary would from spatially 5 to 30 vary from 5 to 30cm depending on location cm depending as shown on location in Figure as 2.1-3. shown in Figure 2.1-3.Figure 3.2 ■ Land Elevations, 2002 versus 2050 Land Subsidence640000660000640000680000660000700000680000640000700000660000640000680000660000700000680000700000!(Pathum Thani!(Pathum Thani!(Pathum Thani!(Pathum Thani!(Nonthaburi!(Bangkok!(Samut Prakan14800001480000150000015000001520000152000015400001540000!(Nonthaburi!(Bangkok!(Samut Prakan14800001480000150000015000001520000152000015400001540000!(Nonthaburi!(Bangkok!(Samut Prakan14800001480000150000015000001520000152000015400001540000!(Nonthaburi!(Bangkok!(Samut Prakan14800001480000150000015000001520000152000015400001540000640000660000640000680000660000700000680000640000700000660000640000680000660000700000680000700000Legend!( ProvinceProvince BoundaryRiver/Canal Network-1 - 00 - 11 - 2Legend8 - 10 !( Province10 - 15 Province Boundary15 - 20 River/Canal Network-1 - 00 - 11 - 28 - 10Ê10 - 150 1.5 3 6 9 1215 - 20Legend!( ProvinceProvince BoundaryÊ-1 - 00 - 1River/Canal 0 1.5 Network 3 6 9 121 - 2Legend8 - 10 !( Province10 - 15 Province Boundary15 - 20 River/Canal Network-1 - 00 - 11 - 2Ê8 - 1010 - 150 1.5 3 6 9 1215 - 20Ê0 1.5 3 6 9 12Elevation (m.MSL)-7 - -5-5 - -3-3 - -12 - 33 - 44 - 55 - 820 Elevation - 25 (m.MSL)25 - 35-7 - -5-5 - -3-3 - -12 - 320 - 25Kilometers3 - 425 - 35Land Elevation in 20024 - 5(Surveyed 5 - 8 benchmark)Elevation (m.MSL) Kilometers-7 - -5Land Elevation in 2002-5 - -3(Surveyed benchmark)-3 - -12 - 33 - 44 - 55 - 820 Elevation - 25 (m.MSL)25 - 35-7 - -52 - 320 - 25 KilometersKilometers3 - 425 - 35Land Elevation in 2050 Land Elevation in 2050-5 - -34 - 5(Forecasted -3 - -1 land 5 - 8 subsidence) (Forecasted land subsidence)Source: Panya Consultants (2009).Source: Royal Thai Survey Source: Department Royal Thai (RTSD) Survey and Department Panya Consultants’ (RTSD) and forecast Panya Consultants’ forecastdistribution patterns (1995). There is an implicitassumption that the underlying variability (timingand location) will remain the same in 2050. This isconservative in the sense that increased variability(worse floods and droughts) around the mean wouldgenerate greater losses than estimated in the study.Main Findings fromHydrological Analysis andGIS Mapping for BangkokFlood-prone area likely to increase inBangkok in the futureFigure 2.1-3 Land Figure Elevation 2.1-3 in Land 2002 Elevation and 2050 in due 2002 to Land 2050 Subsidence due to Land Subsidence2.2 CLIMATE 2.2 CLIMATE2.2.1 General 2.2.1 General2-2remain under water for over a month. The increaseis likely to be in the western areas, where existingflood protection infrastructure is likely to be insufficient.There is also an increase in the maximumwater depth between the base year and 2050 forevents of different return periods. Figure 3.3 forexample, shows a comparison of the maximumdepths of flooding for the 1-in-30-year flood under2008 and the 2050 A1FI scenario. The implication isthat the people living in Bangkok will be facing morefrequent events that significantly disrupt daily life.2-2A flood frequency graph obtained by plottingthe inundated area versus return frequency for theBangkok City case study results for the 2008 and 2050The climate of the The Chao climate Phraya of River the Chao Basin Phraya belongs River to the Basin tropical belongs monsoon. to the tropical The average monsoon. annual The average annualrainfall over the basin rainfall is 1,130 over the mm, basin varying is 1,130 from mm, 1,000 varying to 1,600 from mm 1,000 and to registering 1,600 mm higher and registering in the higher in thenortheastern region northeastern of basin. region According of the to basin. the rainfall According pattern, to the about rainfall 85% pattern, of the average about 85% annual of the average annualrainfall occurs between rainfall May occurs and between October. May Tropical and October. cyclones Tropical occur between cyclones September occur between and September andOctober and may October strike the and basin. may In strike this the case, basin. rainfall In this continues case, rainfall for a long continues period for of time a long in period a of time in arelatively wide area. relatively The peak wide river area. discharge The peak is river registered discharge in October, is registered the end in of October, the rainy the season, end of the rainy season,and severe flood and damage severe may flood arise damage with high may tide arise in with this period. high tide The in mean this period. temperature The mean ranges temperature rangesfrom 26 o C to 31 o C. from Its 26 maximum o C to 31 o temperature C. Its maximum is in temperature April and its is minimum April and is in its December. minimum The is in December. Theevaporation (Class-A evaporation Pan) in (Class-A the basin Pan) is normally in the basin at its is highest normally in April at its and highest lowest in April in October and lowest in Octoberwith an average annual with an value average of about annual 1,700 value mm. of about 1,700 mm.The hydrological analysis shows that for events ofdifferent return periods, the area inundated willincrease in 2050 for both the B1 and the A1FI scenarios(Table 3.4).For instance, the current (2008) estimated annualinundated area in Bangkok and Samut Prakarn isabout 550 km 2 , which will increase to 734 km 2 in 2050under the A1FI scenario for a 1-in-30-year flood. Thatis an increase of 184 kms, or approximately a 30 percentincrease in the area inundated. The inundationcould be for varying depths and varying number ofdays, but about 7 percent of these provinces couldTable 3.4 ■ Bangkok Inundated Areaunder Current Conditionsand Future ScenariosFrequency2008(km 2 )Scenario2050 B1(km 2 )2050 A1FI(km 2 )1/100-year 737 893 9271-in-30-year 550 719 7341-in-10-year 359 481 518Source: Panya Consultants (2009), Appendix k, Table K2.4-1.28 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Climate Change Impact and Adaptation Study for Bangkok Metropolitan RegionFinal ReportChapter 3: Model Development and SimulationFigure 3.3 ■ Maximum Water Depth for 1-in-30-year event, 2008 and 2050, A1FI62000064000066000068000070000062000064000066000068000070000015600001560000156000015600001580000158000015800001580000!(Pathum Thani!(Pathum Thani15400001520000!(Nakhon Pathom!(Nonthaburi!(Bangkok1540000152000015400001520000!(Nakhon Pathom!(Nonthaburi!(Bangkok154000015200001500000!(Samut Sakhon!(Samut Prakan15000001500000!(Samut Sakhon!(Samut Prakan15000001480000Legend620000!( ProvinceProvince BoundaryRiver/Canal NetworkDikeElevation (m.MSL)-7 - -5-5 - -3Source: Panya Consultants’ calculationSource: Panya Consultants (2009).640000-3 - -1-1 - 00 - 11 - 22 - 33 - 46600005 - 88 - 1010 - 1515 - 2020 - 25> 254 - 5 Max. Water Depth (cm)6800000 - 1010 - 5050 - 100100 - 200200 - 300300 - 400> 400700000Ê0 2 4 8 12 16KilometersMax. Water DepthC2008-T301480000Figure 3.2-2 illustrates the maximum water surface profile along the Chao Phraya River from theA1FI river further mouth demonstrates to Bang Sai how district, the flood Ayutthaya hazard will at 10, 30 Figure and 100-year 3.4 flood ■ Bangkok return periods. Flood Comparing Hazardincrease Case dramatically C2008-T100 in and Bangkok. C2050-LS-T100, As the graph it reveals in that the maximum Relationshipwater level will be reducedFigure due 3.4 to below land illustrates, subsidence a of flood about event 0.20 in m, 2008 considering the sea level at the river mouth is not1,000with increased. an inundated Comparing area of 625 Case kmC2050-LS-T100 2 has a frequency and C2050-LS-SR-A1FI-T100,900the maximum waterof 0.02 level or will return be period increased of due 1-in-50 to the years. increasing By 2050 of flood water 800 from the upper basin ~39% and increase the sea level700Increased625underrise.theComparingA1FI scenario,CasefloodC2050-LS-SR-A1FI-T100events inundatingand C2050-LS-SR-SS-A1FI-T100, the maximum600water level will be increased at the river mouth due to storm surge,frequencybut the effect willareas of 625 km 2 would be expected to occur with an500Increased appear up toabout 50 km from the river mouth.400450 Intensityincreased frequency of 1-in-15 years. The corollary of300this is that the intensity of flooding for a 1-in-15-year200event would increase from about 450 km 2 to 625 km 2 .1001/50–year1/15–year00.00 0.02 0.04 0.06 0.08 0.10 0.121480000LegendInundated area (km 2 )620000!( ProvinceProvince BoundaryRiver/Canal NetworkDikeElevation (m.MSL)-7 - -5-5 - -3640000-3 - -1-1 - 00 - 11 - 22 - 33 - 45 - 86600008 - 1010 - 1515 - 2020 - 25> 254 - 5 Max.Water Depth (cm)6800000 - 1010 - 5050 - 100100 - 200200 - 300300 - 400> 400700000Ê0 2 4 8 12 16KilometersMax. Water DepthC2050-LS-SR-A1FI-T30Figure 3.2-1 Maximum Water Depth of Case C2008-T30 and C2050-LS-SR-A1FI-T301480000Population exposed to flooding will increasefor 1-in-30-year flood for both A1FI and B1scenariosSource: Panya Consultants (2009).Flood frequency2008 2050 A1FISevere floods inundate the lower Chao Phrayafloodplain in which Bangkok is situated. The hugevolume of water (over 31,000 million cubic metersfor a 1-in-30-year event) associated with thesefloods can take months to drain as Bangkok acts asa “bottleneck” to the discharge reaching the Gulfof Thailand. This can leave many areas flooded formore than 30 days. The flood protection schemes forBangkok are generally designed to protect againstEstimating Flood Impacts and Vulnerabilities in Coastal Cities | 29

1-in-30-year floods. There are steep increases inpersons affected as inundation levels exceed the2008 1-in-30-year design standards. For example,the number of persons who would be affected by1-in-100-year flood event in 2008 is nearly doublethe number affected by a 1-in-30-year flood, asshown in Table 3.5. Similarly, the number of personsaffected in 2050 by a 1-in-30-year event will rise sharplyfor both B1 and A1FI scenarios, with 47.2 percent and74.6 percent increases, respectively.Areas in Bangkok vulnerable to floodingAlthough Bangkok’s flood embankment andpumped drainage will continue to offer substantialprotection for the interior area, land outside the embankmentto the north and the southwest is likely toexperience significantly worse flooding. These areareas which are undergoing housing, commercial,and industrial development. Within BMR, Bangkokand Samut Prakarn are the two provinces that aremost affected by climate change. In a C2050-LS-SR-SS-A1FI-T30 scenario, almost 1 million people inBangkok and Samut Prakarn would be impactedby floods. The impact will be profound for peopleliving on the lower floors of residential buildings.Nevertheless, people living on higher floors in theBang Khun Thian district of Bangkok and the PhraSamut Chedi district of Samut Prakarn might alsobe impacted. District-wise, Don Muang district innorth Bangkok has the highest number of peopleaffected by floods (approximately 90,000) owing toits higher population density. In the western part ofthe Chao Phraya River, about 200,000 people in BangKhun Thian, Bang Bon, Bang Khae, and Nong Khamdistricts might be impacted. The maps in Figure 3.5aand 3.5b below show the differences in flooding forthe 1-in-30-year event, currently and in 2050.Table 3.5 ■ Exposure of BangkokPopulation to FloodingFrequencyPopulation affected for >30 days2008 2050 B1 2050 A1FI1-in-100 year 1,002,244 1,187,803 1,271,3061-in-30 year 546,748 805,055 954,389Source: Panya Consultants (2009).Impact of the floods on people living incondensed housingFigures 3.5a and 3.5b show the impact of floodingunder the 1-in-30-year A1FI flood in 2050 in comparisonto 2008, with an overlay of poor housingin Bangkok. It shows increased flooding in severalcondensed housing areas where the poor live. Asthe study points out, “ about 1 million inhabitantsof Bangkok and Samut Prakarn will be affected bythe A1FI climate change condition in 2050. One ineight of the affected inhabitants will be from the condensedhousing areas where most live below the poverty level.One-third of the total affected people may be subjected tomore than a half meter inundation for at least one week.This marks a two-fold increase of that vulnerablepopulation. The impact will be critical for the peopleliving in the Bang Khun Thian district of Bangkokand the Phra Samut Chedi district of Samut Prakarn”(Panya Consultants 2009).Key sectors likely to be affected by floodingin 2050 but buildings and housing mostaffectedThe most significant impacts will be felt by theresidential, commercial, and industrial sectors withmore than a million buildings being affected by a1-in-30-year event in 2050. About 300,000 buildingsin areas to the west of Bangkok—like BangKhun Thian, Bang Bon, Bang Khae, Phra Samut,and Chedi districts—will also be impacted. Forthe 1-in-30-year event, approximately 1,700 kmof roads would be exposed to flooding. The LatKrabang water supply distribution station andthe Nongkhaem solid waste transit station wouldboth be subject to 50–100 cm inundation, althoughthe main water and wastewater treatment facilitiesare protected. In addition, 127 health care facilitieswould be subject to flooding, with inundation levelsranging from 10 to 200 cm.Storm surge and sea level rise haverelatively small role in contributing toflooding in BangkokAnalysis carried out in the Bangkok study founda linear relationship between future precipitation30 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Figure 3.5a ■ Affected Condensed (Poor) Community of Case C2008-T30Figure 3.5b ■ Affected Condensed (Poor) Community of Case C2050-LS-SR-SS-A1FI-T30Source: Panya Consultants (2009).and flood volume in the Chao Phraya River. However,it found that flood peak discharge in the ChaoPhraya River will increase by a larger percentagethan precipitation, due to unequal travel times offloods from upstream catchments. Further, whilestorm surges and sea level rise are important, theywill have less effect on flooding in Bangkok. Whilestorm surges do occur in the Gulf of Thailand andcontribute to flooding the BMR area, it was estimatedthat the flood-prone area in Bangkok andSamut Prakarn will increase by about 2 percentdue to a storm surge striking the western coast ofthe Gulf of Thailand.Estimating Climate-relatedImpacts in ManilaThe Manila study (Muto et al. 2010) focuses onMetro Manila and includes Manila and 16 munici-Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 31

15°N10°NThis map was produced by the Map Design Unit of The World Bank.The boundaries, colors, denominations and any other informationshown on this map do not imply, on the part of The World BankGroup, any judgment on the legal status of any territory, or anyendorsement or acceptance of such boundaries.120°E 125°E5°N 120°E125°E5°N15°N10°Npalities. It is a low-lying area crisscrossed by thePasig River and its tributaries, which flows northwestwardfrom Laguna de Bay, the largest lake inthe Philippines, to Manila Bay in the west. MetroManila lies on a swampy isthmus and is markedby three quite diverse hydrological characteristics.These include the Pasig Mariquina area, Kamanavaarea, and the west of Managhan area facing theLaguna de Bay (Figure 3.6). The Pasig Mariquinaarea has several river systems and a catchment areaof 651 sq kms. It includes 10 cities/municipalitiesof Metro Manila. The Kamanava area along thelow-lying coast is flat, prone to typhoons, and haselevations ranging from around sea level to 2–3meters above sea level. Before the 1960s, it mainlyconsisted of lagoons, but has been filled up and currentlycomprises commercial districts, residentialareas, and fishponds. The West of Managhan areais 39 square kms and covers five cities. There area number of drainage channels draining into theLaguna Lake or Napindan River.Importance of Metro Manila to the regionaleconomyFigure 3.6 ■ Metro Manila and itsWatershedJANUARY 20100 5KILOMETERS14°40'MANILAManilaBayBacoor121°00'MeycauayanCaloocanPasig RiverMandaluyongParañaque121°00'San JoseDel MonteSan JuanQUEZONCITYPaterosTaguigMarikinaPasigR iverMarikinaLagunaBayRodriguezSan MateoSource: Philippines map IBRD 37476.AntipoloBinangonanBosobosoLagunaBayPililla121°20'PHILIPPINESPASIG – MARIKINA RIVERS121°20'MALAYSIATuguegaraoSan FernandoBaguioSan FernandoMANILA QuezonArea of MapSulu SeaPROJECT RIVERSWATERSHEDSELECTED CITIESREGION CAPITALNATIONAL CAPITALIloiloZamboangaIliganCotabatoCelebes SeaLegaspiTaclobanCebuREGION CAPITALSNATIONAL CAPITALREGION BOUNDARIESINTERNATIONALBOUNDARIESPhilippineSeaButuanCagayande OroDavaoIBRD 37476Like Bangkok, Manila is also a capital city and amajor economic center. It is located on the easternshore of Manila Bay, on the western side of theNational Capital Region. It is a central hub of thethriving Metropolitan Manila area. With a populationof 11 million, it is ranked as one of the mostdensely populated cities in the world. The PasigRiver bisects the city in the middle before draininginto Manila Bay. Metro Manila, the broader urbanagglomeration, has the highest per capita GDP inthe country. In 2007, its population was estimated tobe over 20 million people. In 2008, it was ranked asthe 40 th richest urban agglomerations in the world,with a GDP of $149 billion, indicating the economicimportance of Metro Manila. Manila is also a majortourist destination, with over 1 million tourists visitingthe city each year.Unregulated expansion into fragile areasand informalityWith large in-migration and rapid populationgrowth, the city expanded to the suburbs, surroundingmunicipalities, and to areas risky forhabitation (e.g., swampy areas, near or above esterosor water canals, along the river or earthquakefault lines, etc.). A large part of the developmentsin informal settlements are unregulated. Manystructures are built on dangerous and risky areas,such as near the seashore or flood zone, or onground prone to landslides. Socioeconomic factorslike land use practices, infrastructure development,building standards/codes and practices, and urbandevelopment policies and programs have greatlyshaped the settlement and building patterns of thecity. These forces generate an environment thatposes high risks to residents and infrastructurealike, especially in the low-lying flood-prone areas.As in other cities in developing countries, there aresignificant wealth disparities in Manila that arereflective of the country as a whole, with 97 percentof GDP controlled by 15 percent of the population(WWF 2009). Within Metro Manila, 10.4 percent ofthe population in lives below the poverty line, accordingto 2006 official statistics. 32 Approximately61 percent of the population in Metro Manila doesnot have access to basic services, according to the2008 Philippine Asset report report card (citedin Manila study). There is an estimated housingbacklog of almost 4 million.32http://www.nscb.gov.ph/poverty/2006_05mar08/table_2.asp32 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Manila’s current climate—Tropical withdistinct wet and dry seasonsFigure 3.7 shows the four main climate regimesin the Philippines and the relative frequency oftyphoons making landfall. Manila falls into Type1, which is defined as having two distinct seasons,with the dry season running from November toApril and the wet season occurring in the othermonths. About 16 percent of the typhoons crossingthe Philippines pass in the general vicinity of Manilaand its nearby areas. The average temperaturesand precipitation in Manila are shown in Table 3.6.Manila has an average annual precipitation of about1,433 mm, but there is significant variability.Main climate-related drivers of flooding—driven by single-storm events, usuallyrelated to typhoonsFigure 3.7 ■ Different ClimaticRegimes in the PhilippinesTable 3.6 ■ Manila: Monthly AverageTemperature andPrecipitationMonthTemperatureAverage Low( o C)TemperatureAverage High( o C)Averageprecipitation(mm)January 24° 30° NegligibleFebruary 24° 30° 12.1March 25° 32° 9.1April 27° 33° 15.9May 27° 33° 133.0June 26° 32° 150.8July 26° 31° 292.9August 26° 31° 305.8September 26° 31° 237.5October 26° 31° 137.2November 25° 31° 81.3December 24° 30° 58.0Total 1433.6Source: http://weather.uk.msn.com/monthly_averages.aspx?wealocations=wc:RPXX0017Unlike Bangkok, where floods are generated byheavy seasonal precipitation over 1 to 3 months,extreme flood events in Manila are caused by heavyprecipitation events over 1 to 3 days generally associatedwith typhoons and storm surge. The watershedarea for Manila’s Pasig-Marikina River System is651 km 2 , which includes the San Juan River catchment.This is relatively small in comparison to thewatersheds of other cities in the study. The precipitationpattern causing flooding is often associatedwith strong winds and flash flooding, which canbe devastating for informal housing located alongdrainage ways. Seeking refuge during such events isextremely difficult and hazardous to the population.Other climate-related factors contributing to floodinginclude a combination of high tide, excess runoff fromrivers, heavy rains, and sea level rise. A schematic ofthe watersheds around Manila is shown in Figure 3.8.Non-climate related drivers of flooding!Even though land subsidence is an important issuein Metro Manila, it was not possible to consider itfor the hydrological simulation for the Manila studybecause of lack of availability of reliable data. 33 TheSource: Muto et al. (2010).33One difficulty in Metro Manila is that the station that issupposed to measure land subsidence of Manila de Bay islocated on one of the old piers of Manila Port, which itselfis sinking gradually.Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 33

severe flooding caused by Typhoon “Ondoy” inSeptember 2009 has raised public awareness of theunderlying causes of flooding. Recent analysis pointsto mostly anthropogenic causes behind the extremeflood event, including (a) a decrease in river channelcapacity through encroachment of houses, siltationfrom deforestation, and garbage; (b) disappearance of21 km of small river channels; (c) urbanization acceleratingrunoff concentration and reducing infiltrationlosses; (d) loss of natural retention areas; and (e) landsubsidence. Among these, land subsidence is the leastunderstood but important cause. It is being drivenby groundwater pumping and possibly geologicprocesses associated with the West Marikina ValleyFault (Siringan 2009). Land subsidence continuesdecades after the groundwater pumping stops, asillustrated by the Bangkok city case study.Flood control and mitigationprojects—1990 Master PlanFlood control and mitigation projects have beenplanned and implemented in Metro Manila overthe years. 34 A key feature of flood control in Manilais the Mangahan Floodway, which was constructedin 1985 and diverts flood flows from the MarikinaRiver into Lake Laguna as a method of attenuatingpeak discharges and protecting Manila’s urban areasfrom excessive inundation. JICA supported thedevelopment of a Flood Protection Master Plan in1990, which still serves as a basis for current and futureflood protection planning. The flood protectionmeasures undertaken as part of these projects areprimarily infrastructure projects involving attenuatingpeak flood flows in Laguna de bay, improvingconveyance of flood waters through Manila to ManilaBay, and preventing overbank spillage where theflood profile is higher than the adjacent river bank.Climate change flood simulation parametersTo simulate the future climate-change-induced flooding,the city study applied the statistical downscaledresults from the University of Tokyo’s IntegratedResearch System for Sustainability Science (IRS3)for temperature and extreme precipitation increases.For SLR in 2050, the study applied approximately50 percent of the increase in sea level rise in 2100projected in the IPCC AR4. 35 Historic storm surge inManila Bay was calculated at 0.91 m during majortyphoons; for 2050, a 10 percent factor increase wasapplied based on Ibaraki University methodology, 36bringing the storm surge estimate for 2050 to 1.0 m.34See Manila study for a list of planned and ongoing floodcontrol projects and studies.35In terms of boundary conditions for the hydrologicalmodel, the SLR values were added to high tidal level; thatis, mean spring high water level in Manila Bay was set asthe base line water level for the flood simulation, to whichwas added an SLR increase in accordance with the twoIPCC scenarios.36The university provided a detailed technical methodologyfor calculating the increase in storm surge based on ananalysis of historic surges and forecasted increased typhoonintensity.Figure 3.8 ■ Major Watershed and Drainage Areas of ManilaPasig-SanJuanMarikinaWest MangahanSource: Muto et al. (2010)34 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 3.6 ■ Manila Climate Change ParametersSimulation CaseTemperature Rise (°C)(downscaled)Sea Level Rise(cm) (global)Increased Rate of Rainfall24-hr event (%)Storm Surge Height(m) at Manila Bay1 Status quo climate (no change in climate0.0 0.0 0.0 0.91dependant variables)2 B1 with no change in storm surge 1.17 19 9.4 0.913 B1 with strengthened storm surge level 1.17 19 9.4 1.004 A1FI with no change in storm surge 1.80 29 14.4 0.915 A1FI with strengthened storm surge level 1.80 29 14.4 1.00Source: Muto et al. (2010).The factors applied are summarized in Table 3.7. The24-hr precipitation factor increase is reasonable as itis close to the time of concentration of flood flows forthe Manila watershed. Similarly, SLR is conservative,since it is taken from the IPCC estimates and doesnot include polar ice cap melt.Infrastructure scenarios assumed in baseyear and in 2050Box 3.2 ■ What does Metro ManilaLook Like in the Future?For the hydrological mapping, the team used 2003 topography datafor the structure of the city. It was assumed that in terms of numberof buildings and road network, Metro Manila in 2050 would lookvery similar to what it is today. Further, while the team intendedto use future land use in the hydrological and GIS mapping, thiswas not possible due to logistical reasons. Regarding populationgrowth, Manila uses a linear extrapolation of current trends to projectpopulation to the future. The city is expected to be a megacity witha population greater than 19 million by 2050. Commercially growingsections of Manila are expected to see a decline in population,with growth occurring in other parts of the city. Areas such as thePasig-Marikina river basin are seeing a decline in population, whichis expected to continue; growth will likely continue to occur in WestMangahan and Kamanava. Future regional GDP is not projected, buta real growth rate of 5 percent is applied to a variety of economicvariables and damage costs.Source: Muto et al. (2010).For the simulation runs, the study examined twoflood infrastructure alternatives based on the 1990Master Plan. The first was to assume the implementationof the Master Plan would be halted in 2008,the base year—referred to as “existing” infrastructureand abbreviated as “ex.” The other alternativewas that the full plan would be implemented by2050—referred to as “business as usual” and abbreviatedas “BAU.” To do this and include the climatechange factors noted above, the team developed 22simulation runs (Muto et al. 2010).Assumptions made regarding populationincrease, land use, urbanization in 2050Numerous assumptions regarding what Metro Manilamight look like in 2050 compared to the baseyear were made in the study (Box 3.2).Findings from theHydrological Analysisand GIS Mapping for MetroManilaSignificant increase in area exposed toflooding by extreme events (1-in-100-yearevent) in 2050 under B1 and A1FI scenariosThe flood pattern in Manila for 2050 is likely to besignificantly affected by the extent to which the 1990Master Plan is implemented. Table 3.8 shows that thearea exposed to flooding from a 1-in-100-year eventunder current conditions will increase from 82.62km 2 to 97.63 km 2 under the 2050 A1FI 1-in-100-yearevent, an increase of 18.2 percent. This represents asignificant increase in the inundated area. With theEstimating Flood Impacts and Vulnerabilities in Coastal Cities | 35

Table 3.8 ■ Manila: Comparison of Inundated Area (km 2 ) with 1-in-100-year flood for2008 and 2050 Climate Change Scenarios with only ExistingInfrastructure and with Completion of 1990 Master PlanNo change to existing infrastructureAll 1990 Master Plan infrastructure implemented2008 (km 2 ) 2050 B1 (km 2 ) 2050 A1FI (km 2 ) 2008 (km 2 ) 2050 B1 (km 2 ) 2050 A1FI (km 2 )Pasig-Marikina 53.73 63.19 67.97 29.14 40.09 44.14West Mangahan 10.65 11.09 11.42 7.79 8.16 8.30KAMANAVA 18.24 18.24 18.24 — — —Total 82.62 92.53 97.63 36.93 48.25 52.44Relative increase 12.0% 18.2% 30.7% 42.0%Source: Muto et al. (2010).implementation of the 1990 Master Plan, infrastructurein the inundated areas will be reduced considerablyfor the current flood hazard (i.e., an inundatedarea of only 36.93 km 2 from a 1-in-100-year event in2008), but the infrastructure will not afford the levelof protection in 2050 for the return periods on whichthe designs are based. Climate change impacts underthe A1FI 1-in-100-year storm event would increasethe inundated area by 42 percent (from 36.93 km 2 to52.44 km 2 ) compared to the base year 2008. The mainpoint is that that the master plan will not offer theplanned level of protection.Increase in percentage of populationexposed to flooding under high and lowemissions scenariosAs would be expected from the significant increasein inundated area under the B1 and A1FI climatechange scenarios, the number of people affectedby the floods increases significantly as well, asillustrated in Figure 3.9. Thus, for a 1-in-100-yearflood in 2050, under the AIFI scenario more than2.5 million people are likely to be affected, assumingthat the infrastructure in 2050 is the same as inthe base year. In a scenario where the flood protectioninfrastructure under the 1990 Master Plan isimplemented, the level of protection expected by itsimplementation will not be achieved under the A1FI1-in-100-year storm event, where approximately 1.3million persons would be affected.An overlay of population density maps withGRID data on inundation shows areas of highFigure 3.9 ■ Comparison of PopulationAffected by Floodingunder Different Scenarios*3,000,0002,500,0002,000,0001,500,0001,000,000500,0000EX SQ EX B1 EX A1FI BAU SQ BAU B1 BAU A1FISource: Muto et al. (2010).* Scenarios include 1-in-100-year event (P100) in 2008 (SQ) with the 2050 B1and 2050 A1FI scenarios with only the existing infrastructure (EX) and with the fullimplementation of the 1990 Master Plan (business as usual or BAU).population density that are likely to face seriousflooding risk (Figure 3.10). These include ManilaCity, Quezon City, Pasig City, Marikina City, SanJuan, and Mandaluyong City.Areas in Metro Manila likely to be at highrisk from flooding due to extreme events in2050The hydrological analysis shows that some local governmentunits (LGUs) will be much more affectedthan others under different scenarios. As Figure 3.11shows, municipalities in the Pasig Marikina Riverbasin (namely Manila, Mandaluyong, and Marikina)and Kamanava areas (namely Malabon and Navotas)are more likely to be vulnerable to flooding comparedto those in the West Managhan area.36 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Figure 3.10 ■ Areas of High PopulationDensity and with HighRisk of Inundation underA1FI ScenarioImpact on roads, rail network, power andwater supply facilities and transportationThe hydrological analysis shows the length of roadsthat are likely to be flooded (in terms of depthand duration) under different scenarios. Thus forinstance, under both emissions scenarios—A1FIand B1, more roads are flooded under “existinginfrastructure scenario” where no additional infrastructuralimprovements are made, compared to ascenario where there is continued implementationof flood protection infrastructure in 2050 (Table 3.9).As a reference, it is useful to note that a depth ofmore than 26 cms is considered impassable for mostvehicles in Manila.The hydrological analysis also shows specificlocations where the power distribution system inManila is vulnerable. It notes that the elevated railsystem in Manila, although above the flood levels,is vulnerable to loss of power from the power utilityMerelco, and that a number of the stations wouldbe inaccessible during extreme floods. The impact ofhigh winds in these situations was not detailed. Theconsultants held discussions with the water supplyproviders who indicated that their facilities were notvulnerable to floods.Source: Muto et al. (2010).Figure 3.11 ■ Areas at High Risk from Flooding under Different Scenarios%80706050403020100CityMandaluyongManilaMarikinaQuezonSan JuanPassayMakatiPasigTaguigPaterosKalookanMalabonNavotasRegion Pasig – Marikina River Basin West Mangahan Area KAMANAVA AreaEx–SQ Ex–B1 Ex–A1F1 BAU–SQ BAU–B1 BAU–A1F1Source: Muto et al. (2010).Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 37

Table 3.9 ■ Affected Length of Road by Inundation DepthRoad Length by Inundation Depth (kms)8–20 cm 21–50 cm Above 50 cmFlood Scenario (Ex)Major Minor Major Minor Major MinorTotalStatus Quo 4.5 3.9 22.1 23.8 31.9 39.8 125.9B1 5.4 9.7 13.6 15.1 47.9 55.6 147.3A1FI 5.3 6.9 14.6 18.2 53.6 60.3 158.98–20 cmRoad Length by Inundation Depth (kms)21–50 cm Above 50 cmFlood Scenario (BAU)Major Minor Major Minor Major MinorTotalStatus Quo 3.78 4.33 6.40 10.45 7.45 13.42 45.82B1 7.24 8.15 9.54 15.73 12.07 20.82 73.55A1FI 9.45 9.05 12.62 16.28 14.97 25.63 87.99Source: Muto et al. (2010).Impact of existing flood-related events onthe poor 37In Metro Manila, there is a strong link betweenwhere the urban poor reside and their vulnerabilityto flooding. Interviews with 300 poor households(in 14 communities in three river basins) to assessthe impact of floods (conducted as part of theManila study) show that they primarily live inenvironmentally fragile, low-lying areas and/orswamps or wetlands that are highly vulnerable tostorm and tidal surges. Households interviewedhave a median monthly income of about 44 pesos/day (less than $1/day). Further, most of them livein slum or squatter settlements and do not havetenure security or access to basic services such asclean water, electricity, sanitation, and drainage.About two-thirds reported suffering regular lossesdue to typhoons, floods, and storm surges. Twentysevenpercent of the households interviewed havesubstandard latrines (dug latrines) or none at all(and use neighbor’s latrine or direct to the river/sea). Those who do have toilets complained abouttoilets overflowing and garbage being carried duringfloods contributing to waterborne diseases.About 65 percent of the households buy water fromneighbors or suppliers. During floods, householdexpenditures on water and transportation increase.Coping strategies include addition of stilts oradditional levels by households to escape risingflood waters, use of styrofoam boats, installationof pumps or “bombastic” to help drain flood water(done in some municipalities by the mayor).The survey details many of the public healthhazards associated with flooding . However, for someof the respondents, the consciousness of health risksposed by flooding is not very high. They say “Hindika naman namamatay dahil sa baha!” (You do notdie from floods or rising waters here!). For example,the risk of catching an infection like the deadly effectsof rat’s urine (i.e., leptospirosis) is not high intheir consciousness. Following Typhoon “Ondoy”in September 2009, however, there was an epidemicof leptospirosis that infected over 2,000 people, withover 160 people dying and many requiring dialysis.Estimating Climate-relatedImpacts in Ho Chi Minh City,VietnamHCMC is a tropical coastal city located on theestuary of the Saigon–Dong Nai River system ofVietnam. While not a capital city, Ho Chi Minh City37A detailed analysis of the impact of future flooding on thepoor was not undertaken as part of the Manila study andneeds to be carried out as a follow-up to this report.38 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

is also a key economic and financial center and acore part of the Southern Economic Focal region inVietnam. The city’s official population is estimatedto be about 6.4 million people. The actual currentpopulation, which includes temporary and unregisteredmigrants, is estimated to be 7–8 million. Thepopulation of HCMC has been growing at a rateof 2.4 percent per year, which is a faster rate thananywhere else in the country. The city accounted forover 23 percent of the country’s GDP in 2006 (ADB2010). Rapid economic growth—11.3 percent annuallybetween 2000 and 2007—has been the centraldriver behind the city’s expansion, as the increasingnumber and magnitude of income earning opportunitiesattract migrants from throughout Vietnam.Key economic growth sectors include industrialand service sectors, which have grown at a rate of11.9 percent and 11.1 percent annually respectivelyTable 3.10 ■ HCMC District PovertyRates, 2003Urban districtsRural districtsPoverty RatePoverty RateDistrict(%) District(%)District 1 2.4 Cu Chi 6.9District 2 4.5 Hoc Mon 6.9District 3 2.8 Binh Chanh 4.4District 4 5.6 Nha Be 12.9District 5 3.8 Can Gio 23.4District 6 5.5District 7 3.0District 8 7.4District 9 5.0District 10 3.0District 11 4.6District 12 6.3Go Vap 6.9Tan Binh 5.5Binh Thanh 5.0Phu Nhuan 3.7Thu Duc 7.8HCMC average poverty rate 5.4Source: Inter-ministerial Poverty Mapping Task Force, 2003, cited in ADB (2010).between 2000 and 2007. During the same period, agricultureand related activities have grown relativelyslowly at a rate of 4.8 percent annually (ADB 2010).Nature of poverty in HCMCIn 2006, HCMC had an overall poverty rate of 0.5percent (GSO Vietnam 2006). 38 While this rate is thelowest in the country, the absolute number of poorin the city is still high at between 30,000 and 40,000persons. In addition, the number of households livingin inadequate housing and poor environmentalconditions is likely to be much higher than the officialpoverty rate suggests. If unregistered migrantsare primarily poor, then the poverty rate would alsobe higher due to the large numbers of migrants missingfrom official statistics. Spatially disaggregateddata for the city shows that there is significant variationin poverty levels in different parts of HCMC.In general, rural districts have higher poverty levelscompared to more urban districts. The highest povertylevels are in the relatively sparsely populatedrural districts of Can Gio and Nha Be to the southof the city toward the coast. These are also the twodistricts that are the most prone to flooding in thecity (Table 3.10).Climate and precipitation in the Dong NaiRiver BasinHCMC has a tropical monsoon climate with pronouncedwet and dry season variations in precipitation.There is a large variation in monthly precipitation,but relatively constant daily temperatures withan annual average in the range of 26–27 o C and avariation between months in the range of 4° to 5 o C.As shown in Table 3.11, precipitation is highest inthe months from May to November, accounting for90 percent of the annual rainfall. Typhoon eventscan pass over HCMC. The dry season runs fromDecember to April and can cause drought conditionsregularly. Extreme droughts have occurred in1993, 1998, and 2002.38This is calculated on the basis of the expenditure typicallyrequired to meet a minimum daily calorific requirementof 1,700 Kcals. A number of qualifications need to be made,as the official figure does not include migrants.Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 39

Table 3.11 ■ Ho Chi Minh City: MonthlyAverage Temperature andPrecipitationMonthTemperatureAverage Low(oC)TemperatureAverage High(oC)Averageprecipitation(mm)January 22° 32° 5.6February 23° 33° 4.1March 25° 34° 14.2April 26° 35° 42.4May 26° 34° 138.9June 25° 33° 209.8July 25° 32° 204.7August 25° 32° 186.7September 25° 32° 178.3October 24° 31° 222.0November 23° 32° 88.9December 22° 31° 23.6Total 1313.7Source: http://weather.msn.com/monthly_averages.aspx?&wealocations=wc%3aVMXX0007&q=Ho+Chi+Minh+City%2c+VNM&setunit=Cit include seasonal monsoonal rainfall and tides.Extreme flooding occurs when tropical storms andstorm surges combine with tidal influences andmonsoon rainfall to create extreme weather conditions.Storm surge has been identified as a key driver forextreme events in HCMC. Further, warmer temperaturesin the South China Sea are expected to increasethe frequency of tropical storms and typhoons thatland in southern Vietnam. SLR is likely to have animportant influence on the inland reach of tidal floodingin HCMC. Currently 154 of the city’s 322 communesand wards have a history of regular flooding, affectingsome 971,000 people or 12 percent of theHCMC population (ADB 2010). With an extensivearea subject to regular flooding, it is not surprisingthat more extensive flooding can be induced bytides, storm surges, heavy rains on the city directlyor in the upper watershed, or a combination of thoseevents associated with a typhoon.Figure 3.12 ■ HCMC: FrequentlyFlooded Areas underCurrent ConditionsLow elevation topography contributes tofloodingMuch of HCMC is located in low-lying lands thatare prone to frequent flooding associated withheavy rains or even just high tides. As reportedin the HCMC case study (ADB 2010), about 40–45percent of land cover in HCMC is at an elevationof between 0 and 1 m. The mangrove forests of CanGio district in HCMC are a key natural resource andprovide considerable storm protection. Figure 3.12shows the areas of HCMC that flood frequently;areas colored in blue and red indicate areas floodedby tides (and hence is salty) and rain (freshwater)respectively.Main climatic factors contributing tofloodingHCMC is subject to both regular and extremeflooding. Regular floods refer to floods that occurthroughout the year on a daily and seasonal basis.Some of the main climatic factors that contribute toSource: ADB (2010).40 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Box 3.3 ■ HCMC in 2050Future land use and development projections for HCMC were guided by a number of sectoral and urban development plans and strategies. As perthe adjustment plan, the population of HCMC in 2025 is expected to be about 10 million; economic growth in HCMC is expected to increase 8 percentannually 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 by215 percent. The vision underlying the adjustment plan includes distributing a population of about 10 million in several satellite cities. It assumesgrowth in both the industrial and service sectors until 2020. In accordance with this vision, a substantial amount of land in HCMC is being reallocatedfor industrial purposes. An industrial master plan prepared by the Ministry of Industry estimates this to be around 11,000 ha by 2025. While theadjustment plan recognizes the importance of green space, it also estimates that open areas will decline by about 25 percent. The adjustment planrefers 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 twopopulation estimates, one low estimate of 12 million people, and a high estimate of 20.8 million people by 2050. The high population estimateis likely to be more realistic. Future population distributions were also estimated based on current patterns and trends and are expected to beinfluenced by the hollowing out of the city center, higher growth and increased population densities in peripheral areas, and the developmentof 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 numberof 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 climatechange 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 numberand magnitude of income earning opportunities attract migrants from throughout Vietnam. The HCMC study bases its GRDP numbers on an “AdjustmentPlan” developed by the government. The master plan, which was developed for 2025, contained projections to 2050 for GDP. Accordingly, economicgrowth in HCMC is expected to be 8.7 percent between 2011 and 2025, and 8 percent between 2026 and 2050. Industrial zones and export promotionzones are expected to increase and the service sector, including financial and tourism services, is expected to grow rapidly.Source: ADB (2010).Importance of non-climate-related factorsin contributing to floodingThe HCMC study shows that a number of nonclimate-relatedfactors also contribute to and exacerbatethe impacts of flooding. This includes domesticsolid waste that is often dumped in canals and waterways,as well as poor dredging of canals. Canalsare often choked with water weeds and garbage,so that in addition to accumulating sediment, theircapacity to drain storm waters is severely limited(ADB 2010). Further, land subsidence is also animportant problem in HCMC, but was not includedin the flood modeling simulations because like Manila,land subsidence data were not reliable. Thisis an important limitation of this study. The upperwatershed of the Saigon River and Dong Nai River,both of which drain HCMC, are well-regulated withdams and reservoirs, including the man-made DauTieng Lake and Tri An Lake, which are able to reducepeak flood flows associated with extreme precipitationevents. Given the intense management of theHCMC upper watershed, upstream rainfall is notconsidered to be a major factor affecting floodingin HCMC. 39 However, the watershed of the DongNai River basin has been subject to considerabledeforestation. Deforestation leads to increased runoff,erosion, and sediment release and exacerbatesflooding in HCMC; this needs to be addressed.Urban planning and assumptions about thecity in the futureAs with the other studies, analysis carried out forHCMC built its understanding of the future cityon the basis of a vast amount of data relating tocurrent conditions and a number of sectoral andurban development plans and strategies developedby the government about the future (Box 3.3).Briefly, the planning system in HCMC is shapedby several strategic plans, including the urban39The Dong Nai and Saigon rivers are heavily regulated byhydropower and irrigation dams. They are the main sourceof energy and water supply for HCMC. However, climatechange has not been factored into their construction or operation.Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 41

Box 3.4 ■ Overview of Downscaling and Hydrological Analysis Carried out forHCMC StudyGHG Emission Scenarios PRECIS IR3SHigh +7.5 (A2) +4.4 (A1FI)Low +1.5 (B2) +2.9 (B1)Low-resolution outputs (~2o) from ECHAM’s (European Centre for Medium Range Weather Forecast-University of Hamburg) atmosphere-ocean-coupledglobal circulation model (GCM) Version 4 were downscaled by PRECIS (Providing Regional Climates for Impacts Studies, and a dynamic modeling tooldeveloped by the UK Met Office Hadley Centre for Climate Prediction and Research, precis.metoffice.com). Two IPCC SRES greenhouse gas emissionscenarios 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 themiddle ranges. However, other GCM’s are also available and need to be included in subsequent studies. PRECIS was used because it is the regionalmodeling tool that has been most widely used globally, as well as being the tool that the UNDP National Communication Support Unit recommendsfor 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 (1994–2003) and a future decade (2050–59), which represent the year 2050, were takenas a key input for the hydrodynamic model—HydroGIS—to calculate the water and flood regime of Ho Chi Minh City. PRECIS gave the total andseasonal rainfall anomalies that are comparable, though with a slightly wider range than the JBIC/IR3S ensemble, which was used by the other citystudies in this collaborative initiative (see table below). By using PRECIS, the daily outputs as well as spatial distribution of climate variables were alsoobtained 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 ChiMinh City for the year 2050 (%change relative to baseline)Monsoon-driven seasonal rainfall years with maximum and minimumannual rainfall for each decade were selected to represent “wet” and“dry” years of the baseline and future A2 and B2 decades.Event-based rainfall extremes for the baseline and future timeslices were derived from the 1-, 3-, 5-, and 7-day maximum rainfall fromPRECIS over the HCMC urban area. Extreme rainfalls of 30- and 100-yearreturn periods were extrapolated from the log-log (power) regressionbetween 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 rainfalldata were rescaled to the observed extreme rainfalls at Tan Song Nhat Airport. The scaling factors for each extreme period were used for scaling therespective future extreme rainfall.Source: Adapted from ADB (2010).master plan prepared by the Department of Planningand Architecture, the land use plan preparedby DoNRE, and the socioeconomic developmentplan prepared by the Department of Planningand Investment. These are prepared by differentagencies under different time scales with littlecoordination between them, thus posing a difficultproblem in estimating future infrastructure andurbanization scenarios. In terms of a developmentscenario for 2050, HCMC has an urban master planthat was approved in 1998. In 2007, a study wascompleted to adjust the HCMC master plan upto 2025 (commonly referred to as the AdjustmentPlan). 40 The city case study used the Adjustment Planas the basis of the most likely development scenario for2050 (ADB 2010).In terms of land use, key trends include a decliningpopulation in central HCMC; doughnut-shapedurbanization, with higher population growth within10 kms from the center of the city; and industries beingrelocated and established in industrial zones. Agricultureand forest land (primarily made up by theCan Gio biosphere reserve) situated in the rural andsuburban areas, has declined from 64 percent in 1997to 59 percent in 2006. According to DONRE’s landuse plans, open space is expected to decline by 25percent by 2020 (ADB 2010). Rivers and canals coverclose to 15 percent of the total land use, reflecting the40Even though it is in the process of being formally approvedby the government, in the interim, it is the recognizedupdate for the 1998 Master Plan.42 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

original swamp land on which the city was settled.Roads and other transport facilities occupy around 3percent; land for industry and commerce is expandingrapidly and now represents about 5 percent ofthe city’s land use. There are 90 parks in HCMC. Inaddition to the urban master plan and the HCMCland use plans, the department of transportation hasprepared a transport master plan for HCMC until2020, which has been used for modeling impacts ofclimate change on the transport sector (ADB 2010).HCMC also has a power development plan and ahealth sector master plan, both to 2020. Similarly,a water supply master plan until 2025 is currentlyunder preparation (ADB 2010) and has been used tomodel impacts on the water sector in 2050.Flood protection infrastructures assumedfor the hydrological analysisHCMC has a planned flood protection dyke andsluice system to protect much of the city from floodingwith an estimated cost of $650 million. This is tobe implemented in phases, which will allow lessonslearned during implementation to be applied to thelater stages. The 2050 modeling has developed scenariosassuming implementation and no implementationof the proposed flood protection measures.Dynamic downscaling technique applied tomodel 2050 climate changeThe HCMC study used the SRES A2 and B2 scenariosas the enveloping case for high and low projections.Further, unlike the other cities in the study, whichused statistical downscaling, the HCMC case studyused both statistical downscaling and also a PRECISdynamic downscaling technique to model futureclimate change parameters (Box 3.4). The dynamicdownscaling approach allowed the study to addressthe complex hydrometeorological and oceanographicchanges. Specifically, PRECIS allowed simulationof mean (minimum and maximum temperatures)for the base case and A2 and B2 scenarios, simulationof daily rainfall data in the base year (averagedover 1994–2003) and for 2050 (based on averageof 2050–59). Based on this, the percent increase inrainfall in terms of average increase and increase forextreme events (1-in-30 year, 1-in-100 year, etc.) wascalculated. Thus, it was estimated that future extremerainfall under the high emissions scenario will increase bymore than 20 percent for a 1-in-30-year flood and by 30percent for a 1-in-100-year flood. This information wasthen used as input into the hydrological modeling.Inputs and outputs for the hydrologicalmodelingThe main drivers of the water regime in the HCMCinclude the following: seasonal monsoon-drivenrainfall, extreme rainfall due to typhoons and tropicalstorms in the vicinity of the city, local SLR, stormsurge, upstream-downstream inflow as a function ofcatchment basin hydrology, and various land usessuch as water management infrastructure, hydrologic/hydrodynamicconductivity, and water demandby sectors and geographic locations. Some of theseare related to regional climate change and are more afunction of land use and development. For instance,“over 75 percent of flooding points in HCMC haveoccurred following rainfall of 40mm even duringebb tide,” which indicates that surcharge from stormdrains is a major factor contributing to flooding. Forthe hydrological simulations, it was assumed that thepopulation and land use in 2050 were the same as thebase year (see HCMC annex). For the HCMC study,a hydrodynamic modeling tool—Hydro-GIS—wasused to integrate information about these driversand derive output variables (such as flood depth,duration, salinity distribution, etc) for an assessmentof risks and impacts in 2050.Estimates for sea level rise and storm surgeIn addition to using dynamic downscaling to estimateprecipitation changes, the study used the DIVA(Dynamic Interactive Vulnerability Assessment) tooldeveloped by the DINAS-COAST consortium. 41 Theresults from that tool yielded SLR of 0.26 m and 0.24 m41In DINAS-COAST, the DIVA method was applied toproduce a software tool that enables its users to producequantitative information on a range of coastal vulnerabilityindicators, for user-selected climatic and socioeconomicscenarios and adaptation policies, on national, regional andglobal scales, covering all coastal nations. http://www.pikpotsdam.de/research/research-domains/transdisciplinaryconcepts-and-methods/project-archive/favaia/divaEstimating Flood Impacts and Vulnerabilities in Coastal Cities | 43

Table 3.12 ■ Climate Change Parameter Summary for HCMCIPCC ScenarioTemperatureincrease (oC)Precipitation 24-hreventPrecipitation 3–5days Sea Level Rise (m) Storm Surge (m)Land subsidence(m)B2 +1.4 –25% Insignificant change 0.24 1.08 (from 0.53 2008) Not consideredA2 +1.4 +20% +20% 0.26 1.08 (from 0.53 2008) Not consideredfor the A2 and B2 scenarios, respectively. The studyalso notes that the city’s low topography creates asituation where a tipping point 42 of around 50 cmswould considerably increase the impact of SLR, withsignificant areas potentially becoming permanentlyflooded. For storm surge, the study assessed thatthe historic storm surge lasts about 24 hours andreached about 0.5 m. For storm surge from intensifiedtyphoons in 2050, the study examined adjustedhistoric tracks that would increase the storm surgeand estimated future storm surges associated withextreme events at 1.08 m, which would be associatedwith a track making landfall at HCMC.Flood simulations limited to regular andextreme 1-in-30-year flood events for currentand the 2050 A2 high emission scenarioIn assessing the impact of extreme flooding inHCMC, the study focused mainly on the 1-in-30-yearextreme event for 2008 and the 2050 scenario. Also,the study focused mainly on the A2 scenario becausethe team found that the extreme precipitation eventsunder the B2 scenario were not significantly differentthan 2008 events. A summary of the simulationsconducted is provided in Annex B. A significantadvantage of the dynamic downscaling techniqueis that it gives temporal and spatial information onprecipitation patterns on a daily basis, which allowsassessment of not only floods, but also droughtevents. While this was not defined as a focus of thestudy for the four cities, it is important in the contextof HCMC. Drought scenarios also were modeled.Main Findings fromHydrological Analysis andGIS Mapping for HCMCFlood hazard increases, for both regularand extreme events and number of personsexposed to flooding rises dramaticallyLarge areas of HCMC (1,083 km 2 ) 43 flood annually;extreme floods (1-in-30-year) inundate 1,335 km 2 .Similarly, much of HCMC’s population alreadyexperiences regular flooding with about 48 percentof the communities affected annually. Table3.13 shows a comparison of regular (annual) andextreme (1-in-30-year) flooding under the currentand the 2050 A2 scenario for inundated areas andcommunities affected. The analysis shows that thearea inundated increases for regular events from 54percent to 61 percent in 2050, and for extreme events42A point at which the changes or impacts rapidly increase,possibly irreversibly.431 km 2 = 100 ha.Table 3.13 ■ Summary of Flooding at Present and in 2050 with Climate ChangeNumber of communes affected(from a total of 322)Present 2050Regular flood Extreme flood Regular flood Extreme flood154 235 177 265Area of HCMC flooded (ha) 108,309 135,526 123,152 141,885% of HCMC area affected 54 68 61 71Source: ADB (2010).44 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

from 68 percent to 71 percent; that is, an increaseof 7 percent and 3 percent for regular and extremeevents respectively in 2050. This is also indicatedthrough GIS maps (Figures 3.13a and 3.13b).Figure 3.13a ■ HCMCCity Case Study:Comparison of 1-in-30-year Flood for 2008Significant increase in both flood depth andduration for regular and extreme events in2050A key finding of the hydrological analysis is thatthere is a significant increase in both depth andduration for both regular and extreme floods overcurrent levels in 2050. The average maximum flooddepth in HCMC is about 35 cm and the averagemaximum flood duration is 64 days. As a rule ofthumb, more than 50 cm of flood water is consideredthe threshold for safe operation of vehiclesand bikes. Flood duration of more than three daysat 50 cm depth is considered to cause significantdisruption in the city. Results from the hydrologicalanalysis estimate that there will be a 20 percent and43 percent increase in average maximum depth forregular and extreme events respectively, and a 21percent and 16 percent increase in average maximumflood duration during regular and extremeevents respectively in 2050. That is, the averagemaximum flood duration increases from 64 to 86days for regular events, and from 18–22 days forextreme events. Similar to past events, the depthof extreme floods in 2050 throughout the city willbe higher (72–100 cm) compared to the depth forregular flooding (35–44 cm).Figure 3.13b ■ HCMCCity Case Study:Comparison of 1-in-30-year Flood for 2050 A2ScenarioIncrease in populations at risk from floodingby regular and 1-in-30-year event in 2050Results from the hydrological modeling and GISmapping show that an increasing number andproportion of the population 44 will be affected byflooding from regular and extreme events in 2050(Table 3.13). For regular floods, about 15 percentof HCMC’s population is affected for the baselinescenario (in 2007). This increases to 49 percentwithout implementation of proposed flood controlmeasures. With the implementation of the proposedflood control measures, the percentage of affectedpopulation is reduced to 32 percent (ADB 2010).Climate change will also impact daily life in HCMCSource: ADB (2010).44As mentioned earlier, the official population of HCMCin 2007 was 6.4 million people and about 8 million if unregisteredmigrants are included. The team considered two(high and a low) population growth scenarios, but assumeda high population growth scenario for the vulnerabilityanalysis. The high case includes unregistered migrants intothe baseline; that is, it assumes current population to beabout 8 million. This seems more realistic compared to theassumption in the low-growth scenario, which does not assumeunregistered migrants currently in the city. Further,the existing rate of 2.4 percent population growth assumedin the low-case scenario is low compared to the experienceof other large cities in the region.Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 45

y 2050 under the A2 scenario. As Table 3.14 shows,currently about 26 percent of the population wouldbe affected by a 1-in-30-year event. However, thisis expected to rise to 62 percent of the population(about 12.9 million people) by 2050 under the A2scenario without implementation of the proposedflood control measures (ADB 2010). With the constructionof the flood control system, it is estimatedthat under the A2 scenario, for a 1-in-30-year flood,52 percent of the population—or about 10.8 millionpeople—will be affected. Thus with the implementationof the flood control measures, 2 million fewerpeople are likely to be affected. However, even withthe flood control measures being implemented,more than half of the projected 2050 population isstill at risk from flooding during extreme events.Districts and areas most at risk fromflooding due to 1-in-30-year event in 2050Another important finding from the hydrologicalanalysis is that the spatial distribution of floodsis also likely to change in 2050 under extremeconditions, depending on implementation of dif-Table 3.14 ■ District Population Affected by an Extreme Event in 2050Population 2007 Population 205020072050—without floodcontrol measures2050—with flood controlmeasuresDistrict(1,000)(1,000)No. % No. % No. %District 1 201 232 88 44 100 43 64 27District 2 130 1,492 85 66 1,410 95 1,160 78District 3 199 148 17 9 42 28 35 23District 4 190 125 99 52 125 100 81 64District 5 191 128 42 22 83 65 54 42District 6 249 216 34 14 193 90 26 12District 7 176 1,071 94 53 1,071 100 691 65District 8 373 575 88 24 574 100 171 30District 9 214 3,420 39 18 2,322 68 2,311 68District 10 239 172 67 28 76 44 76 44District 11 227 154 46 20 36 24 26 17District 12 307 1,583 65 21 795 50 788 50Go Vap 497 592 73 15 149 25 103 17Tan Binh 388 671 9 2 20 3 20 3Tan Phu 377 482 0 0 19 4 17 4Binh Thanh 450 623 64 14 511 82 502 81Phu Nhuan 176 146 0 0 4 3 0 0Thu Duc 356 1,433 133 37 756 53 733 51Binh Tan 447 1,557 64 14 776 50 318 20Cu Chi 310 1,804 83 27 489 27 502 28Hoc Mon 255 1,483 77 30 728 49 728 49Binh Chanh 331 1,926 282 85 1,744 91 1,601 83Nha Be 75 437 75 100 437 100 369 84Can Gio 67 393 66 99 393 100 393 100Total 6,425 20,863 1,690 26 12,851 62 10,766 52Source: ADB (2010).46 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 3.15 ■ Districts Affected by Flooding in Base Year and in 2050Present 2050Regular Extreme (Linda) Regular ExtremeCommune nameDist (means District) Area (ha)Floodedarea (ha)% floodedareaFloodedarea (ha)%floodedareaFloodedarea (ha)%floodedareaFloodedarea (ha)%floodedareaDist.1 762 42 5.56 215 28.24 54 7.13 327 42.94Dist 2 5,072 3,134 61.80 4,525 89.21 3,691 72.78 4,784 94.32Dist 3 471 0 0.00 57 12.14 0 0.00 133 28.17Dist 4 408 61 14.93 397 97.35 110 27.03 408 100.00Dist 5 432 15 3.46 183 42.28 16 3.68 282 65.17Dist 6 713 16 2.25 224 31.41 56 7.88 638 89.53Dist 7 3,554 1,068 30.05 3,178 89.42 2,040 57.39 3,552 99.94Dist 8 1,969 185 9.42 1,700 86.38 846 42.99 1,964 99.78Dist 9 11,357 7,165 63.09 7,829 68.93 7,350 64.72 8,103 71.34Dist 10 584 37 6.25 144 24.57 37 6.25 258 44.08Dist 11 508 18 3.55 70 13.77 18 3.55 120 23.67Dist 12 5,463 2,559 46.85 2,621 47.98 2,563 46.92 2,742 50.20Dist Go Vap 2,010 171 8.50 343 17.05 219 10.92 506 25.19Dist Tan Binh 2,226 0 0.00 41 1.86 0 0.00 65 2.92Dist Binh Thanh 2,094 594 28.36 1,619 77.33 942 44.99 1,718 82.02Dist Phu Nhuan 468 0 0.00 1 0.12 0 0.00 12 2.56Dist Thu Duc 4,692 1,514 32.27 2,140 45.62 1,813 38.64 2,256 48.08Dist Cu Chi 43,246 7,313 16.91 10,842 25.07 7,985 18.46 11,735 27.14Dist Hoc Mon 10,838 3,657 33.74 5,311 49.00 3,855 35.57 5,322 49.10Dist Binh Chanh 25,422 16,924 66.57 22,340 87.88 20,863 82.07 23,057 90.70Dist Nha Be 10,005 8,272 82.68 9,984 99.79 9,876 98.71 10,005 100.00Dist Can Gio 61,284 55,148 89.99 60,413 98.58 59,734 97.47 61,284 100.00Dist Tan Phu 1,575 0 0.00 0 0.00 0 0.00 63 3.98Dist Binh Tan 5,192 415 7.98 1,349 25.98 1,082 20.84 2,551 49.12HCMC TOTAL 200,346 108,309 54.06 135,526 67.65 123,152 61.47 141,885 70.82Source: ADB (2010).ferent adaptation options. As Tables 3.14 and 3.15indicate, the districts are not evenly affected due toflooding. Districts 6 and 8 and Binh Than district,for example, will experience a significant increasein the proportion of population affected, as well asan increase in the area flooded compared to otherdistricts. Some areas—such as Nha Be and CanGio—will be severely affected by extreme events,with 100 percent of its area flooded. With the floodcontrol system implemented, some districts (District9, Binh Than) will experience significant increasesin flooding, while others (Districts 1, 7, 11) will seea reduction in flooding.The poor are at greater risk from flooding;in some cases, both the poor and non-poorare at riskIn general, poorer areas in HCMC are more vulnerableto flooding. As Figures 3.14a and 3.14b show,some of the districts that are more vulnerable toflooding (dark purple in Figure 3.14b, are also theEstimating Flood Impacts and Vulnerabilities in Coastal Cities | 47

Figure 3.14a ■ HCMC Poverty Rates byDistrictrural poor living in the south of the city are directlydependent on natural resources for their livelihood.For instance, 60 percent of agricultural land wouldbe affected by saline intrusion during regularfloods. Inundation by salt water can damage cropsand reduce the productivity of agricultural land,thus impacting the livelihoods of the rural poor. Inurban areas of the city, the poor typically live alongcanals and drainage ditches—for example, NhieuLoc Thi, Nghe canal, Tan Hoa-Lo Gom canal, Doi-Te canal—or in slum areas near industries. Theseareas are among those at higher risk from flooding,in part due to underdeveloped infrastructure andpoor drainage and sanitation facilities.Economic activity is highly vulnerable toflooding now and in the futureFigure 3.14b ■ Districts Vulnerable toFloodingThe industrial sector is an important and growingpart of the HCMC economy. Climate change willaffect economic activity and industrial productionboth directly through inundation of production areas,and indirectly through inundation of essentialinfrastructure linked to strategic economic assetsand effects on the availability of key productioninputs (e.g. water, primary resources). Fifty percentof industrial zones (IZs) are at risk 45 of flooding fromextreme events with the proposed flood controlmeasures in place, and 53 percent without the systemin place (ADB 2010). An additional 20 percentof IZs will be located within 1 km of likely inundationif the proposed flood control measures are inplace (22 percent if the flood control measures arenot in place), meaning that they are likely to sufferindirect impacts. In terms of land area assigned toindustrial activities, about 67 percent of the areapotentially under industrial use in 2050 is likely tobe affected (Table 3.16).The city’s existing and planned transportnetwork is also likely to be exposed to increasedSource: ADB (2010).districts with higher poverty rates (indicated by redand orange in Figure 3.14a). However, in some ofthe areas that are most at risk—such as Can Gio andNha Be—the poor and non-poor are both at risk. The45Risk from impacts due to climate-related impacts—namely,flooding— in the study was determined by calculatingthe distance of an infrastructure component or facility fromthe flood zone that is inundated and assigning it a risk factor.Thus a distance of 0km from the flood zone shows it isinundated,

Table 3.16 ■ Effects of Flooding onFuture Land Use under2050 A2 Extreme EventFuture landuse type% future land useaffected without floodcontrol system% future land useaffected with floodcontrol systemUrban 61 49Industrial 67 63Open space 77 76Source: ADB (2010).flooding. Under the A2 scenario, a 1-in-30-yearevent would inundate 76 percent of the axis roads,58 percent of the ring roads, and nearly half of thecity’s main intersections, even with the flood dykeprotection system. Further, 60 percent of the city’swaste water treatment plants and 90 percent of thelandfill sites are at risk of flooding (Figure 3.15).The environmental consequences of flood watersinundating landfill sites and carrying pollutantsto nearby fish ponds is a potential risk that needsfurther consideration.Figure 3.15 ■ Impact on WasteManagement SectorImpact of saline intrusion on naturalresource dependent sectorsIt is estimated that the impact of flooding on agriculture,forestry, and primary industries will besevere. In 2050, it is expected that there will be asignificant increase in saline intrusion. Saline intrusionis mainly a problem during regular flooding,which is dominated by tidal influences (ADB 2010).It is estimated that close to 60 percent of agriculturallands are expected to be affected by increased salinityin 2050. Increased storm activity and sea levelrise are expected to increase the level of salinity insurface waters and also groundwater resources.However, since agriculture is of declining importancein HCMC, this has a more direct impact onthe poor and farmers directly dependent on theseresources rather than the economy as a whole. Figure3.16 shows an overlay of the 2050 A2 1-in-30-year flood on projected land use patterns. It showsthat flooding will impact broad areas and sectors,but particularly agricultural and protected areas.Figure 3.16 ■ HCMC 2050 A2 1-in-30-year Flood InundationOverlaid on ProjectedLand Use PatternsSource: ADB (2010).Source: ADB (2010).Estimating Flood Impacts and Vulnerabilities in Coastal Cities | 49

Potential increase in drought in 2050 underlow emission scenarioAn initial analysis of the data collected for thestudy shows that “the frequency of dry seasondrought in 2050 is likely to increase by the order of10 percent under the low-emission scenario withlittle change under high-emission scenario” (ADB2010). Further, the incidence of drought is likely toFigure 3.17 ■ HCMC Droughts andSalinity Intrusion in2050Source: ADB (2010).increase under the low-emission scenario in both thedry and the wet season. These assessments requiremore in-depth analysis. However, they do providean indication of the order of magnitude of changeand illustrate which scenarios are more susceptibleto drought.ConclusionTo sum up, all three megacities face considerableclimate-related risks in terms of changes in temperatureand precipitation. These climate risksare compounded by risks posed by non-climatefactors (such as land subsidence, poor drainage,and deforestation in upper watersheds), thus increasingthe likelihood of urban flooding. Underdifferent scenarios, in all three cities, there is likelyto be an increase in the area exposed to floodingand an increase in the percentage of populationaffected by extreme events. Given that all threecities are megacities with high population growthrates, the results warrant serious consideration.Despite data limitations, it is apparent that bothclimate and non-climate factors are important andneed to be considered in future adaptation efforts.While flood protection infrastructures are eitherin place or planned, the analysis shows that in allthree megacities, while these are likely to reducethe impact of flooding, they are not sufficient toprovide the expected level of protection from futureclimate events.50 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

4Assessing Damage Costsand Prioritizing AdaptationOptionsChapter 3 described the scale of physicalimpacts in terms of area, population, andsectors that are likely to be affected underdifferent climate and infrastructure scenarios inthe coastal cities. In this chapter, we discuss (a) thenature of the damages and monetary costs borneas a result of these physical changes, and (b) adaptationoptions considered by the city studies. Theoverall approach to estimating costs was discussedin chapter 2 and is further elaborated in this chapter.Bangkok: Analysis of DamageCosts Related to Floodingin 2008 and 2050In Bangkok, climate-change-induced flooding willlikely result in physical damages to buildings andhousing, income losses to individuals and firms,revenue losses to public utilities, and might alsoadversely affect public health. In estimating costs,the Bangkok study examines direct and indirecttangible damages (Table 4.1) associated with potentialfloods under 16 scenarios. Damage costs areevaluated for the base year of 2008 and for 2050. Aspreviously described, the scenarios include threeclimate change scenarios (current climate, B1 andA1FI), three levels of flood intensity (1-in-10 year,1-in-30 year, and 1-in-100 year) and additional scenariosthat include land subsidence, sea-level rise,and storm surge. Estimates for area, depth, and durationof flooding derived from the hydrological analysis wereTable 4.1 ■ Summary of DamagesAssessed in the BangkokStudyDamage Type Sector Damage MechanismDirectResidential units Damage to buildings and assetsDamagesCommercial units Damage to buildings and assetsIndustrial units Damage to buildings and assetsIndirectDamagesTransportation,public health, energy,water supplyand sanitationPopulationCommercial unitsIndustrial unitsTransportationPublic healthEnergyWater supply andsanitationSource: Panya Consultants (2009).No direct damage to theseinfrastructures expected in thefutureLoss of incomeLoss of incomeLoss of incomeLoss of revenue (negligible)Additional costs of medical careLoss of net revenueLoss of net revenuekey inputs for the damage cost assessment. In addition,the damage cost assessment makes numerous assumptionsregarding prices, population growth,GDP, and the location of major infrastructure such asroads, electric utilities and so on related to Bangkokin 2050, as discussed in chapter 2. While assessingdamage costs, prices are held constant in real terms51

over the period of analysis. These assumptions aredescribed in chapter 2. Damages costs are evaluatedin Thai Baht and converted to US dollars, using anaverage exchange rate for 2008 of 1 US$ = 33.31 THB.Bangkok likely to witness substantialdamage costs from flooding in 2050, rangingfrom $1.5 billion to $7 billion under a seriesof climate and land use scenariosTable 4.2 presents the costs incurred from damagesto buildings, income losses, health costs, and revenuelosses to public utilities for a range of scenarios.A comparison across scenarios shows that the costsare likely to significantly increase with climate andland use change. In the base case scenario (2050-LS-T10), a 1-in-10-year flood is estimated to resultin damages of 52 billion THB ($1.5 billion), but a1-in-100-year flood with climate change (2050-LS-SS-SR-A1FI) could cost THB 244 billion ($7 billion).The increase in costs from flooding are best illustratedby looking at the implications of a mediumsizedflood. A 1-in-30-year flood in 2008 (2008-T30),for instance, is estimated to result in damage costsof 35 billion THB ($1 billion). In 2050 with climatechange (2050-LS-SS-SR-A1FI-T30), a similar 1-in-30-year flood would cost 148 billion ($4.5 billion) 46 .Thus, in a future climate change (A1FI) scenario, a 1-in-30-year flood in 2050 could lead to a four-fold increasein costs to Bangkok.To illustrate the nature of different costs associatedwith floods, Figure 4.1 breaks down the costsassociated with a 1-in-30-year flood in the future(2050-LS-SS-SR-A1FI). As the figure shows, thedirect costs to buildings are the highest costs borneas a result of flooding, followed by income lossesto commercial establishments.It is useful to understand more carefully therelationship between the costs of hazards such asfloods and the probability of their occurrence. Figure4.2 plots total damage costs to Bangkok fromfloods in different scenarios against the probabilityof their occurrence. As this figure suggests, the costsof damage increase for higher intensity floods, butthis also means that there is a lower probability ofsuch floods occurring. The area beneath each floodexceedance curve represents the average total expectedflood damage cost from floods of differentFigure 4.1 ■ Damage Cost Associatedwith a 1-in-30-year Flood(C2050-LS-SR-SS-A1FI-T30)Damage costs (millions of THB)60,00050,00040,00030,00020,00010,000055,432Residence14045,945CommerceSource: Panya Consultants ( 2009).21,67112,3029,726IndustryIndirect ImpactPublic HealthDamage to Residencial BuildingDamage to Commercial BuildingDamage to Industrial BuildingIncome Loss of Daily Wage EarnerIncome Loss of CommerceIncome Loss of IndustryHealth Care CostIncome Loss of EnergyIncome Loss of Water Supply and SanitationTotal Damage Cost2,012 1,145 13EnergyDirect ImpactWater Supply& Sanitation55,432 MBaht45,94512,30214021,6719,7262,0121,14513148,386 MBahtintensity in any one year. The difference in the areasbetween different exceedance curves representsthe incremental annual damages as a result of a newclimate/land use scenario. These annualized valuesare discussed later in the chapter.Figure 4.2 ■ Loss Exceedance Curves,BangkokDamage costs (mill THB)300,000250,000200,000150,000100,00050,00000.00 0.05 0.10 0.15Probability of flood ocurrence2008 2050-LS 2050-LS-SR-B12050-LS-SR-A1FI 2050-LS-SR-SS-A1FISource: Based on calculations in Panya Consultants (2009).461 USD=33.3133 THB, which was the average exchangerate in 2008.52 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 4.2 ■ Summary of Flood and Storm Damages, Bangkok (million 2008 THB)Damaged Items 2008-T102050-LS-T102050-LS-SR-B1-T102050-LS-SR-A1FI-T102050-LS-SR-SS-A1FI-T10 2008-T302050-LS-T302050-LS-SR-B1-T302050-LS-SR-A1FI-T302050-LS-SR-SS-A1FI-T30 2008-T100Damage building 12,166 40,277 54,231 63,982 70,058 27,281 75,500 97,761 105,575 113,679 61,037 142,432 170,368 176,328 179,695 188,2152050-LS-T1002050-LS-SR-B1-T1002050-LS-SR-SS-B1-T1002050-LS-SR-A1FI-T1002050-LS-SR-SS-A1FI-T100Residence 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 88,628Commerce 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 82,147Industry 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 17,440Income/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 51,629lossDaily wageearner27 47 71 80 89 92 102 128 137 140 167 176 205 205 205 205Commerce 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 3,508Industry 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 13,158Energy 149 383 456 508 517 254 598 1,057 1,104 1,145 659 1,954 2,340 2,340 2,341 2,341Water supply &sanitation6 9 11 12 12 6 11 12 12 13 6 15 17 17 17 17Health 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 4,022Total 15,248 51,819 70,823 82,939 91,118 35,284 99,324 128,737 139,001 148,386 79,178 185,759 223,087 230,357 233,863 243,866Total 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 7,320Source: 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 intensitiesAssessing Damage Costs and Prioritizing Adaptation Options | 53

Damage costs in 2050 are largelyattributable to land subsidenceLand subsidence as a result of groundwater utilizationis a major problem in Bangkok. While there isreason to worry about climate change, the impactsof land subsidence are even bigger. For example,if we look at the impacts of a 1-in-30-year flood inTable 4.2, with current (2008) climate conditions, thecosts are about THB 35 billion ($1 billion). However,this increases to THB 99 billion ($3 billion) in 2050,just from expected land subsidence (2050-LS). Thus,there is a nearly two-fold (181 percent) increase inflooding costs between 2050 and 2008 as a result ofland subsidence.If we add climate change to this projectedchange in land subsidence and estimate damagecosts in the context of a 2050-LS-SR-SS-A1FI scenario,the costs of a 1-in-30-year flood are THB 148billion ($4.5 billion). Climate change results in afurther 49 percent increase in flood damage costsin 2050 (relative to 2050-LS). However, the bulk ofthe increase (67 percent) in flooding costs in 2050is attributable to land subsidence. Thus, in order toreduce the costs of flooding in 2050, a top priority shouldbe policies to reduce land subsidence.Damage costs vary by sectors, with morethan 75 percent of the costs attributable tobuildingsDamage to buildings dominates flood-related costsin Bangkok. In the context of a 1-in-30-year flood,in each of the five climate and land scenarios consideredfor Bangkok, 76–77 percent of total damagecosts are attributable to damage to buildings. AsTable 4.2 shows, in 2008, a 1-in-30-year flood couldcause building damages to the extent of THB 27 billion($800 million) in Bangkok. A similar 1-in-30-yearflood could cost up to THB 75 billion ($2.3 billion) in2050 after taking into account land subsidence andeconomic development. If we add climate changeimpacts, the damage costs further increase by 51percent to THB 110 billion ($3.4 billion). 47 There is asteep increase in building costs between 2008 and2050 (177 percent) simply because of land subsidence(Table 4.2). The increase in damages from climatechange is also significant, but less steep.Damage costs to transport are likely to belimitedThe Bangkok study examined different types of infrastructureand sought to understand the impact offloods on public investments. In general, new publicinfrastructure in Bangkok has taken the possibilityof flooding into account. For example, because ofthe height at which highways and the metro rail arebuilt, transportation will generally be unaffected byflooding. This does not mean that small streets inBangkok will not be flooded, but the economic coststo transport infrastructure is expected to be limited.The Bangkok study was unable to estimate the trafficdelay costs associated with flooding, and thereforethis potentially important cost is not included.There are unlikely to be direct water and sanitationinfrastructure-related costs because of theirlocation at higher than flood levels. However, thereare likely to be revenue losses to water and sanitationagencies if flood waters rise above 2 meters.Under these conditions water supply may becomedysfunctional and the agencies will likely incur revenuelosses. 48 The revenue loss to water supply andsanitation utilities is expected to more than doubleto THB 13 million ($390, 235) in 2050, given climatechange (A1FI), sea level rise, storm surge, and landsubsidence. However, the bulk of this increase incosts is attributable to land subsidence.If flooding is higher than 1 meter, the MetropolitanElectricity Authority will lose revenues tothe extent of THB 1.1 billion ($33 million) in 2050(A1FI-SL-SR-SS-T30). The damages with climatechange are twice the damages that are likely to occurin a no-climate-change scenario (2050-LS-T30).However, infrastructural damages are limited becauseenergy to Bangkok is supplied by two powerplants that are either being retired or will be retiredby 2050. Future infrastructure is expected to be47These values are all in 2008 THB and reflect simply thephysical changes that would be wrought as a result of climatechange, economic development, and land subsidence.48The average losses are calculated by multiplying watercharges by the average water used per day and scaling thisup for the duration of floods and number people affected ifflood waters were higher than 2 meters. Similarly, sanitationcosts = No. of affected people x waste generation (solid andliquid) per capita per day x treatment cost x flood duration54 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

uilt outside the cities and are not expected to beimpacted by floods.The Bangkok study estimates that health carecosts could double with climate change (Table 4.2).The costs from a 1-in-30-year flood could be aboutTHB 2 billion ($60 million) in 2050 with climatechange (2050-LS-SS-SR-A1FI), sea level rise, stormsurge, and land subsidence. This is about twicethe cost that would occur in 2050 without climatechange and only land subsidence (2050-LS). Thisincrease in health costs is primarily triggered bythe larger number of people likely to be affectedby floods.Daily wage earners will see significantincreases in income lossesClimate change will affect the population of workersliving in condensed housing areas and deliversignificant income losses to daily wage earners. AsTable 4.3 shows, losses to daily wage earners arelikely to increase by 25 percent when we consideran A1FI climate change scenario (2050-LS-SR-SS-A1FI-T30) in 2050 relative to a scenario where thereis only land subsidence (2050-LS-T30).Table 4.3 ■ Changes in Income Lossesto Wage Earners,Commerce, and IndustryChanges in income losses* from extreme weather (1/30 flood)under different scenarios2050-LS-SR-SS-Scenarios 2008-T30 2050-LS-T30 (%) A1FI-T30 (%)Daily wage— 11 25earnerCommerce — 266 45Industry — 168 38*Each column represents the percentage change in income relative to the previouscolumn.However, as expected, the overall cost to lowincomeworkers is very small relative to the estimatedcosts to industry and commerce (Table 4.2).Commercial income losses come second only tobuilding damages as a source of losses from floods.For instance, in the case of a 1-in-30 -year flood in aclimate change scenario (2050-LS-SR-SS- A1FI-T30),income losses to commerce are estimated at 21 billionTHB ($630 million).Box 4.1 ■ Examining Building Damages, Income Losses, and Health Costs inBangkokIn order to estimate the damage to buildings, the Bangkok study developed GIS maps of areas likely to be flooded. Different categories of buildingslikely to be impacted—identified using National Housing Authority data (2004)—were overlaid on these maps. Of particular concern was condensedhousing, 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. Damagecosts 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 bedamaged. Average asset values associated with residential buildings (obtained from census data) are THB 328,889 for Bangkok and THB 220,180 forSamut 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 threetypes 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 businesssurveys. This was then adjusted to reduce operational expenses. The average net commercial income was estimated to be THB 4,930 per establishmentper day. A similar accounting was used to estimate average income to industrial establishments. These average values are multiplied by the number ofdays of flooding and the number of buildings affected by floods to estimate income losses from commerce and industry. Income losses to daily wageearners 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 durationand nature of the areas affected. But few estimates are available of the likelihood of these outbreaks and what populations may be affected. In theabsence of disease and cost information, the average cost of hospitalization in Bangkok and Samut Prakhan was used as a proxy for public healthcosts. 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 theaffected population to get health damages.Assessing Damage Costs and Prioritizing Adaptation Options | 55

Damage costs constitute approximately 2–8percent of 2008 GDP for events of differentintensitiesSince Bangkok and Samut Prakarn are the twoareas that are mainly affected, flood damages arecompared to their gross regional domestic product(GRDP). Under assumptions of high emissions(2050-LS-SR-SS-A1FI), the impacts of a 1-in-30-yearflood in 2050 will amount to THB 148 billion (5percent of current GRDP). A bigger 1-in-100-yearflood would cost about 8 percent of current GRDP.49Similar sized floods would cost 1–3 percent ofGRDP under a no-climate-change scenario (2008),as indicated in Table 4.4.To carefully estimate the costs of climate change,Table 4.4 compares 2050 flood damages with climatechange (2050-LS-SR-SS-A1FI) to a scenario wherethere is land subsidence but no climate change(2050-LS). The net impact of climate change is thenthe difference in costs between these two scenarios.This amounts to THB 49 billion ($1.5 billion) in 2008values, or approximately 2 percent of 2008 GRDP.Prioritization ofAdaptation Options inBangkokAs discussed earlier, Bangkok has a long history offloods and its residents have developed numerousadaptation measures to cope with the risk of floods.A wealth of well-tested experience, information, andtechnology is already available to identify viableadaptation strategies. The Bangkok study reviewedand analyzed some of the more prominent practicesand potential adaptive interventions, such as improvingflood forecasting and preparing strategiesfor flood warnings, evacuation, transport duringfloods, and post-flood recovery; implementingprotection (e.g. construction of dikes/seawalls) orretreat strategies for coastal development; changesin coastal development and land use; and publicinformation campaigns and training exercises. Themunicipal agencies in BMR have also developedseveral plans for flood mitigation and drainageworks that include both structural and non-structuralmeasures. However, these plans do not includeclimate change considerations.The Bangkok study proposes a portfolio of adaptationoptions in the form of structural measuresthat can mitigate the impact of 1-in-30-year or 1-in-100-year floods.■■The results of the simulations indicate that thecrest elevations of dikes around Bangkok andalong both banks of the Chao Phraya River willnot be high enough to cope with flooding ofmore than a 10-year return period in the future.Moreover, the protected area to the west of theChao Phraya River has insufficient pump capacityto drain the floodwater into the Tha ChinRiver and the Gulf of Thailand. Thus, the firstset of options includes dike and pumping capacityimprovement.49In terms of GDP, the Bangkok study identifies the current2008 GDP for Bangkok and Samut Prakarn to be 2,916 billionTHB and estimates that GDP in 2050 will be 19,211 billionTHB. All damage costs are represented in 2008 THB, thus toestimate percent of GDP, 2008 GDP values were used.Table 4.4 ■ Damage Costs in Bangkok and Regional GRDPDamageCosts Million THB (2008) % of 2008 GRDPc2050-LS-SR-SS- Costs of Climatec2050-LS-SR-SS- Costs of ClimateMill THB C2008 C2050-LS A1FIChange C2008 C2050-LSA1FIChangeT 10 15,248 51,819 91,118 39,299 0.52 1.78 3.12 1.35T30 35,284 99,324 148,386 49,062 1.21 3.41 5.09 1.68T100 79,178 185,759 243,866 58,107 2.72 6.37 8.36 1.99Source: Based on estimates in Panya Consultants (2009).56 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

■■■■■■For the eastern part of Bangkok, pumping capacityto drain floodwater into the Bang Pakong Riverand the Gulf of Thailand needs to be increased. Basedon simulations, the study considers investmentsthat would increase pumping capacity from737 to 1,065 m 3 /sec. The total capacity of canalswould be improved from 607 to 1,580 m 3 /sec.In western Bangkok, there are three majorpumping stations at Khlong Phasi Charoeng,Sanam Chai, and Khun Rat Phinit Chai, whichdrain the floodwater into the Tha Chin Riverand the Gulf of Thailand with a total capacity of84 m 3 /sec. This current capacity is inadequateto cope with future climate change. Based onsimulation results, increased pumping capacitiesand canal improvements are proposed.The BMA has proposed coastal erosion protection,including the rehabilitation of mangrove forest alongthe shoreline of Bang Khun Thian. In the easternarea of the Chao Phraya River, there are plansto construct rock-pile embankments along the shorelineto protect the industrial community area fromcoastal erosion and waves. These considerationswould cost 35 and 49 billion THB to protectagainst floods of a 30- or 100-year return periodrespectively.Table 4.5 presents the investment costs requiredto protect Bangkok against a 1-in-30-year flood and a1-in-100-year flood in the context of an A1FI climatescenario. The Bangkok study estimates the totalinvestment costs of making structural adaptationsinvestments to be 35.3 billion THB ($1.06 billion)for a 30-year return flood protection project, and49.5 billion THB ($1.5 billion) for 100-year returnflood protection. The total annual operation andmaintenance cost for 30- and 100-year return floodsis estimated at 584 ($17.5 million) and 874 millionTHB ($26 million) respectively.Proposed adaptation options can reduceflooded areas by 51 percentThe maximum inundation area corresponding to a30-year return period flood with and without theproposed structural adaptation measures is presentedin Figure 4.3. The maximum inundation areaof Bangkok and Samut Prakarn will reduce withadaptation measures from 744.34 to 362.14 km 2 , adecrease of 382 km 2 or 51 percent.Table 4. 6 shows flood damage costs in Bangkokwith and without an adaptation-related infrastructureinvestment project. The difference in flooddamage costs “with” and “without” adaptationinvestments represents the benefits of the adaptationinvestment projects. The expected benefit ofthe project is the expected annual reduction in flooddamage cost. Box 4.2 describes how the annualbenefits are estimated as the incremental benefitsof having flood control projects of different returnperiods. The average annual benefits (or reductionin flood damage cost) are estimated at 4.4 and 5.9Table 4.5 ■ Investment Costs for Adaptation Projects in Bangkok (million THB)Investment Cost for 30-year Return PeriodInvestment Cost for 100-year Return PeriodYear FS & DD Civil Work Pump Total FS & DD Civil Work Pump Total1 21 21 30 302 42 42 59 593 42 42 59 594 1,981 1,981 2,405 2,4055 3,962 3,962 4,811 4,8116 5,944 3,083 9,027 7,216 5,065 12,2817 5,944 6,166 12,110 7,216 10,130 17,3478 1,981 6,166 8,148 2,405 10,130 12,536Source: Panya Consultants (2009).Notes: FS=feasibility study; DD=detailed design.Assessing Damage Costs and Prioritizing Adaptation Options | 57

Figure 4.3 ■ Maximum Inundation Area Without and With the Proposed AdaptationSource: Panya Consultants (2009).Box 4.2 ■ Expected Annual Benefits from Adaptation in BangkokAnnual Flood Damage Cost with and without the ProjectFlood damage cost300,000250,000200,000150,000100,00050,000The figure above shows flood damage costs in Bangkok with and withoutan adaption-related infrastructure investment project. A climate changescenario of A1FI is assumed.Expected annual benefits from flood control investments are basedon the probability of floods of different intensity (return period) occurring.Thus, the expected annual cost of flood damage is the area under the floodexceedance curve, which plots the probability of occurrence against costs ofdamage. Thus, the expected annual benefit from a flood protection projectis the difference in the area without and with the project.In the case of Bangkok, the expected annual benefit with aninvestment project for 30-year return period (or a flood infrastructureproject 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 periodproject is compared to the costs of the project. This allows decision makers to look at net discounted benefits and make a decision on whichinvestment to undertake.Source: Zhang and Bojo (2009).EDCB00.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10Probability of ocurrenceWithout Project Without ProjectWithout Projectfor 30-yr Return Period for 100-yr Return PeriodA58 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 4.6 ■ Flood Damage Costs With and Without a 30-year Return Period FloodProtection Project (million THB)Flood Damage CostFlood Damage Cost by ReturnReturn Period (Year) Probability of occurrence Without Project With Project Without Project With Project10 0.100 91,145 45,465 9,115 4,54720 0.050 123,308 59,337 6,165 2,96730 0.033 148,412 71,811 4,947 2,39450 0.020 148,412 148,412 2,968 2,968100 0.010 243,902 243,902 2,439 2,439Source: Panya Consultants (2009).billion THB ($132 to $177 million) for investmentsthat would protect against floods of 30-and 100-year return periods respectively. These benefits areestimated for the case of an A1FI climate scenario.The viability of these investments was determinedby estimating the net present value (NPV) ofbenefits for two scenarios. A base case considered thereal value of annual benefits to be constant throughoutthe analysis period. The second case incorporatedgrowth in the real value of infrastructure damage ; itwas assumed that damages (or benefits from floodcontrol) would grow at an average rate of 3 percentper year. Discount rates of 8 percent, 10 percent, and12 percent were used to estimate the NPV. 50Investments to reduce the impacts of a1-in-100-year flood economically viable forBangkokTable 4.7 presents the results of the NPV calculationsassuming that there is 3 percent growth in the realvalue of damage costs. The results indicate that theflood protection was economically feasible for bothfloods of 30- and 100-year return periods if the opportunitycost of capital is not more than 10 percent.From this preliminary evaluation and with theunderstanding that Thailand uses a discount rateof 8 percent for public investments, the Bangkokstudy proposes that flood infrastructure shouldbe designed to protect against a 100-year returnperiod flood as it provides a higher net return(NPV=13.4 billion THB, or $0.4 billion). Such aninvestment will be economically efficient givenan A1FI climate change scenario. However, if adiscount rate of 10 percent is applied, Bangkokshould opt for the adaptation project aimed at a30-year return period.50The period of analysis was 38 years (2012–50), of which thefirst 8 years are for studying, designing, and construction,and 30 years is the economic benefit period of the project.Table 4.7 ■ Net Present Value of Adaptation Measures to Provide Protection Againsta 1-in-30 and 1-in-100-year Flood (million THB)Designed Flood Protection Improvement Project forDescription30-Year Return Period100-Year Return PeriodDiscount Rate (%) 8 10 12 8 10 12Present Value of Costs (million Baht) 24,950 21,578 18,831 35,117 30,276 26,349Present Value of Benefits (million Baht) 36,354 25,439 18,286 48,521 33,954 24,406Net Present Value (NPV) (million Baht) 11,404 3,862 –545 13,405 3,678 –1,944Benefit-Cost Ratio (B/C Ratio) 1.46 1.18 0.97 1.38 1.12 0.93Source: Panya Consultants (2009).Assessing Damage Costs and Prioritizing Adaptation Options | 59

Analysis of Damage CostsRelated to Flooding inMetro ManilaAs discussed in the previous chapter, Manila Cityis a semi-alluvial plain formed by sediment flowsfrom four different river basins. Its drainage andlocation between Manila Bay in the west and alarge lake, Laguna de Bay to the southeast, makesit “a vast drainage basin” that is subject to frequentoverflowing of storm waters. Manila is vulnerable todifferent types of flooding, including overbanking,storage-related floods, and interior floods.The government of the Philippines is well awareof the flooding problems in Manila; a master plan forflood control infrastructure was developed over adecade ago. If this plan is implemented, the city willsee a significant decline in flood impacts. Thus, theManila study discusses the implications of climatechange in the context of the master plan as well aswhat would happen without it. Like Bangkok, theManila study examines flooding impacts in threeclimate change scenarios (no change, B1, and A1FI).Overall, the Manila study identifies impacts in termsof damage costs in relation to 18 different scenarios.In evaluating costs in dollars, an average exchangerate for 2008 of 1USD = 44.47 PHP was used.The Manila case study, like Bangkok, uses asectoral approach to estimate costs. Flood-relatedcosts are a result of (a) building damages; (b) losses topublic infrastructure and utilities; (c) income lossesto firms and residents; and (d) income losses fromtransportation blockages. It undertakes a more detailedanalysis of the transport sector and identifiesincreased costs associated with flood-related trafficdisruptions. The study assumes that the value ofvarious costs in 2050 are much higher than 2008, andthat these costs generally increase at a rate of 5 percentper year. However, in the paragraphs below, the2008 values of all costs are reported in order to enablea comparison across cities. Like the Bangkok studyand as noted in chapter 3, the Manila case studymakes a series of assumptions in valuing damagesand uses best available information to obtain costestimates. Thus, damage cost estimates are betterviewed as an indicator of damages rather than aspoint estimates of actual costs that may be incurred.Flooding in Manila can cause varyingdamages ranging from $109 million to$2.5 billion in different current and futurescenariosFlood costs as a result of various climate and infrastructurescenarios are presented in Table 4.8. Thetable presents flood damages associated with floodsof three different intensities (1/10, 1/30, and 1/100)under current no-climate-change conditions (SQ),under scenarios where flood control infrastructureis in place (MP) or not (EX), and under two climatechange scenarios (A1FI and B1). Flood-related costsrange from 5 billion PHP ($109 million)—in a scenariowhere this a 1-in-10-year flood, master planinfrastructure is in place, and there is no climatechange (10-SQ-MP)—to 112 billion PHP ($2.5 billion)in a situation where the planned infrastructureis not in place and climate change contributes to a1-in-100-year flood (100-A1FI-EX).Figure 4.4 shows the effects of different intensityfloods in different climate scenarios, given Manila’sexisting flood control infrastructure. It is useful toconsider the impact, for example, of a medium-sized1-in-30-year flood. A 1-in-30-year flood would cost$0.9 billion in a no-climate-change scenario (30-SQ-EX). These costs would increase by 72 percent to $1.5billion in the case of climate change (30-A1FI-EX).In a B1 climate change scenario, the cost wouldincrease by 55 percent to $1.4 billion.Figure 4.5 presents the loss exceedance curvesfor Manila. This shows the cost of floods increasingFigure 4.4 ■ Flood Costs under ThreeReturn Periods and TwoClimate Scenarios (PHP)3,000,000,0002,500,000,0002,000,000,0001,500,000,0001,000,000,000500,000,0000SQ EX B1 EX A1FI EX1/10 1/30 1/100Source: Based on estimations in Muto et al. (2010).60 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 4.8 ■ Flood Damage Costs in Manila (2008 PHP)Cost in 2008 in Pesos P10 SQ EX P10 SQ MP P10 B1 EX P10 B1 MP P10 A1FI-EX P10 A1FI MPDamages to Buildings Residential 785,486,988 320,880,033 842,295,372 491,606,130 595,395,243 546,225,294Commercial 8,641,501,748 610,789,400 9,658,314,207 1,611,046,487 13,750,520,244 2,326,289,074Institutional 66,863,814 20,189,999 91,707,535 37,209,916 96,826,650 37,268,296Industrial 2,890,401,496 1,173,449,757 2,414,697,965 1,461,799,749 1,756,641,760 1,346,409,219Maintenance cost Current roads 1,162,100 346,199 1,587,787 463,132 2,632,955 543,277Future roads 44,014 30,219 44,014 31,532 91,969 38,102Vehicle operating costs 8,823,186 2,628,501 12,055,195 3,516,306 19,990,575 4,124,802Travel time cost savings 33,199,847 8,380,787 45,754,992 11,655,307 71,672,669 13,646,330Loss to Firms Sales 2,816,137,180 2,704,662,851 2,961,770,824 2,822,212,152 3,044,628,088 2,881,793,868Income loss of settlers Formal settlers 32,629,500 20,444,625 49,763,250 26,401,500 49,437,000 29,098,125Informal settlers 85,652 51,072 151,620 51,072 255,892 72,352Total 15,276,335,523 4,861,853,444 16,078,142,760 6,465,993,284 19,388,093,046 7,185,508,737Total USD 343,484,797 109,317,627 361,513,243 145,386,332 435,936,694 161,564,467Cost in 2008 in Pesos P30 SQ EX P30 SQ MP P30 B1 EX P30 B1 MP P30 A1FI-EX P30 A1FI MPDamage to Buildings Residential 1,802,689,882 399,849,739 3,660,228,253 549,439,668 4,210,760,389 637,339,590Commercial 22,710,938,518 2,273,492,105 35,692,199,142 7,069,333,943 39,538,199,655 10,143,817,110Institutional 158,250,637 23,533,947 270,248,699 85,001,479 334,199,868 96,920,697Industrial 4,216,676,982 1,330,430,240 9,932,796,023 2,657,311,465 11,606,388,976 3,456,942,255Maintenance cost Current roads 5,286,655 1,102,956 6,846,841 1,937,811 7,482,737 2,313,418Future roads 244,376 244,376 302,185 302,185 329,119 329,119Vehicle operating costs 40,138,658 8,374,141 51,984,296 14,712,729 56,812,303 17,564,506Travel time cost savings 374,633,321 31,760,926 421,032,785 74,184,136 573,888,428 85,170,808Loss to Firms Sales 10,756,786,447 3,281,670,824 11,832,564,006 4,515,810,393 12,434,679,407 5,075,470,880Income loss of settlers Formal settlers 93,848,625 39,640,500 184,246,875 49,636,125 196,321,500 51,926,625Informal settlers 4,731,076 92,036 5,367,880 111,188 5,750,388 118,636Total 40,164,225,177 7,390,191,790 62,057,816,985 15,017,781,123 68,964,812,770 19,567,913,643Total USD 903,083,118 166,166,717 1,395,355,360 337,671,262 1,550,657,529 439,979,917(Continued to next page)Assessing Damage Costs and Prioritizing Adaptation Options | 61

Table 4.8 ■ Flood Damage Costs in Manila (2008 PHP)(Continued)Cost in 2008 in Pesos P100 SQ EX P100 SQ MP P100 B1 EX P100 B1 MP P100 A1FI-EX P100 A1FI MPDamage to Buildings Residential 3,688,647,788 1,045,670,772 6,022,893,816 1,326,288,039 7,517,544,912 2,101,690,472Commercial 37,699,327,245 15,298,341,749 63,871,514,594 25,506,211,401 68,021,524,157 37,713,082,264Institutional 298,785,692 158,994,559 485,447,235 173,893,911 1,874,981,233 253,765,175Industrial 8,650,623,155 5,694,313,706 16,556,719,073 5,532,356,399 17,850,618,995 9,193,023,327Maintenance costs Current roads 8,143,240 3,010,272 9,677,159 4,831,659 10,443,791 5,780,183Future roads 360,001 360,001 485,467 485,467 524,226 524,226Vehicle operating costs 50,729,576 22,855,337 62,246,103 36,684,130 68,001,872 43,885,751Travel time costs 706,986,380 277,477,558 1,082,134,984 197,675,748 1,420,426,406 340,173,579Loss to Firms Sales 13,403,412,143 6,567,976,899 14,085,687,162 7,745,705,319 14,639,854,088 8,339,388,091Income loss of settlers Formal settlers 214,933,500 67,473,375 230,942,250 95,140,125 481,092,750 105,586,875Informal settlers 6,050,968 584,668 6,881,952 1,247,540 7,089,432 2,091,824Total (PHP) 64,727,999,688 29,137,058,896 102,414,629,796 40,620,519,739 111,892,101,862 58,098,991,768Total (USD) 1,455,393,788 655,139,889 2,302,768,766 913,342,794 2,515,867,487 1,306,342,1101 USD=44.474561 PHPSource: Muto et al. (2010).62 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Figure 4.5 ■ Loss Exceedance Curvesfor Manila (PHP)Damage costs120,000,000,000100,000,000,00080,000,000,00060,000,000,00040,000,000,00020,000,000,00000.00 0.02 0.04 0.06 0.08 0.10 0.12Probability of floodsSQ EX SQ MP B1 EX B1 MP A1FI EX A1FI MPSource: Based on estimations in Muto et al. (2010).as the probability of their occurrence decreases. Aspreviously discussed, the area beneath each curverepresents the average total annual expected damagecosts from floods of all intensities (also see adaptationsection). Note that the loss exceedance curve ishighest for the A1FI emissions scenario (A1FI-EX).Climate change increases the extent of damages inall types of floods.Implementing existing Master Plansrelated to infrastructure development willsignificantly reduce flooding related costsFigure 4.6 ■ Damage Costs Associatedwith Different Scenarios(PHP)120,000,000,000100,000,000,00080,000,000,00060,000,000,00040,000,000,00020,000,000,0000SQ EX SQ MP B1 EX B1 MP A1FI EX A1FI MP1/10 1/30 1/100Source: Based on estimations in Muto et al. (2010).As Figure 4.6 shows, under each climate scenario,implementing the master plan will result in a significantreduction in costs. In an A1FI climate changescenario, for example, a 1-in-30-year flood will resultin damages to the extent of PHP 69 billion ($1.5 billion)without the master plan and PHP 19 billion($427 million) with the master plan implemented, soimplementing the master plan would reduce flooddamages by over 70 percent, given climate changeand the possibility of a 1-in-30-year flood. Thus, avery good starting point for Manila, in terms of itsresponse to climate change, would be to reconsiderand evaluate the master plans that are already onthe books.Building damages make up a major portionof flood-induced costsThe single most important contributor to totaldamage costs in each of the climate scenarios isdamage to buildings. Approximately 72 percent ofthe costs from a 1-in-30-year flood (averaged acrossall scenarios), for instance, result from damage tobuildings. Figure 4.7 shows the damage to buildingsin the case of a 1-in-30-year flood. The damagesare significant. Implementing the master planwill reduce damages significantly. In the case of a1-in-30-year flood with an A1FI emissions scenario,implementing the master plan will reduce damagesto buildings by 74 percent (relative to A1FI-EX).Income losses from climate changeassociated floods will increase by 9 to 16percentFloods will result in income or revenue losses toindividuals and firms. A 1-in-30-year flood withFigure 4.7 ■ Damages to Buildingsfrom a 1-in-30-year Flood(2008 PHP)60,000,000,00050,000,000,00040,000,000,00030,000,000,00020,000,000,00010,000,000,0000P30SQ EXP30SQ MPP30B1 EXSource: Based on estimations in Muto et al. (2010).P30B1 MPDamages to BuildingsP30A1FI EXP30A1FI MPAssessing Damage Costs and Prioritizing Adaptation Options | 63

Table 4.9 ■ Income and Revenue Losses to Individuals and Firms Associated withFloods (2008 PHP)Income Costs of Climate Change % increase in income costs fromFlood Intensity SQ EX A1FI-EX(A1FI)Climate ChangeT 10 2,848,852,332 3,094,320,980 245,468,648 9T 30 10,855,366,148 12,636,751,295 1,781,385,147 16T 100 13,624,396,611 15,128,036,270 1,503,639,659 11Source: Muto et al. (2010).an AIFI climate scenario, for instance, will resultin income losses to the extent of 12.6 billion PHP($284 million) (Table 4.9). Significant income lossesare expected for residents of parts of Manila. Thepopulation in the Pasig-Marikina River basin inManila and Malabon and Navatos in Kamanavaare most likely to be affected.Examining a with (A1FI) and without (SQ)climate change scenario, Table 4.9 shows that climatechange is likely to contribute to a 9 percentincrease in income losses given a 1-in-10-yearflood, a 16 percent increase in income losses froma 1-in-30-year flood, and an 11 percent increasein income losses if a 1-in-100-year flood occurs.More than 95 percent of the flood-related incomelosses, however, will be a result of losses borne byfirms. 51 Given a 1-in-30-year flood, firms will seea 16 percent increase in costs in a climate change(T30 A1FI-EX) relative to a no-climate-change scenario(T 30 SQ EX). Comparing the same scenarios,formal settlers or residents 52 will see an over 100percent increase in costs. Informal settlers will seetheir damage costs from flooding increase by 22percent as a result of climate change associatedwith a 1-in-30-year flood.Sectoral Impacts will VaryMost of the power stations that distribute energyto Manila will not be damaged by floods becauseof their location on high ground. However, a fewsubstations in the flood-prone areas may be shutdown for a few hours if flood waters reach theground level. This would mostly lead to some revenuelosses because of closure. In terms of waterand sanitation services, piped water supply is notexpected to be affected by flooding. Water pressurewithin pipes is expected to be strong enough to preventinfiltration by contaminated water. Pumpingstations are also above flood level. Floods will affectmany roads in Manila and will render them notpassable. A 1-in-100-year flood in a climate changescenario, for example, is likely to inundate over 30km of roads. Some parts of Manila’s rail system willbe affected by flooding, particularly if power cutsoccur during floods. Under some flooding scenarios,the railway system, LRT1 will be affected by powercuts and could be stopped.Increase in cost of flooding on roadnetworksRoad and transportation delay-related damagesfrom a 1-in-30-year flood with climate change(A1FI-EX) would be 638 million PHP ($14 million).This represents an over 52 percent increase in costsrelative to a no-climate change scenario (SQ EX).Over 90 percent of the road and transport-relatedcosts are related to time-cost delays from trafficdisruptions.51Because of lack of data, the Manila case study uses revenuesinstead of net revenue losses to firms. Thus, firm-relatedlosses are likely to be overestimated. A 2008 survey onbusiness income losses to firms is used to examine the costsof flooding to business. The average income or sales datawas obtained for firms that undertake different economicactivities such as manufacturing, construction, hotels, andso on. Affected buildings were classified according to differentuses and the sales losses from each of the buildingsobtained, assuming one firm per floor.52The income of individuals directly affected by floods isdifficult to measure. The Manila study assesses income byfirst estimating the number of households that are affectedby flooded buildings. The income loss to these householdsis then estimated using average per capita income datafrom the National Statistics Office.64 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Box 4.3 ■ Increased Time Costs and Health Risk from Flooding in ManilaThe Manila study aggregates two different estimates to establish the cost of flooding on road networks. It estimates the cost of maintenance onflood-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 offloods are categorized into vehicle operating costs and time costs. Vehicle operating costs are estimated by examining the increased operating costsas 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 timeto 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 annualrisk in gastrointestinal diseases in the areas that are likely to be flooded. The risk of gastrointestinal diseases from accidental ingestion of water rosefrom 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).Floods of different intensity can costbetween 3 percent and 24 percent of GDPFigure 4.8 ■ Flood Costs as a Percentof 2008 GDPTable 4.10 compares a “without climate change”(SQ EX) and “with climate change” (A1FI-EX) scenario,indicating that the costs of flooding can rangefrom PHP 15 billion ($337 million) (SQ-EX-10) toPHP 111 billion ($2.5 billion) (A1FI-EX-100). Sincethese numbers are presented in 2008 values, we cancompare them to Metro Manila’s regional GDP of468 billion PHP (National Statistical CoordinationBoard). Damage costs range from 3 percent of GDP(SQ-EX-10) to 24 percent (A1FI-EX-100).The costs of climate change range from PHP 4billion ($89 million) (1/10 flood) to 47 billion PHP60,000,000,00050,000,000,00040,000,000,00030,000,000,00020,000,000,00010,000,000,0000P30SQ EXP30SQ MPP30B1 EXSource: Based on estimations in Muto et al. (2010).P30B1 MPDamages to BuildingsP30A1FI EXP30A1FI MPTable 4.10 ■ Damage Costs from 1-in-10, 1-in-30, and 1-in-100-year Floods in DifferentScenarios (2008 PHP)FloodIntensity SQ EX A1FI-EXClimate ChangeDamage Costs(2008 PHP) withan A1FI Scenariowith EX1/10 15,276,335,523 19,388,093,046 4,111,757,522.591/30 40,164,225,177 68,964,812,770 28,800,587,593.151/100 64,727,999,688 111,892,101,862 47,164,102,174.61Source: Muto et al. (2010).(approximately $ 1 billion) (1/100 flood). As Figure4.8 shows, climate change costs represent 1 percent(1-in-10 flood), 6 percent (1-in-30 flood) and 10 percent(1-in-100 flood) of GDP.Prioritization of AdaptationOptions in ManilaThe Manila case study looked at a variety of adaptationoptions that could reduce the impact of 1-in-30and 1-in-100-year floods. The adaption options examinedincluded improving current practices, capacitybuilding and better coordination among local governmentand national flood management agenciesand structural measures such as dam construction,Assessing Damage Costs and Prioritizing Adaptation Options | 65

aising dikes, and improved pumping capacity thatwould reduce and/or eliminate the impacts of floods.Manila targets adaptation options thatwould almost eliminate floods in the Pasig-Marikina River basinSeveral structural measures are examined in orderto analyze the economic implications of these investments.The Manila study focused on analyzingadaptation options that could eliminate floods to theextent possible. 53 These include:■■■■Pasig-Marikina River: In this area, in order toprevent overbanking and flooding from theriver, investments in raising the embankmentare considered. Embankment raising is consideredboth with the possibility of buildingthe Marikina Dam and without. The MarikinaDam was initially proposed as part of the 1990master plan and would be able to prevent a1-in-100-year flood.West of Mangahan and KAMANAVA area: Inthese areas, improved pumping capacity andstorm surge barriers are required to reducefloods. The investment required for installingpump capacity to control flooding under anallowable inundation depth of 30 cm is considered.For preventing overflows from Manila Baycaused by typhoons, the study looks at optionsfor increasing the height of storm surge barriersin coastal areas.The Manila study examines adaptation optionsamong a number of different scenarios involvingdifferent types of construction in different areas. Table4. 11 identifies the options that were consideredunder each scenario. Notably, while the Bangkokstudy looks at adaptation only in the context of anA1FI scenario, the Manila case study looks at adaptationoptions in the context of no-climate-change(SQ), B1 and A1FI scenarios. The Manila study alsoexplores scenarios in which a major investment inconstructing the Marikina dam is undertaken (wD)and cases where the dam is not constructed (nD).The costs of adaptation are also considered in thecontext of whether the existing master plan (MP) isimplemented or not (EX).Table 4.12 presents the estimated initial costs ofeach adaptation option considered. 54 In this analysis,the construction period considered is 5 years startingfrom 2010 and the project’s life is expected tolast until 2060. Investments after the completion ofthe master plan are assumed to start from 2014. Nomaintenance cost was included.The incremental benefits of moving fromthe current situation to full adaptation andfrom implementing the master plan andthen moving to full adaptation consideredThe investment costs (assumed to occur over a5-year period) are compared with the incrementalexpected annual benefits from flood reduction associatedwith a 1-in-10-year, 1-in-30-year, and 1-in-100-year flood protection projects. The incrementalbenefits emerge from adaptation investments thatwould take Metro Manila from its:■■■■existing infrastructure (EX) to full adaptationlevel (difference between EX and full adaptationin Figure)1990 master plan level (MP) to full adaptationlevel (difference between MP and full adaptationin Figure)The annual benefits are identified in Figure 4.9.The area between the existing infrastructure curveand full adaptation level are the annual benefitsfrom making investments in the current situationwhere the master plan has not been implemented.The area between the master plan curve and the fulladaptation level refers to the annual benefits fromgoing beyond the master plan.In the cost benefit analysis of the differentinvestments, flood control benefits are assumed togrow at an annual rate of 5 percent. The net presentvalue (NPV) of benefits was obtained using a dis-53For KAMANAVA and West of Mangahan areas, totalelimination is not possible because of their low height.Rather, pumping capacity improvement is considered tominimize the duration of the flooding.54For each scenario only certain types of adaptation investmentsare considered. For example, in a B1EXwD scenario,there is no investment considered for a 1/10 flood since thisscenario already includes a dam, which is expected provideprotection for a 1/10 flood.66 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 4.11 ■ Adaptation Investments Considered for Different Return Periods andClimate ScenariosClimateScenarioAdaptation Investment Options/Costs Considered1/10 flood 1/30 flood 1/100 floodSQ SQ EX wD Pasig Marikina River embankment + Marikina Dam Pasig Marikina River embankment+Marikina DamSQ EX nD Storm surge barrier Pasig Marikina River embankmentSQ MP wDMarikina Dam in addition to investments alreadyconsidered under the MPSQ MP nD0 costs because additional flood prevention investmentsare already incorporated under the MPB1 B1 EX wD Pasig Marikina River Embankment + Additionalembankment in Marikina River + Marikina Dam +storm surge barrierB1 EX nDStorm surge barrier + Pumpcapacity improvement (KA-MANAVA +West of Mangahan)Pasig Marikina River embankment +additional embankmentin Marikina River + storm surge barrierB1 MP wD Additional embankment in Marikina River +Marikina DamB1 MP nDAdditional embankment in Marikina riverA1FI A1FI-EX wD Pasig Marikina River embankment +additionalembankment in Marikina River + Marikina Dam +storm surge barrierA1FI-EX nDStorm surge barrier + Pumpcapacity improvement (KA-MANAVA +West of Mangahan)Pasig Marikina River embankment +additionalembankment in Pasig and Marikina River + stormsurge barrierA1FI MPwD Additional embankment in Marikina River +Marikina DamA1FI MP nDAdditional embankment in Pasig and Marikina River0 costs because flood prevention costs,including Marikina Dam, area alreadyincorporated into the MPPasig Marikina River embankment +Marikina Dam +additional embankmentin Pasig and Marikina River basinAdditional embankment in Pasig andMarikina River basinPasig Marikina River embankment +Marikina Dam +additional embankmentin Pasig and Marikina River basinAdditional embankment in Pasig andMarikina River basinTable 4.12 ■ Investment Costs and Net Present Value of Benefits Associated withDifferent Flood Control Projects in Manila (PHP) using a 15 percentdiscount rate1/10 Flood 1/30 Flood 1/100 FloodClimate Scenario Investment Cost NPV Investment Cost NPV Investment Cost NPVSQ SQ EX wD NA 14,121,102,133 809,000,801 13,501,553,721 3,763,340,285SQ EX nD 42,887,291 209,952,438 10,943,489,020 2,939,371,200 NASQ MP wD NA 3,177,613,113 1,199,401,356 0.00 4,755,005,784SQ MP nD NA 0.00 3,329,771,755 NAB1 B1 EX wD NA 14,232,087,722 5,634,294,466 13,604,450,310 10,150,493,344B1 EX nD 1,003,222,253 (349,951,115) 11,054,474,609 7,764,664,865 NAB1 MP wD NA 3,216,390,949 3,512,076,308 102,896,589 7,587,409,494B1 MP nD NA 38,777,837 5,642,446,707 NA(Continued to next page)Assessing Damage Costs and Prioritizing Adaptation Options | 67

Table 4.12 ■ Investment Costs and Net Present Value of Benefits Associated withDifferent Flood Control Projects in Manila (PHP) using a 15 percentdiscount rate (Continued)1/10 Flood 1/30 Flood 1/100 FloodClimate Scenario Investment Cost NPV Investment Cost NPV Investment Cost NPVA1FI A1FI-EX wD NA 14,248,304,696 7,092,797,122 13,640,673,269 12,011,488,435A1FI-EX nD 1,409,166,226 (581,704,127) 11,099,925,438 9,203,568,238 NAA1FI MP wD NA 3,216,390,949 5,509,802,569 139,119,548 10,411,715,022A1FI MP nD NA 68,011,692 7,620,573,684 NASource: 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 isassumed that these costs are incorporated in the master plan.count rate of 15 percent, which is what is used by thePhilippines National Economic and Developmentauthority to estimate project feasibility. 55Dam construction emerges as aneconomically viable option in both climatechange scenarios, followed by embankmentbuilding in the Pasig-Marikina basinTable 4.12 presents the present value of net benefitsfrom different investments considered. The NPVat12 billion PHP ($269 million) is highest among allthe different scenarios. This suggests that constructingthe Marikina Dam would maximize benefitsrelative to other adaptation options in the case ofan A1FI or B1 scenario. The investments that wouldbe required in this scenario are building the Pasig-Marikina River basin embankment, the MarikinaDam, and some additional embankments along thePasig and Marikina rivers. These investments wouldlargely eliminate floods in this part of Metro Manila.However, given that constructing the dam is adecision that may or may not be taken, it is usefulto consider what alternative options emerge. In thiscase, in an A1FI scenario it is recommended that thePasig Marikina River embankment be built withsome additional components along the Pasig andMarikina rivers, and that storm surge barriers beconstructed. This basically means that investmentsunder the current master plan in the Pasig-MarikinaRiver basin should be prioritized and continued andsome additional investments need to be made forFigure 4.9 ■ Annual Benefits fromAdaptation Investments inMetro ManilaDamage cost (PHP in million)120,000100,00080,00060,00040,00020,00000.00 0.02 0.04 0.06 0.08 0.10ProbabilityExisting infrastructure(EX)1990 master plan level(MP)Source: Based on estimations in Muto et al. (2010).Full adaptationlevelfull adaptation to an A1FI climate. This is what iscurrently being undertaken by the government ofthe Philippines in implementing the Pasig-MarikinaFlood Control Project Phase II to avoid damages fromP30 floods. The recommendations in the context ofa B1 climate change scenario are similar. In sum,the first priority is to control flooding in the Pasig-55It is important to note that the total avoided damages(gross) in chapter 3 do not directly feed into the NPV calculations.This is because there are three geographical areas,and depending on the return period and adaptation investments,only the relevant benefits are considered for eachscenario. (For example, the embankment for Pasig-MarikinaRiver does not affect KAMANAVA coastal area and thereis thus no benefit in that area).68 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Marikina River areas through a dam, embankments,and storm surge barriers. While adaptation to climatechange would require significant investments, in thecase of Metro Manila, these are investments that arealready planned. Thus, climate change adaptationwould require the government to commit to implementingplans that are currently on the books.Analysis of Damage Costs inHCMCHCMC has a long history of extreme weather events.Between 1997 and 2007, almost all of the districts ofHo Chi Minh City have been directly affected bynatural disasters to some extent. The total value ofdamage to property from natural disasters over thelast 10 years is estimated at over $12.6 million (202billion VND). Most impacts have been concentratedin the predominantly rural Can Gio and Nha Bedistricts. However, with increased levels of floodingand extreme events due to climate change, urbanareas are likely to suffer increasing levels of damage.This is likely to increase costs significantly.The HCMC study used a different approachfrom Bangkok and Manila to estimate the aggregatecosts of climate change impacts to 2050. It does nottake a sectoral approach and estimate in detail thecosts that will be incurred as a result of climatechange. While areas of vulnerability and risks areclear, there has been no attempt to value specificrisks directly. Rather, a first approximation of thelikely costs of climate change is undertaken basedon macro data. Further, the HCMC study estimatesthe present value of flooding from the current periodup to 2050. Thus, it presents cost estimates todaythat reflect repeated flooding over a long period oftime. In contrast, the Manila and Bangkok studiesestimate the costs of specific single events of floodingin different scenarios.As discussed in chapter 2, the HCMC studyuses two approaches to estimate the costs of climatechange: (1) cost estimates based on expected lostland values; and (2) cost estimates based on aggregateGDP loss. To recap briefly, the land valuemethod estimates how climate change may affectthe value of the land stock in HCMC. The first stepin the analysis was to determine land prices. Thearea subject to flooding both in extreme eventsand regular flooding was then determined underfuture scenarios using HydroGIS modeling. Onceaverage land price and flooding extent and durationwere estimated, the relationship between theflooding and the decline in the economic value ofthe flooded land was determined to calculate thecost of flooding due to climate change. The valueof land affected by flooding is assessed assuminglinear and quadratic relationships between floodduration and land values. The second approachused by HCMC is to calculate costs on the basis ofexpected losses in production (proxied by GDP) dueto climate-change-induced flooding. For each districtand for each year (2006 to 2050), the annual costof flooding was calculated and the discounted costssummed for the whole period to find the presentvalue of expected lost GDP. In the following pagesand the rest of this report, an average exchange ratefor 2008 of 1 USD = 16302.25 VND is used to convertVietnamese dong to U.S. dollars.Macro level analyses suggest that there willbe significant costs as a result of climatechangeThe study estimates that the losses in the economicvalue of land affected by 2050 climate change wouldrange from VND 100,358 billion ($6.15 billion) forregular flooding to VND 6,905 billion ($0.42 billion)for extreme flooding, assuming a quadraticrelationship between flood duration and land value(Table 4.13). 56 The HCMC study also reports climatechange costs assuming a linear relationshipbetween flood duration and land values. In thiscontext, the cost of climate change is estimated tobe VND 369,377 billion ($22.7 billion) for regularfloods and VND 111,678 billion ($6.9 billion) forextreme events. The highest costs (using either method)are borne by Binh Chanh district, district 9, Can Gio,and Nha Be district.56In all estimates, the costs of extreme events are muchsmaller than those of regular flooding because (a) floodingdue to extreme events lasts for a smaller number of daysthan that for regular flooding; and (b) the calculation offlooding assumes a return period of 30 years for extremeflooding. The expected value in any given year is therefore1/30th of the cost of an extreme event.Assessing Damage Costs and Prioritizing Adaptation Options | 69

Uncertainty in the estimation of costsThese land-based estimates represent a first approximationof potential losses. There remainsconsiderable uncertainty as to how land valuesmay be impacted by increases in the frequency orduration of flooding. To the extent that the impactsof existing flood events (climate variability) havealready been capitalized in the price of land, theabove results should be interpreted as the possiblechanges in land values resulting from additional(and as of now unexpected) days of flooding resultingfrom climate change.Using the GDP loss estimation method (Table4.14), the cost of regular flooding in terms of GDPloss is estimated to be about VND 806,831 billion($49.5 billion) in present value terms. The cost ofextreme flooding is estimated to be approximatelyVND 7,978 billion ($0.49 billion).Table 4.13 ■ Expected Cost of Flooding based on Quadratic Relationship betweenDuration of Flooding and Land Values in HCMCFlooded area (ha) Average duration (days) Land valueExpected cost (billion VND)District Regular Extreme Regular Extreme (1,000VND/sq.m) Regular Extreme1 34 249 81 12 22,410 380 602 3,036 4,115 127 22 1,556 5,719 2333 0 105 0 0 17,407 0 04 78 348 21 6 7,369 19 75 12 243 95 10 12,946 103 246 45 594 4 2 8,508 0 27 1,004 2,451 52 9 4,070 830 618 655 1,768 12 4 4,318 31 99 5,877 6,696 140 28 1,621 14,014 63910 30 198 88 11 10,759 189 1911 14 99 73 11 7,245 40 712 2,241 2,465 140 29 2,043 6,735 318Go Vap 184 455 65 12 4,072 237 20Binh Thanh 685 1,383 62 10 10,127 2,001 105Phu Nhuan 0 9 0 0 9,234 0 0Thu Duc 1,593 1,913 95 21 3,995 4,312 253Cu Chi 7,234 10,875 117 26 717 5,330 396Hoc Mon 3,538 5,040 155 35 1,237 7,892 573Nha Be 7,948 8,421 95 20 1,778 9,572 450Can Gio 46,435 48,486 88 20 423 11,417 616Tan Phu 0 45 0 0 4,796 0 0Tan Binh 0 58 0 1 4,934 0 0Binh Tan 917 2,300 29 12 1,932 112 48Binh Chanh 18,890 22,057 83 24 3,217 31,423 3,068Total 100,450 120,372 VND 100,358 6,905Total (USD) 6.156083 0.4235612Source: ADB (2010). Some differences due to exchange rate variations.1 USD=16302.2570 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Table 4.14 ■ Present Value of the Cost of Floods up to 2050 using the GDPEstimation MethodDistrictRegular FloodingGDP Loss (VND)Extreme FloodingGDP Loss (VND)TotalGDP Loss (VND)District 1 4,262,678,022,700 118,930,920,400 4,381,608,943,100District 2 91,872,408,828,200 684,250,653,100 92,556,659,481,300District 3 — 1,045,570,300 1,045,570,300District 4 2,111,546,363,400 79,830,492,300 2,191,376,855,700District 5 2,036,113,916,300 63,099,964,700 2,099,213,881,000District 6 102,479,383,000 23,505,283,400 125,984,666,400District 7 23,901,066,736,700 319,603,147,100 24,220,669,883,800District 8 3,089,545,724,600 162,117,638,100 3,251,663,362,700District 9 203,476,562,493,300 1,573,188,396,600 205,049,750,889,900District 10 5,608,083,746,600 104,469,939,500 5,712,553,686,100District 11 1,631,296,447,400 36,684,956,800 1,667,981,404,200District 12 107,560,392,251,700 816,151,762,100 108,376,544,013,800District Go Vap 13,980,278,430,700 30,422,891,000 14,010,701,321,700District Tan Binh — 1,135,984,100 1,135,984,100District Binh Thanh 43,871,270,790,100 414,883,343,800 44,286,154,133,900District Phu Nhuan — 2,004,000 2,004,000District Thu Duc 59,101,002,606,400 533,992,286,200 59,634,994,892,600District Cu Chi 28,955,319,716,000 365,741,723,400 29,321,061,439,400District Hoc Mon 44,762,782,087,100 565,835,862,900 45,328,617,950,000District Binh Chanh 97,469,697,076,900 1,276,837,444,400 98,746,534,521,300District Nha Be 42,592,220,679,200 331,433,560,700 42,923,654,239,900District Can Gio 24,169,804,222,200 206,645,566,200 24,376,449,788,400District Tan Phu — 23,680,600 23,680,600District Binh Tan 6,276,999,023,500 268,297,958,200 6,545,296,981,700Total 806,831,548,546,000 7,978,131,029,900 814,809,679,575,900USD 49,492,036,286 489,388,338 49,981,424,624Source: ADB (2010). Some differences due to exchange rate variations.1 USD=16302.25Table 4.15 presents a summary of the costestimates from HCMC based on different methodologies.In summary, a first approximation of thecosts of climate change for regular flooding eventsis between $6.15 billion and $49.5 billion in presentvalue terms, and between $0.42 billion and $6.9billion for extreme flooding. 57The results of the two damage valuations showconsiderable divergence, with GDP figures suggestingdouble the damage costs achieved usingthe land value methodology. This may result fromtwo different factors. First, it is well known that administrativelydetermined land prices (as opposed57The HCMC numbers are much higher than the damagecosts associated with flooding in Manila and Bangkok.However, as previously noted, these numbers reflect thesum of a series of annual damages that are expected to occurfrom now up to 2050 and are not directly comparablewith the damage costs of one event in Bangkok or Manila.Assessing Damage Costs and Prioritizing Adaptation Options | 71

Table 4.15 ■ Summary of Present Valueof Climate Change Costsin HCMC (USD)Regular Flooding Extreme Flooding TotalLand (linear) 22,658,038,001 6,850,465,427 29,508,503,427Land6,156,082,749 423,561,165 6,579,643,914(quadratic)GDP 49,492,036,285 489,388,337 49,981,424,623Source: ADB (2010). Some differences due to exchange rate variations.1 USD=16302.25to market prices for land) undervalue land in thecity. On this basis alone, had market prices for landvalues been used instead of administratively determinedland prices (which were effectively used), theestimated cost of climate change using land valueswould have been higher than presented earlier.Second, GDP per capita figures at the district levelwere unavailable. This may overestimate the GDPloss across the city as rural districts in the southsuch as Can Gio and Nha Be are responsible forvery little GDP production, and yet which may bemore vulnerable to climate change. Further detailedinvestigations and data collection would be requiredto refine these figures.Box 4.4 ■ Rough Estimate ofViability of Proposed Flood ControlMeasuresHCMC has already proposed to build a system of flood defensesthat will significantly alter the pattern of flooding and the hydrologyof the city. This project has an estimated capital cost of $750million and is expected to be completed by 2025. The effects of theproposed flood control measures on regular flooding would seemto be significant in reducing the population exposed to flooding in2050 from an estimated 10.2 million to 6.7 million, a reduction of 35percent. Based on the GDP method of estimation, the project wouldproportionately decrease damage costs by 35 percent, making thisan economically viable project. However, further economic analysesof this project needs to be carefully undertaken.Source: ADB (2010).Analysis of Adaptation inHCMCUnlike the Bangkok and Manila studies, the HCMCstudy did not undertake a detailed cost benefitanalysis to prioritize adaptation options. Whileit did provide an estimate of the viability of thegovernment’s flood protection project (Box 4.4), themain focus was to identify institutional mandateswith respect to urban planning, flood protection,and adaptation and propose a range of adaptationoptions that need to be undertaken in coordinationwith different sectors.Broadly, adaptation measures in the context ofmanaging floods can be categorized as those thatinvolve (a) protection against predicted climatechange, (b) accommodation to improve resilience,(c) retreat to reduce exposure, and (d) improvedmanagement (see Annex C). 58 For instance, infrastructuralmeasures such as construction of floodembankments, polders, sea walls, and pumpeddrainage are common engineering solutions andare being implemented in all three cities discussedin this report. “Accommodation” measures thatseek to minimize vulnerability include measuressuch as raising houses on stilts, adjusting croppingpatterns, and revising building codes for housingand industry. In some cases, where the risks of lossof life or assets is severe, “retreat” as a planning optioncan reduce exposure to extreme events. Usedin conjunction with restoring natural ecosystemsthat provide flood protection benefits, it can be auseful planning tool. Finally, flooding needs to beconsidered in the context of overall water basinmanagement and the institutional capacity to managethe resource.Developing sector specific adaptationoptions with focus on the poorIn the HCMC study, a combination of these measuresis proposed. Specifically, the study argues fordevelopment of sector specific adaptation optionswith a focus on the poor. For instance, in terms58Drawn from IPCC, AR4, chapter 6, Coastal systems andlow-lying areas, Figure 6.11 Evolution of planned coastaladaptation measures, pg. 342, 2007.72 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

of reducing vulnerability of the poor, the studyrecommends livelihood protection and livelihooddiversification schemes, improved early warningsystems, improved building construction requirements,land use planning and use of open spacefor flood management, zoning controls to ensurethat low income housing is located outside offlood prone zones. HCMC already has a plannedprogram of resettlement away from rivers anddrains and this may require expansion to includevulnerable areas (Can Gio and Nha Be districts)identified by the HCMC analysis. Priority adaptationactions in the transport sector include reviewand revision of design standards for roads, bridges,and embankments so they are consistent with expectedflooding and climate conditions. Further,in light of the findings of the study, new transportinfrastructure needs to be reassessed. Given thatmost industrial zones and clusters in HCMC willbe at direct risk of flooding with or without theproposed flood control plan, a number of adaptationmeasures—such as locating industrial zonesoutside of vulnerable areas and retrofitting existinginfrastructure—are proposed.Combining infrastructure based solutionswith eco-system based adaptation measuresAn important recommendation of the HCMC studyis to combine infrastructure-based solutions withthe use of ecosystem-based adaptation measures,which provide a buffer against climate risks. Themangrove forests to the south of HCMC providesignificant protection against storm surges, but areunder severe pressure due to land use change andencroachment. Natural systems of the Dong NaiRiver basin provide a range of ecosystem servicessuch as regulation of hydrological flows, freshwaterstorage, erosion control, and water purification.These too are under threat due to land use changeand urban development. Adaptation approachessuggested by the study include reforestation ofthe Dong Nai River basin watershed, restorationof wetlands, rehabilitation of canals and rivers,strengthening zoning regulations to protect ecosystemresilience, and planting of buffer zones alongdykes and riverbanks, including dykes proposedby the planned flood control system.Strengthening institutional capacity toadapt to climate related risksThe HCMC study argues that comprehensive adaptationplanning is needed to provide the overallframework and direction for adaptation at the citylevel, in line with national level goals and targets.A key institutional recommendation arising fromthe study is the development of an HCMC ClimateChange Adaptation Plan. The HCMC Peoples Committee(PC), the study argues, can take a proactiveleadership role in this context and in driving climatechange adaptation in the city. Further, there are severalkey agencies that shape overall land use, spatialzoning, environmental quality, and natural disasterresponse management in the city, each of whichcan pursue a range of actions within their sectoraldomain (Table 4.16). Since one sector or area’s planhas implications on activities carried out by othersectors, coordination between different adaptationplans and planning processes is critical.ConclusionGiven the magnitude of climate change costs, adaptationto climate change clearly needs to be a seriousconsideration. As discussed, adaptation investmentsare already under way in all three cities. Eachof the cities will be better protected against climatechange with the flood control measures that areeither already planned or are slowly being implemented.However, additional policy, institutional,and ecosystem-based measures also need to be putin place and prioritized by the city governments.The analysis undertaken in the three casestudies needs to be viewed as an initial attempt atestimating the impacts and damage costs relatedto climate change. As discussed in chapter 2, thereare a number of uncertainties associated with eachlevel of analysis. The climate downscaling and thehydrological models provide results that are by nomeans certain. The economic analysis overlays alarge number of assumptions related to prices, GDP,population distribution, growth, the structure ofcities and so on, over and above these uncertainties.Therefore, they should be viewed as a preliminaryattempt to be followed by improved studies.Assessing Damage Costs and Prioritizing Adaptation Options | 73

Table 4.16 ■ Proposed Implementation Arrangements for HCMCAuthorityHCMC DONREHCMC Environment Protection Agency (HEPA)Department of Planning and Architectureand its Institute for Urban PlanningDepartment of Construction with MOCLine departments and institutes responsiblefor (a) transport, (b) power supply, (c) watermanagement and supply, (d) water qualityand sanitation, (e) industry, (f) agricultureand fisheries, and (g) public healthHCMC Steering Committee for Flood andStorm ControlPriority actions(i) Revise the HCMC land use strategy and action plan to incorporate climate change issues and adaptationmeasures(ii) Prepare assessment guidelines for reviewing sector and spatial plans adaptation requirements andconsistency with the city adaptation plan(iii) Prepare assessment guidelines for integrating adaptation in SEA and EIA when applied to developmentplans and project proposalsPrepare adaptation monitoring and audit guidelines to keep track of adaptation performanceRevise the HCMC urban strategy and plan to set out the spatial land uses, controls, and safeguardsfor adaptationRevise and pilot the Building Code in the city so that it responds to climate change(i) Audit existing infrastructure and development plans and orientations(ii) Retrofit adaptation measures in existing infrastructure(iii) Define sector-specific adaptation options(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(i) Support each commune and district in reviewing and revising their specific contingency plans toprotect and cope with more extreme flooding and storm events, and identifying the key assets andresidential 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 majorflood events(iv) Develop an early warning system for traffic and alternative transport routes in the event of floodsSource: ADB (2010).74 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

5Conclusions andPolicy ImplicationsThe studies undertaken in the three cities showthe challenges faced by coastal megacitieswith respect to regional and global climatechange and variability. These studies highlight theimportance of achieving sustainability in urban areasthrough a better understanding of the relationshipbetween urbanization, local hydrological systems,and climate. All of the cities considered in this reportare megacities with (registered and unregistered)populations close to or greater than 10 million. Eachis a driver of national and/or regional economicgrowth and all depend on the capacity of local hydrologicalsystems—such as rivers and streams,watersheds, and wetlands—to provide a range ofservices that are critical to the sustainability of thesecities and its residents. Given that these are all coastalmegacities, they face increased climate-related riskssuch as sea level rise and increased frequency ofextreme events. In this chapter, we conclude byhighlighting key findings and policy implicationsthat would be useful for urban policy makers toconsider with respect to adaptation at the city level.Key Findings and Lessons forPolicy MakersMajor risks are posed by increasedtemperature and precipitation in coastalmegacitiesAll three cities are likely to witness increases intemperature and precipitation linked with climatechange and variability. In Bangkok, temperatureincreases of 1.9 ° C and 1.2 ° C for the A1FI and B1scenarios respectively are estimated for 2050 andare linked with a 3 percent and 2 percent increasein mean seasonal precipitation for the high and lowemissions scenarios. A flood that currently occursonce in fifty years may occur as frequently as oncein 15 years by 2050, highlighting the potential foran increase in the frequency of extreme events. InManila, the mean seasonal precipitation is expectedto increase by 4 percent and 2.6 percent for thehigh and low emissions scenarios. In HCMC, theaverage annual temperature is expected to rise by1.4 ° C in 2050 (with only slight differences betweenthe high and low emissions scenarios). Further, a20 percent and 30 percent increase in precipitationis expected for a 1-in-30-year and a 1-in-100-yearevent respectively under the high emission (A2)scenario. While extreme rainfall linked with stormsis expected to become more common, preliminaryanalysis suggests that the frequency of dry seasondrought in HCMC is also likely to increase by 10percent under the low emissions scenario. Theserisks need to be recognized and better understoodby urban planners in the context of specific coastalcities, particularly in developing countries.Increase in flood-prone areas due to climatechange in all three citiesIn all three megacities in 2050, there is an increase inthe area likely to be flooded under different climatescenarios compared to a situation without climatechange. In Bangkok, for instance, under the conditionsthat currently generate a 1-in-30-year flood butwith the added precipitation projected for a highemissions scenario, there will be approximately75

a 30 percent increase in the flood-prone area. InManila, even if current flood infrastructure plansare implemented, the area flooded in 2050 will increaseby 42 percent in the event of a 1-in-100-yearflood under the high emission scenario comparedto a situation without climate change. In HCMC,the area inundated increases (from 54 percent ina situation without climate change to 61 percentin 2050 with climate risks considered) for regularevents and for extreme (1-in-30 year) events from68 percent (without climate change) to 71 percent(with climate risks considered) in 2050 under thehigh emission scenario. Further, there is a significantincrease in both depth and duration for both regularand extreme floods over current levels in 2050. Theanalysis also highlights areas that will be at greaterrisk of flooding in each metropolitan area. In MetroManila, for instance, areas of high population densitysuch as Manila City, Quezon City, Pasig City,Marikina City, and San Juan Mandaluyong City arelikely to face serious risks of flooding.Increase in population exposed to floodingdue to climate changeIn all three cities, there is likely to be an increase inthe number of persons exposed to flooding in 2050under different climate scenarios compared to asituation without climate change. For instance, inBangkok, in 2050, the number of persons affected(flooded for more than 30 days) by a 1-in-30-yearevent will rise sharply for both the low and highemission scenarios by 47 percent and 75 percentrespectively compared to those affected by floods ina situation without climate change. In Manila, for a1-in-100-year flood in 2050, under the high emissionscenario more than 2.5 million people are likely tobe affected assuming that the infrastructure in 2050is the same as in the base year, and about 1.3 millionpeople if the 1990 master plan is implemented. InHCMC, currently, about 26 percent of the populationwould be affected by a 1-in-30-year event. However,by 2050, it is estimated that approximately 62 percentof the population will be affected under the highemission scenario without implementation of theproposed flood control measures. Even with theimplementation of these flood control measures,more than half of the projected 2050 population isstill likely to be at risk from flooding during extremeevents. How to plan for such large percentages ofpopulation being exposed to future flooding needsto be seriously considered.Costs of damage likely to be substantialin and can range from 2 to 6 percentof regional GDPIn Bangkok, the increased costs associated withclimate change (in an A1FI scenario) from a 1-in-30-year flood is THB 49 billion ($1.5 billion) orapproximately 2 percent of GRDP. These are the additionalcosts associated with climate change. Theactual costs of a 1-in-30-year flood would includecosts resulting from land subsidence and would beeven higher. In Manila, a similar 1-in-30-year floodcan lead to costs of flooding ranging from PHP 40billion ($0.9 billion) given current flood controlinfrastructure and climate conditions to PHP 70billion ($1.5 billion) with similar infrastructure butfuture A1FI climate. Thus, the additional costs ofclimate change from a 1-in-30-year flood wouldbe approximately PHP 30 billion ($0.65 billion) or6 percent of GRDP. The HCMC study adopts a differentmethodology to analyze costs and its resultscannot directly be compared to the costs of Manilaand Bangkok. The HCMC study uses a macro approachand estimates a series of annual costs up to2050. The flood costs to HCMC, in present valueterms, range from $6.5 to $50 billion. The “annualized”costs of flooding would likely be comparableto the costs of Bangkok and Manila.Damages to buildings is an importantcomponent of flood-related costsDamage to buildings emerges as a dominant componentof flood-related costs, at least in Bangkokand Manila. In these cities, over 70 percent offlood-related costs in all scenarios are a result ofdamage to buildings. Cities are, almost by definition,built-up areas full of concrete structures. Itis, therefore, not surprising that the main impactsof floods is on these structures and the assets theycarry. In HCMC, 61 percent of urban land use, 67percent of industrial land use, and 77 percent of76 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

open land use is expected to be flooded in 2050in an extreme event if the proposed flood controlmeasures are not implemented. Future floodingalso has major sector implications. For instance, inthe transportation sector in HCMC, even with theimplementation of the proposed flood protectionsystem, a significant percentage (76 percent of axisroads, 58 percent of ring roads) of the existing andplanned roads will be exposed to increased flooding.Thus, as cities develop over the next 40 years,it would be important to design their commercial,residential, and industrial assets and zones to minimizethese costs.Impact on the poor and vulnerable willbe substantial; however, even better-offcommunities will be affected by floodingIn Bangkok and Samut Prakarn, the study estimatesthat about 1 million inhabitants will beaffected by the A1FI climate change condition in2050. One out of eight of the affected inhabitantswill be those living in condensed housing areaswhere the population primarily lives below thepoverty line. Of the total affected population, approximatelyone-third may have to encounter morethan a half-meter inundation for at least one week,marking a two-fold increase in the vulnerablepopulation. People living in the Bang Khun Thiandistrict of Bangkok and the Phra Samut Chedi districtof Samut Prakarn will be especially affected.In HCMC, in general, poorer areas in HCMC aremore vulnerable to flooding. However, in some ofthe areas, the poor and non-poor are both at risk.The Manila study shows that while low-incomeresidents have devised many coping strategies toextreme events, state and non-state actors need toplay a more proactive role in addressing vulnerabilityof the urban poor.Urban plans and flood protectioninfrastructure need to take climate risksinto accountFlood protection plans are already in place in allthree mega-cities considered. However, in all threecities, these have been prepared without consideringfuture climate-related risks, which need to be consideredfor the infrastructure to provide the expectedlevel of protection from future climate risks. InHCMC for example, which has a long history of respondingto natural disasters, a number of measuressuch as dykes and early warning systems (which arealso important aspects of a climate change adaptationstrategy) have been developed. However, thelikelihood of increasing frequency and intensity ofextreme events has not been built into these disastermanagement plans. Consideration of climate-relatedrisks has also not been sufficiently integrated intourban planning and in various sectoral plans in thecities. As the HCMC case shows, in addition to theurban master plan, there are a number of city-levelsectoral plans with little coordination between them.The main policy implication is that climate changeadaptation measures need to be well-integratedinto urban, sectoral, and flood protection plans andthe coordination between these plans needs to bestrengthened.Reduction of land subsidence should beconsidered an important part of urbanadaptationOne of the main findings of this study is that nonclimate-relatedfactors such as land subsidence areimportant and in some cases even more importantthan climate risks in contributing to urban flooding.In Bangkok, for instance, there is nearly atwo-fold increase in damage costs between 2008and 2050 due to land subsidence. Further, almost70 percent of the increase in flooding costs in 2050in the city is due to land subsidence. While datafor land subsidence was not available for Manilaand HCMC and this issue was not considered inthe hydrological modeling for these two cities,available literature suggests that it is an importantfactor in all three cities and should be consideredin follow-up studies. Even though the megacitieshave already undertaken a number of measuresto slow down land subsidence, further regulatoryand market incentives are clearly required to stemgroundwater losses. City governments need tobetter assess factors contributing to land subsidenceand consider options to reduce it. Based onqualitative information provided in the city studies—the impact of other non climate factors such asConclusions and Policy Implications | 77

dumping of solid waste into the city’s canals andwaterways, poor dredging of canals, siltation ofdrains, deforestation and soil erosion in the upperwatershed, and land use patterns—are also criticaland should be part of a broader and multisectoralapproach to addressing urban adaptation.Important to consider variation in climaterisks and factors contributing to floodingAn important finding of this study is that whilea combination of climate-related factors can contributeto urban flooding, some factors are muchmore important than others in different cities. Forinstance, in HCMC, storm surges and sea levelrise are important factors contributing to flooding.However, in Bangkok these factors seem tobe relatively less important. In part this has todo with the location, elevation, and topographyof the city. The policy implication is that adaptationmeasures need to be designed based on thespecific hydrological and climate characteristicsof each city.Need for greater emphasis on ecosystembasedapproachesIn all of the cities considered, an important focusof flood management and control is on hardinfrastructure solutions. However, the studies(Bangkok and HCMC) show that ecosystem-basedapproaches are also important and can be usefullycombined with infrastructure interventions. TheHCMC study, for instance, argues for combiningconstruction of dykes (as approached by theplanned flood control measure) with managementand rehabilitation of the mangrove systems in CanGio, reforestation of the Dong Nai River Basin upperwatersheds, river and canal bank protection,and implementation of basin-wide flow managementstrategies. Urban wetlands provide a rangeof services, including flood resilience, allowinggroundwater recharge and infiltration, and providinga buffer against fluctuations of sea level andstorm surges. Rehabilitation of urban wetlands istherefore critical. Upstream protection of watershedsthrough reforestation is also important inmanaging the release of runoff into the reservoirsthat serve urban residents. These options need tobe considered in developing a comprehensive approachto urban adaptation. Dykes themselves aresusceptible to erosion by rainfall and storms. Assuch, it is also important to consider protection ofdykes through planting.Importance of locally induced climatechange factorsEven though this report does not take into accountlocally induced climate change factors, these arevery significant (Grimm et al. 2008). For example,the rate of increase of temperature in HCMC is almostdouble that of the surrounding Mekong Deltaregion. Between 1997 to 2006, the average temperatureof HCMC increased .34° C compared to .16° Cfor the Mekong Delta region. This increase in temperaturefor HCMC has coincided with increasingurbanization in the city. In HCMC, the heat islandeffect is changing the climate and urbanization ishaving a significant effect on increases in temperature,rainfall, and flooding in and around HCMC.Further analysis is needed to better understandthis issue.Lessons on Methodologyfor Follow-up StudiesEven though analysis similar to that undertakenin this report has been carried out for cities in developedcountries (Rosenzweig et al. 2007), it hasso far not been done in developing country cities.Moreover, preparation of this report has spurredsimilar studies in other cities in the Africa andMiddle East and North Africa regions. It is importanttherefore to conclude with a few lessons forfollow-up studies.Approaches to downscalingIn this study two different approaches to downscalingwere used, namely statistical scaling usedby JICA and use of a dynamic regional model asdone by the HCMC study. Each method has itsstrengths and limitations. Dynamic regional modelscan simulate daily information on a broad range of78 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

hydrometeorological data, yielding information onfloods and droughts and on the joint occurrence ofvariables, like the pattern of precipitation over alarge watershed. Dynamic models that are nestedin AOGCMs are necessarily bound to the parentmodel’s intrinsic biases. There is, for example, awide variation in the results of the AOGCMs inforecasting extreme precipitation, and use of onemodel alone can yield unreliable results. Statisticalmodels yield single parameters and shed little lighton the joint probability of occurrence. Nevertheless,because the results are based on a correlationwith the results of many models, the downscaledparameters have less intrinsic bias than those generatedby a single model. These constraints need to beconsidered in undertaking similar city-level studies.Need to forge stronger links betweenmethodologies developed by disaster riskreduction community and climate changeadaptationThere is a strong global consensus on the increasinglyurgent need to address the underlying causes of climatechange and its impacts through mitigation ofgreenhouse gas emissions and development of effectiveadaptation strategies. This has been underscoredat the recent United Nations Framework Conventionon Climate Change (UNFCCC) Conferences of Parties(COP) in Copenhagen (COP-15), which againhighlighted the need to foster better linkages betweenclimate change adaptation (CCA) and disaster riskreduction (DRR). The Hyogo Framework of Action,signed by 168 countries in 2005, is the guiding documentfor the DRR initiatives to mainstream DRR intothe national, sectoral, and local-level developmentplanning process. The DRR community has wellestablishedprocedures and terms of reference for riskassessment, including probabilistic risk assessment,that allow assessment of risks based on an analysisof hazards, exposure, vulnerability, and capacity(ECLAC 2003). The use of simulation models to developevent-loss relationships are likewise commonlyapplied in the DRR work (e.g., hydrometeorological,earthquake, storm surge, tsunami, and landslidemodels). In addition, the DRR community has thelocal networks and knowledge management systemsthat will provide significant benefit for climatechange adaptation studies. These synergies need tobe recognized and strengthened.Lessons with respect to damage costanalysis and prioritizing adaptation optionsThe three studies were based on available data. Thismeant that various assumptions had to be maderegarding the number of households affected, the“damage rate” or extent of damage to various assetsas a result of floods of different intensity, and thecosts associated with these damages. The underlyingdata and valuation would be much improved ifthere was scope for survey-based data collection. Arather important gap in the studies is the analysisof health impacts. Floods do have serious healthconsequences and while the studies have attemptedto estimate these, the full importance of these impactshas not been captured. To carefully examinethe health impacts of floods would require first anestimation of disease prevalence in the future (takinginto account climate and economic conditions)and second, application of robust dose-responserelationships to establish the extent of flood-relatedillnesses. This is an impact issue that requires morecareful consideration. In terms of adaptation toclimate change, the Bangkok and Manila studiesare unable to incorporate “soft options” such aschanges in legal and institutional issues into theircost-benefit analyses. Thus, it is possible that thereare less costly policy changes that might bringabout the same type of results as the engineeringsolutions identified. However, identifying the fullimplications of different policy reforms and thenincorporating these into a cost-benefit frameworkis difficult. Future studies would do well to lookat behavioral change options, zoning possibilities,legal and institutional reforms, and conservationpossibilities carefully and rank them relative toengineering solutions.Need for detailed institutional analysis infuture studiesIn undertaking similar studies in the future, a moredetailed analysis of institutional and organizationalcapacity for urban adaptation is suggested. Whileall three city studies considered in this report doConclusions and Policy Implications | 79

consider broad institutional issues, a more detailedcapacity analysis could provide a more nuancedunderstanding of policy and regulatory gaps, existingplanning and decision-making processes, coordinationbetween key municipal organizations, andentry points for strengthening state, private sectorand civil society relations with respect to addressingurban adaptation.Uncertainties, limitations, and interpretingthe findings of this studyAny study forecasting conditions four decadeshence will be faced with large uncertainties. Oneissue facing the analysis was what would be thepathway of GHG emissions. To address that issue,the city case studies examined both a high and alow GHG emissions scenario to bracket the likelyfuture conditions. In the climate change downscalingmethodologies, there are uncertainties inforecasting the increase in extreme and seasonalprecipitation under the different scenarios. Thetechniques applied in the statistical downscalingexamined the results from numerous AOGCMs.Robust relationships were identified for temperatureand precipitable water increases, where theinternal error was estimated at ~ 10 percent errorand ~ 10–20 percent error, respectively (Sugiyama2008). Hydrologic models can simulate flood eventswith relatively small errors (

BibliographyAllen, M. R., and W. J. Ingram. 2002. “Constraintson future changes in climate and the hydrologicalcycle.” Nature 419: 224–232.ALMEC Corporation. 2008. Final Report on theStudy on Climate Change Impact over Asian MegaCities (Phase 2): Metro Manila. Final Report to JICA.Manila: ALMEC.American Society of Civil Engineers. 1996. HydrologyHandbook. ASCE Press: New York.Asian Development Bank (ADB). 2010. Ho Chi MinhCity. Adaptation to Climate Change. In collaborationwith the HCMC People’s Committee and DONRE.Prepared by ICEM - International Center for EnvironmentalManagement.Asian Institute of Technology. 2004. An Assessmentof the Socioeconomic Impacts of Floods in Large CoastalAreas. Final Report for APN CAPaBLE Project:2004-CB01NSY-Dutta.CTI Engineering International Co. Ltd. 2008. TheStudy on Climate Impact Adaptation and Mitigationin Asian Coastal Mega Cities -Case of Metro Manila.Final Report to JICA. Tokyo: CTI Engineering InternationalCo. Ltd. (unpublished).ECLAC (Economic Commission for Latin Americaand the Caribbean). 2003. Handbook for EstimatingSocioeconomic and Environmental Effects of Disasters.Mexico: ECLAC.General Statistics Office (GSO) Vietnam. 2006.Vietnam Household Living Standard Surveys, 2002 and2004: Basic Information, Hanoi, Vietnam.Giorgi, Filippo. 2008. “Regionalization of climatechange information for impact assessment andadaptation.” WMO Bulletin 57 (2): 86–92.Grimm, N. B. et al. 2008. “Global Change and theEcology of Cities.” Science 319 (5864): 756–760.Gupta, A. et al. 2009. @ 110 Km/hr. Down to Earth.June 16–30, 2009.Ibaraki University, Institute for Global Change AdaptationScience. 2008. Climate Change Impact overAsian Mega Cities: Advisory for Coastal Engineering andAdaptation. Paper Commissioned by JICA. Ibaraki:Ibaraki University.Intergovernmental Panel on Climate Change(IPCC). 2000. Special Report on Emissions Scenarios.http://www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4_syr.pdfIntergovernmental Panel on Climate Change(IPCC). 2007a. Climate Change 2007: Synthesis Report.http://www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4_syr.pdfIntergovernmental Panel on Climate Change(IPCC). 2007b. Climate Change 2007. Impacts, Adaptationand Vulnerability. Contribution of WorkingGroup II to the Fourth Assessment Report ofthe Intergovernmental Panel on Climate Change(M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. vander Linden and C.E. Hanson, eds.) Cambridge,UK: Cambridge University Press. http://www.ipcc.ch/publications_and_data/ar4/wg2/en/spm.html81

IMF. 2009. World Economic Outlook Database-October2009. Washington, DC: International MonetaryFund.Jonkman, S. N, M. Brinkhuis-Jak, and M. Kok. 2004.“Cost benefit analysis and flood damage mitigationin the Netherlands.” Heron 49 (1): 95–111.Jones, R et al. 2004. Generating High ResolutionClimate Change Scenarios using Precis. NationalCommunications Support Unit Handbook. UNDP,GEF and the Hadley Center. Exeter, UK: Met OfficeHadley Center.McGranahan, G., D. Balk, and B. Anderson. 2007.“The rising tide: assessing the risks of climate changeand human settlements in low elevation coastalzones.” Environment and Urbanization 19 (1): 17–37.Meehl, G. A., T. F. Stocker, W.D. Collins, P. Friedlingstein,A.T. Gaye, J.M. Gregory, A. Kitoh, R. Knutti,J.M. Murphy, A. Noda, S.C.B. Raper, I.G. Watterson,A.J. Weaver, and Z. C. Zhao. 2007. “Global ClimateProjections.” In S. Solomon, D. Qin, M. Manning,Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, andH.L. Miller, eds. Climate Change 2007: The PhysicalScience Basis. Contribution of Working Group I to theFourth Assessment Report of the IntergovernmentalPanel on Climate Change. Cambridge, UK, and NewYork: Cambridge University Press.Mehrotra, S. et al. 2009. “Framework for City ClimateRisk Assessment: Buenos Aires, Delhi, Lagosand New York.” World Bank Commissioned Research.Fifth Urban Research Symposium, Citiesand Climate Change, Marseille, France.Messner, F., E.C. Penning-Rowsell, C. Green, V.Meyer, S.M. Tunstall, and A. Van der Veen. 2007.“Evaluating Flood Damages – Guidance and Recommendationson Principles and Methods.” ReportNo T09-06-01. Floodsite. http://www.floodsite.net/html/partner_area/project_docs/T09_06_01_Flood_damage_guidelines_D9_1_v2_2_p44.pdfMuto, M., K. Morishita, and L. Syson. 2010. Impactsof Climate Change on Asian Coastal Areas: The case ofMetro Manila. JICA Manila Report forthcoming asJICA Research Institute Working Paper.National Economic and Social Development Board(NESDB). 2007. “Poverty Line.” Bangkok. http://www.nesdb.go.th/Default.aspx?tabid=402Nicholls, R.J. et al. 2007. “Coastal Systems andLow Lying Areas.” In M.L. Parry, O.F. Canziani,J.P. Palutikof, P.J. van der Linden, and C.E. Hanson,eds. Climate Change 2007. Impacts, Adaptation andVulnerability. Contribution of Working Group II tothe Fourth Assessment Report of the IntergovernmentalPanel for Climate Change. Cambridge, UK:Cambridge University Press. Available at: Nicholls, R.J. et al. (2008). “Ranking Port Citieswith High Exposure and Vulnerability to ClimateChange: Exposure Estimates.” Environment WorkingPaper No. 1. ENV/WKP (2007)1. Paris: OECD.Panya Consultants Co. Ltd. 2009. Climate ChangeImpact and Adaptation Study for the Bangkok MetropolitanRegion. Final Report Commissioned by theWorld Bank. Bangkok.Rosenzweig, C., D.C. Major, K. Demong, C. Stanton,R. Horton, and M. Stults. 2007. “Managing climatechange risks in New York City’s water system: Assessmentand adaptation planning.” Mitig. Adapt.Strategies Global Change 12: 1391–1409.Satterthwaite, D. et al. 2007. “Building ClimateChange Resilience in Urban Areas and AmongUrban Populations in Low and Middle IncomeNations.” Paper presented at Building for ClimateResilience: Center for Sustainable Urban Development,New York, July 8–13.Smith, K, and Ward. R.1998. Floods: Physical Processand Human Impacts. Chichester: John Wiley andSons.Solecki, W., and C. Rosenzweig. 2008. “ClimateChange Adaptation Strategies and Disaster Risk82 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

Reduction: Inventory and Prospect.” GFDRR WhitePaper. (Draft.)Stern, N. 2007. The Economics of Climate Change – theStern Review. Cambridge, UK: Cambridge UniversityPress.Sugiyama, M. 2008. Study on Climate Change Adaptationand Mitigation in Asian Coastal Mega-cities. IntegratedResearch System for Sustainability Science(IR3S), University of Tokyo. Report Commissionedby JICA.Sugiyama, M. 2009. “Precipitation extreme changesexceeding moisture content increases in MIROCand IPCC climate models.” Proceeding of the NationalAcademy of Sciences of the United States 107(2): 571–575.United Nations. 2007. World Urbanization Prospects:2007 Database. http://www.un.org/esa/population/publications/WUP2005/2005WUP_DataTables11.pdfWorld Bank. 2010. The Cost to Developing Countriesof Adapting to Climate Change: New Methods and Estimates.Washington, DC: World Bank.WWF. 2009. Mega Stress for Mega Cities: A ClimateVulnerability Ranking of Major Coastal Cities in Asia.Gland, Switzerland: WWF International.Yohe, G., J. Neumann, and H. Ameden. 1995. “Assessingthe Economic Cost of Greenhouse-InducedSea Level Rise: Methods and Application in Supportof a National Survey.” Journal of EnvironmentalEconomics and Management 29:S-78–S-97.Yusuf, A.A., and H.A. Francisco. 2009. “ClimateChange Vulnerability Mapping for South Asia.” Singapore:EEPSEA. Available at: Zhang, Yabei, and Jan Bojö. 2009. “Note on the CostBenefit Analysis for the Flood Protection ImprovementProject.” Prepared as a background note to theBangkok Study. (Draft.)Bibliography | 83

AnnexAVulnerability of KolkataMetropolitan Areato increased Precipitationin a Changing ClimateGlobal and RegionalContextStudy ObjectivesThis study sought to:Low-lying coastal areas are highly vulnerable toflooding. A number of recent studies have shownthat coastal areas are vulnerable to a range ofrisks related to climate change, including coastalflooding. Among these coastal areas, the IntergovernmentalPanel on Climate Change specificallyidentifies as hotspots the heavily urbanized megacitiesin the low-lying deltas of Asia (IPCC 2007 b).Among Asian countries, India— with its 7,500 kmlong predominantly low-lying and densely populatedcoastline—is particularly vulnerable. A recentglobal survey identified Kolkata and Mumbai asamong the top ten cities with high exposure toflooding under the current climate change forecasts(Nicholls et al. 2008). The study also showsthat exposure will increase in the future; by 2070,Kolkata is expected to lead the top 10 list in termsof population exposure.In response, the World Bank—in collaborationwith the Asian Development Bank (ADB) and theJapan Japan International Cooperation Agency(JICA)—launched four country-level studies forManila (led by JICA), Ho Chi Minh City (led byADB), and Bangkok and Kolkata (led by the WorldBank).■■■■■■■■■■■■Compile a database with past weather-relatedinformation and damage caused by extremeweather-related episodes.Determine scenarios most appropriate for assessingthe impact of climate change.Develop hydrological, hydraulic, and stormdrainage models to identify vulnerable areasand determine physical damage estimates resultingfrom climate change effects.Assess monetary, social, and environmental impactsresulting from such climate change events.Formulate adaptation proposals to cope withdamage arising from climate change effects.Strengthen local capabilities so that Kolkata’splanning process can account for climate-relateddamage effects in analyzing new projects.In this study, precipitation events in Kolkatabased on available historical rainfall data for 25years were used as a baseline (without climatechange) scenario. For modeling climate change,predictions for temperature and precipitationchanges in Kolkata in 2050 for A1FI and B1 emissionscenarios—from an analysis of 16 GCMs used for theIPCC Fourth Assessment Report—were provided by85

a paper commissioned by JICA (Sugiyama 2008). Asea level rise of 0.27 m by 2050 was also added tothe storm surge for the A1FI and B1 climate changescenarios based on current estimates. All these scenarios(without and with climate change effects)were then modeled to assess the impact in termsof the extent, magnitude, and duration of flooding.Geographic andSocioeconomic ContextsThe geographic area covered in the study is theKolkata Metropolitan Area (KMA), a continuousurban area stretching along the east and west bankof the Hooghly River surrounded by some ruralareas lying as a ring around the conurbation andacting as a protective green belt. KMA has an areaof 1,851 square kilometers and consists of a complexset of administrative entities comprising 3 municipalcorporations (including Kolkata Municipal Corporation,or KMC), 38 other municipalities, 77 nonmunicipalurban towns, 16 out growths, and 445rural areas. KMC, the core of the city, lies along thetidal reaches of the Hooghly and was once mostlya wetland area. The elevation of KMA ranges from1.5 to 11 meters above sea level (masl). The elevationof KMC area ranges from 1.5 to 9 masl with anaverage of 6 masl.With a population of about 14.7 million (including4.6 million in KMC), KMA is one of the 30 largestmegacities in the world (United Nations 2007). Theaverage population density in KMA is 7,950 peopleper km 2 ; in KMC, it is 23,149 per km 2 . The averageper-capita income in KMA in 2001–02 was $341 (at1993–94 prices).The Study AreaA special characteristic of KMA is its large slumpopulation, comprising more than a third of the totalpopulation. These slums not only lack basic infrastructureand services, but are also the hub of manyinformal manufacturing activities, some of which involvehighly toxic industries. Little oversight of suchactivities is carried out by government agencies. Thismixed residential and commercial/industrial landuse in slums make these areas highly vulnerable toextreme weather-related events, especially flooding.MethodologyThe study modeled the impact of climate changeon increased flooding in KMA. The main causes offlooding in KMA are intense precipitation, overtoppingof the Hooghly River due to water inflow fromlocal precipitation as well as that from the catchmentarea, and storm surge effects. Land subsidence—which occurs in a few pockets—was not includedin the study. The study covered three main sourcesthat aggravate flooding in KMA:■■■■■■Natural factors. Natural factors include flattopography and low relief that cause riverineflooding and problems with drainage.Developmental factors. Developmental factors includeunplanned and unregulated urbanization;low capacity drainage; sewerage infrastructurethat has not kept pace with the growth of thecity or demand for services; siltation in availablechannels; obstructions caused by uncontrolledconstruction in the natural flow of storm water;and reclamation of natural drainage areas(marshlands).Climate change factors. Climate change factorsinclude an increase in the intensity of rainfall,sea level rise, and an increase in storm surgecaused by climate change effects.Flooding from intense precipitation was modeledfor three scenarios—30-year, 50-year, and100-year return period flood events—assuming noclimate change effects. The climate change effectswere added to the 100-year flood event using theA1FI and the B1 scenarios.Assumptions about the impacts of climatechange in Kolkata in 2050 (Sugiyama 2008) included(a) a temperature increase of 1.8°C for theA1FI scenario and 1.2°C for the B1 scenario; (b) afractional increase in the precipitation extremes ofabout 16 percent for the A1FI and 11 percent for theB1 scenarios; and (c) sea level rise of 0.27 m by 2050,which was added to the storm surge for the A1FIand B1 climate change scenarios based on current86 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

estimates. All these scenarios (with and withoutclimate change effects) were then modeled to assessthe impact in terms of the extent, magnitude, andduration of flooding.Three separate models were used to capturethe overall effect of natural, developmental, andclimate change factors that lead to flooding in KMA.A hydrological model (SWAT model) was usedto develop the flow series for the whole Hooghlycatchment. The generated data was then fed into ahydraulic model (HECRAS model) to analyze theimplication of the flood passing through the riverstretch. Finally, a storm drainage model (SWMMmodel) was used to determine the flooding that willresult once the river flooding is combined with localprecipitation and drainage capability of the urbanarea under an extreme flood situation.Vulnerability analysisThe models generated the increase in depth, duration,and extent of flooding in Kolkata due to climatechange effects. A separate vulnerability analysiswas done to assess the impact of flooding; this partof the analysis was restricted to only the KMC areabecause of data limitations. The vulnerability analysiswas based on three separate indices: (1) a floodvulnerability index based on the depth and durationof flooding, (2) a land-use vulnerability index basedon the nature of land use, and (3) a social vulnerabilityindex based on the existing infrastructure andthe socioeconomic characteristics of the population.Finally, the three separate indices were combined toform a composite vulnerability index.Vulnerability was assessed at both the ward andsubward levels. To evaluate the impact of flooding inKMC, the analysis identified the 9 most vulnerablewards that may need specific attention in designingadaptation strategies. To assess vulnerabilitywithin each ward in greater detail, the analysis wasextended to the sub-ward level using spatial data.Damage assessmentThe vulnerability analysis was followed by aseparate economic damage assessment. Due todata limitations, this was also restricted to the KMCarea. Damage to stocks measured primarily physicaldamage arising out of water submersion (sectorsincluded residential buildings and property, com-INDIAWEST BENGAL STATEAND THE CITY OF KOLKATAIBRD 37779SEPTEMBER 2010West BengalArea: 88752 sq. km.Population: 80 millionPop. Density: 904/sq. km.Kolkata Metropolitan AreaArea: 1851.41 sq. km.Population: 14.7 millionPop. Density: 7950/sq. km.RiverHooghlyKolkotaI N D I AWestBengalKolkataArabianSeaBay ofBengalKolkata Municipal CorporationArea: 187 sq. km.Population: 4.57 millionPop. Density: 23149/sq. km.HooghlyRiverGarden ReachTiljalaThis map was produced by the Map Design Unitof The World Bank. The boundaries, colors,denominations and any other informationshown on this map do not imply, on the partof The World Bank Group, any judgment onthe legal status of any territory, or anyendorsement or acceptance of such boundaries.INDIAN OCEANVulnerability of Kolkata Metropolitan Area to increased Precipitation in a Changing Climate | 87

mercial and industrial establishments, and majorpublic infrastructure). Flow damage included loss ofincome and increased morbidity, which are primarilylinked to the duration of a flood. The damageestimates were based on data extrapolated to theKMC population in 2050. The analysis assumed noadditional investments in flood protection measuresthat may be implemented in the future to lowerflood damage. Inflation also was not considered,and all estimates used 2009 prices. Damage assessmentswere estimated for the 100 year flood returnperiod and the A1FI climate change scenario todetermine the additional damage caused by climatechange effects.Adaptation analysisFinally, a separate analysis was done to examineadaptation measures in KMC that can alleviatesome of the problems posed by flooding. Theanalysis mainly focused on gains from the completede-siltation of trunk sewers by modeling floodingunder a completely de-silted trunk sewer scenario.The study also examined other proposals to buildnew sewers and upgrade sewers in vulnerable areas,as well as institutional changes that can help copewith future flood damage.Main FindingsThe most vulnerable wards to climate change. Thestudy identified the most vulnerable wards to climatechange—namely, wards14, 57, 58, 63, 66, 67,74, 80, and 108. Six wards (14, 57, 58, 66, 67, and108) in the eastern part of the city and ward 80 arevulnerable because of inadequate infrastructure,unplanned land use, and poor socioeconomic andenvironmental conditions. Infrastructural problemsare getting worse with increased building activity,as these areas have become attractive to developersafter becoming part of KMC.The other two wards—63 and 74—are highlyvulnerable due to their topography. In these wardsthe capacity of the sewerage system has not keptpace with changes in population. These have beenfurther aggravated by inadequate maintenance,as well as the siltation of the existing trunk sewersystems, which have considerably reduced their carryingcapacity. While the sewer networks in KMCunder such partially silted condition still providereasonable hydraulic capacity for carrying the dryweather flow, they are inadequate for carrying stormweather flow, even with normal precipitation duringthe rainy season.Additional losses likely to occur due to climatechange. Damage from a 100-year flood will increaseby about $800 million—to more than $6.8 billion in2050—due to climate change (A1FI scenario). Theimpacts by sector (in Indian rupees at 2009 prices)are shown in the chart below. Local currency wasconverted to dollars using the purchasing powerparity index for India of 2.88 (IMF 2009). The largestdamage components—under both the 100-yearreturn period flood and the A1FI climate changescenario—are for residential property and buildingsand health care. Commerce, industry, and otherinfrastructure like roads and transport services alsoTotal losses in major sectors in KMC (Rs million in 2050)40,00035,00030,00025,00020,00015,00010,0005,0000ResidentialbuildingResidentialpropertyResidentialincome lossCommerce Industry Healthcare Roads Transport ElectricityCurrent climate 100 year floodA1F1 climate 100 year flood88 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

sustain significant damage. Due to data constraints,some impacts could not be quantified in this analysis,so the estimates provided are likely to understatethe overall impact of climate change.Investing in de-silting of trunk sewers willreduce the area and population affected by floods.The study looked at the impact of investing inde-silting trunk sewers both in the town andsuburban systems of KMC. The business as usualscenario considers an average 30 percent silting intrunk sewers. The adaptation scenario considersinvesting in de-silting and reducing it to zero. Thefindings indicate that this simple investment canreduce the area affected by a flood by 4 per centand the population affected by floods by at least5 percent.Adaptation MeasuresThe current adaptation deficit. Climate change islikely to intensify urban flooding through a combinationof more intense local precipitation, riverineflooding in the Hooghly, and coastal storm surgesA major cause of such periodic flooding during therainy season is the city’s current adaptation deficit,including deficiencies in physical infrastructure,problems with land use, and socioeconomic andenvironmental factors.Adaptation strategy. The city needs a comprehensiveand effective strategy that invests inboth soft and hard infrastructure to tackle floodingproblems in Kolkata. The goal of the strategy is to (a)reduce the percentage of people affected by floodingand sewage-related diseases in KMC, and (b)target the most vulnerable areas. The strategy shouldinclude preparedness both before and during theevent, as well as post-event rehabilitation strategies.Investing in hard infrastructure. Investing inhard infrastructure should take into account thefollowing:■■The strategy needs to follow a comprehensiveapproach to planning that recognizes drainagesystem complexity and interconnectivity of its elementssuch as storm water drainage, water supply,wastewater, water pollution control, waterreuse, soil erosion, and solid waste management.■■■■The strategy should protect major urban services,including roads, traffic, water supply, electricity,and telecommunications. It should recognizethe importance of open space and green areas asan integral part of city development.The strategy should spell out the climate risksand mitigating factors needed in operationalplans for key relevant agencies.Investing in soft infrastructure. To ensure longtermfinancial, institutional, and environmentalsustainability, the adaptation strategy should alsoinclude:■■■■■■■■■■■■■■Strengthening disaster management andpreparedness for both pre- and post-disastersituations.Enforcing land use and building codes to reduceobstruction and encroachment of floodplainsand environmentally sensitive areas such ascanal banks and wetlands and to prevent conversionof green spaces and natural areas thatcan act as retaining zones during flooding.Introducing sustainable financing—emphasizingboth cost reduction and cost recovery—forinfrastructure investment and maintenance.Increasing the budget for sewerage anddrainage maintenance and the allocation ofmoney for silt removal and mechanical sewercleaning.Adopting flood insurance that incorporates suitableincentives for adaptation and minimizesflood damage.Strengthening the regulatory and enforcementprocess, including improving institutional managementand accountability.Enforcing pollution management frameworks,including introduction of incentivesand disincentives to ensure compliance withregulations.The government of West Bengal has alreadystarted investing in adaptation. Among the suggestedadaptation measures, a number of projectsare either currently under way or are plannedfor future implementation in KMA under theJawaharlal Nehru National Urban Renewal Mission(JNNURM) and the KEIP scheme funded byVulnerability of Kolkata Metropolitan Area to increased Precipitation in a Changing Climate | 89

ADB. The selection and prioritization of projectsfor adaptation should be based on cost benefitanalysis using the net present value (NPV) approach.Factoring in the additional impact due toclimate change in such cost benefit analysis mayrender many projects—which previously did notshow an adequate return on investment—economicallyviable.90 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

AnnexBScenarios Applied in theHydrodynamic Modeling inthe HCMC studySeasonal ExtremeINPUTSStormUpstreamSCENARIOS Rainfall Rainfall Surge Mean Sea Level Inflow Landuse PURPOSESI.BL.0 Baseline_Max No No Baseline Baseline Baseline Extent and duration of the “regular” seasonalmonsoon floods with current land useI.B2.0 B2_Max No No B2 2050+24cm Baseline BaselineI.A2.0 A2_Max No No A2 2050+26cm Baseline BaselineI.BL.1 Baseline_Max No No No Baseline 2050 Extent and duration of the “regular” seasonalmonsoon floods with planned futureI.B2.1 B2_Max No No B2 2050+24cm Baseline 2050land use and flood control systemI.A2.1 A2_Max No No A2 2050+26cm Baseline 2050I.BL.0_1 Baseline_Max Baseline_30y No No 100y 2050 Extent of the floods caused by 30y stormI.B2.0_1 B2_Max B2_30y No B2 2050+24cm 100y 2050 rainfall (but without storm surge) and 100yextreme upstream inflow; with plannedI.A2.0_1 A2_Max A2_30y No A2 2050+26cm 100y 2050future land use and flood control systemII.BL.0 Baseline_Max Baseline_30y Linda No 100y Baseline Extent of the “extreme” floods driven byII.B2.0 B2_Max B2_30y Linda B2 2050+24cm 100y Baseline combination of tropical storms (typhoons)surge, mean sea level rise, 30y rainfall andII.A2.0 A2_Max A2_30y Linda A2 2050+26cm 100y Baseline100y upstream inflow; with current land useII.BL.1 Baseline_Max Baseline_30y Linda No 100y 2050 Extent of the “extreme” floods driven byII.B2.1 B2_Max B2_30y Linda B2 2050+24cm 100y 2050 combination of storms surge, mean sealevel rise, 30y rainfall and 100y upstreamII.A2.1 A2_Max A2_30y Linda A2 2050 +26cm 100y 2050inflow; with planned future land use andflood control systemII.BL.0_a Baseline_Max No Linda No Baseline Baseline Extent of floods and saltwater intrusionII.B2.0_a B2_Max No Linda B2 2050 +24cm Baseline Baseline caused by storm surge and sea level riseonly in current land useII.A2.0_a A2_Max No Linda A2 2050 +26cm Baseline BaselineII.BL.0_b Baseline_Max Baseline_30y No No 100y Baseline Extent of the floods caused by 100 yII.B2.0_b B2_Max B2_30y No B2 2050 +24cm 100y Baseline extreme upstream inflow and 30y stormrainfall (without storm surge) in currentII.A2.0_b A2_Max A2_30y No A2 2050 +26cm 100y Baselineland useIII.bl.0 Baseline_Min No No No Baseline Baseline Extent of combination effect of drought andIII.B2.0 B2_Min No No B2 2050 +24cm Baseline Baseline sea level rise on water quality especially riskof salinization under current land useIII.A2.0 A2_Min No No A2 2050 +26cm Baseline BaselineI.BL.1_1 Baseline_Max Baseline_30y No No 100y 2050 Extent of floods driven by 30y rainfall, 100yI.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 floodI.A2.1_1 A2_Max A2_30y No A2 2050 +26cm 100y 2050control systemSource: ADB (2010).91

AnnexCAdaptation to IncreasedFlooding: Brief OverviewIn response to the hazard of floods increasingfrom climate change, the Asian coastal cities inthe study have a range of possible adaptationapproaches drawn from existing flood managementtechniques. Each adaptation approach attempts toreduce risk by either reducing exposure or vulnerability,or by increasing capacity. The adaptationmeasures can be broadly grouped as: 59■■■■■■■■ProtectAccommodateRetreatImproved managementProtection—reducing exposure withstructural interventionsFlood protection measures seek to minimize theexposure of urban areas to inundation from floods.Common engineering interventions include floodembankments, polders, and sea walls, frequentlycombined with pumped drainage. To allow thetransfer of a flood wave past a city without damage,other protection techniques include (a) increasingthe hydraulic efficiency 60 of the flood channel withdredging, widening, and removal of obstructions;(b) diverting the flood flows around the city throughdiversion channels; and (c) attenuating 61 the floodflows upstream with reservoirs or through themanaged flooding of the agricultural and wetlandareas. With the exception of planned flooding ofagricultural and wetlands to attenuate flood peaks,the measures outlined above would be classifiedprimarily as structural adaptation measures.Within a city, storm drainage systems are designedto remove storm runoff without flooding.The design discharge for the conveyance systemsare normally based on storms with a return periodbetween 1-in-5-years to 1-in-10-years, as comparedto design standards of 1-in-30 to 1-in-100-year returnperiods commonly used for flood protectionembankments. In an urban setting, roads, buildings,and paved surfaces offer little opportunity forinfiltration of storm rainfall, which means that mostof the precipitation appears as storm runoff. In thisregard, the predicted increase in storm precipitation(ranging from 9 to 15 percent depending on the cityand the scenario) will generate a similar increase inthe design discharge for the storm drainage systems.To provide the same level of protection againststorm-induced urban flooding, the cities will needto adapt by either increasing the discharge capacityof their storm drainage systems or identifyingways to delay or reduce precipitation runoff. Forexample, adaptation measures might include stormattenuate ponds (using inner city parks) or makingstreets, alleys, and other paved areas more porousto rainfall to facilitate infiltration.Flood protection measures always carry a residualrisk of failure, which can be catastrophic as inthe case of New Orleans. An important outcome ofthe study is that even for cities with well-designedflood protection plans (e.g., Bangkok and Manila),the flood protection measures as currently plannedwill not offer the expected level of protection.59Drawn from IPCC, AR4, Chapter 6, Coastal systems andlow-lying areas, Figure 6.11 Evolution of planned coastaladaptation measures, pg. 342, 2007.60Improving the hydraulic efficiency lowers the flood profile.61Attenuation upstream lowers the flood profile downstream.93

Accommodate—a traditionapproach in monsoonal Asiato reducing vulnerabilityAccommodation acknowledges that people andassets are exposed to floods, but seeks to minimizevulnerability. In rural Asia, houses located in floodplains and exposed to frequent flooding are usuallybuilt on stilts, which reduces the vulnerability of thepeople and the household assets to flooding. Croppingcalendars designed around annual monsoonfloods and floating varieties of rice are other waysin which rural Asia accommodates floods. In citieswhere ankle-deep flooding occurs regularly, woodenplanks on blocks are common sights to allow passageand access of pedestrians. In periurban areas, landfillis frequently used to elevate new construction sitesalong major roads above the expected flood levels.Accommodation, however, can be applied ona broader scale. Flood losses in urban areas ariseprimarily from damage to buildings, houses, andcommercial/industrial properties. In this regard,risks could be dramatically reduced if buildingcodes specified minimum elevations for the firstoccupied level for housing, commerce, or industry(based on predicted inundation levels for the zonesin the city). These elevations could be reviewed andincreased as warranted with long-term (5 to 10 year)advance projections to developers. Given the timeframe of the study (40 years), this gradual approachgives urban planners a powerful tool to reduce riskswith minimal economic impact.Accommodation also relates to measures to“flood-proof” structures and building exteriors thatmay be exposed to floods. Flood-proofing dramaticallyreduces damages from floods. These standardsshould be included in urban building codes as partof an adaptation program. The study has demonstratedthat traditional accommodation methodsmay not reduce vulnerability in 2050. In HCMC, forexample, areas with frequent ankle-deep floodingwill be facing knee-to-waist-deep floods, which arenot easily managed by pedestrians. This will reduceaccess and increase losses. Similarly, many housesand buildings that have been built with main floorsabove the current 1-in-30-year flood levels will beinundated by 1-in-30-year floods in 2050.Retreat—when risks aretoo high, retreat reducesexposureRetreat as planning option to reduce exposure can beapplied in urban areas through urban land use plansand zoning codes. The social, environmental, and economicimplications need to be carefully studied. In urbanareas with high property values, it is usually onlyapplicable where the risks of loss of life or assets arehigh enough to warrant the action. In rural areas, retreatis common practice vis-à-vis flood embankmentsas rivers modify courses and erode embankments.Forced retreats are occurring in Bangkok’s coastal areadue to coastal erosion and land subsidence.Used in conjunction with restoring naturalecosystems that provide flood protection benefits,retreat can be a useful adaptation planning tool. Incoastal areas, land that becomes untenable due toSLR could be converted to natural systems (e.g.,mangroves) that can help capture sediment andprotect against SLR and storm surges. In urbanareas, reclaiming land along rivers and convertingthese to linear parks that will inundate during floods(increasing hydraulic capacity and lowering theflood profile) help beautify the city while providinga range of benefits. Similarly, agricultural landsthat may become unviable due to flooding couldbe purchased and converted to wetlands. Retreatsometimes occurs following major disasters whenthe perceived risks outweigh the desire to reconstructamong the private sector. Planners shouldseize such opportunities and develop the lands fortheir ecological and flood protection potential.In addition to the retreat from section of a cityor coast line, facilities that are critical in disastersituations or at particular risk (like schools) can beretired and moved as part of a longer term adaptationprogram.Improved management—building capacity to reducerisksFlooding needs to be considered in the context ofoverall water basin management and the capacity94 | Climate Risks and Adaptation in Asian Coastal Megacities: A Synthesis Report

of people and institutions to manage the resource.Experience has shown that piecemeal or ad hoc approachesare at risk of causing significant adverseimpacts elsewhere. For example, flood embankmentsprotect one area by aggravating the flood inanother area. Water uses and hazards need to beaddressed holistically for the water basin to maximizebenefits.With the hazard of flooding increasing, adaptationmeasures aimed at improved managementcan focus on (a) reassessing operational rule curvesfor reservoirs, given the increasing precipitationvariability; (b) increasing awareness and community-basedadaptation initiatives (e.g., shelters andearly warning systems); and (c) strengthening theplanning and risk-sharing mechanisms related tocontrolled flooding of agriculture lands to attenuateflood peaks when protecting cities.Adaptation to Increased Flooding: Brief Overview | 95

AnnexDComparison of Costsacross CitiesAn interesting question to ask is how climatechange costs compare across the different cities.The figure below provides a comparisonbetween the Manila and Bangkok case studies inthe context of a 1-in-30-year flood. HCMC is notincluded in this analysis because the methods forassessing damage costs were very different. Thefigure below shows that in Bangkok, for instance,the damages from a 1-in-30-year flood wouldincrease by almost 50 percent going from currentclimate to an A1FI climate change scenario, while asimilar flood in Manila will have a larger effect andincrease costs by some 79 percent. However, whilethe Bangkok damage costs account for land subsidence,the Manila costs do not. Thus, accounting forland subsidence may make the increase in costs inthe two cities rather similar.The costs of climate change are accuratelymeasured as the costs in similar scenarios with andwithout climate change. Thus for Manila, the costsof climate change are represented by the differencebetween the 30-A1FI-EX and 30-SQ-EX scenarios.The costs for Manila from a 1-in-30-year flood are$0.67 billion, which amounts to 6 percent of its2008 GRDP. Similarly, the costs for Bangkok—thedifference between 2050-LS-SR-SS-A1FI and 2050-LS scenarios—amount to $1.47 billion, or 2 percentof Bangkok’s 2008 GRDP. Given that Bangkok is amuch wealthier city than Manila, these numbersare credible.How does HCMC stand in relation to the twocities? As previously noted, the costs to HCMC,because they are in present value terms and becauseof the different methodology used, are notcomparable to the estimates from Bangkok andManila. However, to get a rough understanding ofComparing Damage Costs in Bangkokand Manila80%70%60%50%40%30%20%10%0%72%49%Percent increasein costs of 1/30flood due toclimate change15%9% 3% 5% 6% 2%Costs of 1/30flood as a % ofGRDP (withoutclimate change)Costs of 1/30flood as a % ofGRDP (with A1FIclimate change)Manila 2008 Bangkok 2008Additional costs ofclimate change asa % of GRDPwhether they are in the same order of magnitude,we annualize these costs. We find that annualequivalent of the damage costs to HCMC fallsbetween $0.67 and $5 billion, 62 which on the lowerend is comparable to the costs borne by Bangkokand Manila.62These numbers are the equal annual equivalents (EAE)(using a 10 percent discount rate and time period of 42years) of the present value of total costs from regular andextreme flooding. The range reflects the minimum andmaximum present values of the two land and GDP-basedestimations of climate change costs. EAE = NPV [(i(1 + i)n] / [(1 + i)n − 1)] where i=interest rate and n=time period.97

THE WORLD BANK1818 H Street, NWWashington, DC 20344Telephone: 202.473.1000Internet: www.worldbank.orgE-mail: feedback@worldbank.orgContacts:Poonam Pillai, ppillai@worldbank.orgJan Bojo, jbojo@worldbank.orgMaria Sarraf and Susmita Dasgupta, msarraf@worldbank.org, sdasgupta@worldbank.orgAsian Development BankContact: Jay Roop, jroop@adb.orgJapan International Cooperation AgencyContact: Megumi Muto, Muto.Megumi@jica.go.jp