Report No. 43946-IN Climate Change Impacts in Drought and Flood Affected Areas: Case Studies in India June 1, 2008 South Asia Region India Country Management Unit Sustainable Development Department Social, Environment and Water Resources Management Unit Document of the World Bank Disclaimer This report has been discussed with the Government o f India but does not bear their approval for all its contents, especially where the Bank has stated its judgment/opinion/policy recommendations. The findings, interpretations, and conclusions expressed in this paper are based on staff analysis and recommendations and do not necessarily reflect the views o f the Executive Directors o f The World Bank. ... 111 Table of Contents Abbreviations andAcronyms Conversions Disclaimer Acknowledgments ........................................................................................................................................ Glossary.................................................................................................................................................... xiii ...x ExecutiveSummary .................................................................................................................................... 1 A.Backgroundto the Study ................................................................................................................. 1 B. Copingwith andAdaptingto Drought Conditions.......................................................................... 4 C.Elementsof a Strategy for Adaptation to Drought Conditions........................................................ 6 D.ClimateChange andVulnerability to Floods................................................................................ 10 E. Elementsofa Strategy for BuildingFloodResilience .................................................................. 11 F. ConclusionsandRecommendations:............................................................................................. 13 1. Introduction,Context, and Objectives ....................................................................................... 16 1.1 Introduction........................................................................................................................... 16 1.2 Context: ClimateVariability. a DevelopmentChallenge...................................................... 17 1.3 Objectives.Process. andApproach....................................................................................... 20 2. India's Climate: Background, Trends, andProjectionTools .................................................. 25 2.1 India's Current Climate. ItsVariability. andTrends............................................................. 25 2.2 FutureClimateChangeProjections....................................................................................... 26 2.3 Methodology ofthe Study..................................................................................................... 27 3. ClimateVariability and Change:A CaseStudyinDrought-ProneAndhra Pradesh ............36 3.1 IntroductionandBackground................................................................................................ 36 3.2 Characteristicsof Study Area ................................................................................................ 36 3.3 ImpactofDrought ................................................................................................................. 40 3.4 FutureProspectsunder ClimateChange................................................................................ 43 3.5 PullingTogether the Pieces: Policy Implications.................................................................. 49 4. ClimateVariability and Change: A CaseStudyinDrought-ProneMaharashtra .................56 4.1 Introduction andBackground................................................................................................ 56 4.2 Characteristicsof Study Area ................................................................................................ 57 4.3 Impact of Drought ................................................................................................................. 60 4.4 FutureProspects under Climate Change................................................................................ 61 4.5 PullingTogether the Pieces: Policy Implications.................................................................. 66 iv 5. ClimateVariability and Change: A Case Study inFlood-Prone Orissa ................................. 72 5.1 Introductionand Background................................................................................................ 72 5.2 Characteristics of Study Area................................................................................................ 72 5.3 Impacts of Floods .................................................................................................................. 75 5.4 FutureProspectsunder Climate Change................................................................................ 78 5.5 Policy Implications................................................................................................................ 81 6. A Way Forward ............................................................................................................................ 87 6.1 StrengtheningClimate Information Systems andMechanisms ............................................. 88 6.2 Fostering Climate-Resilient Reforms inAgncultureand Water Resource Management......88 6.3 Economic Mechanisms and Instrumentsto Promote IncomeDiversification....................... 89 6.4 ImprovingInstitutionalCapacities andProgram Linkages ................................................... 89 6.5 FutureWork........................................................................................................................... 90 Appendix A Integrated Modeling System-Framework and Analysis ........................................................ 93 Appendix B Methodology usedfor the Design and Analysis of Household Surveys and Data ..............100 Appendix C InternationalConference on Adaptation to Climate Variability and Change...................... 106 Appendix D Programs that Address Droughts and Floods inthe Case Studies....................................... 112 Appendix E A Conceptual Framework for Maharashtra Drought Adaptation Pilot inRainfedAreas ....116 Appendix F Program for StakeholdersConsultation................................................................................ 125 Appendix G Descriptionofthe Economic Model.................................................................................... 127 Appendix H Maps of Location and Climate Change Impacts inStudyAreas......................................... 136 Bibliography.............................................................................................................................................. 143 Boxes Box 1.1 Working Group I1of IPCC: Summary of Rural Impacts............................................................... 18 Box 2.1 EmissionScenarios Selectedfor the Study.................................................................................... 31 Box 2.2 Structure ofthe FarmEconomic Model ........................................................................................ 33 Box 3.1 Groundwater Crisis Looms Large over Anantapur and Chittoor Districts .................................... 37 Box 3.2 Illustrationo fTrends inCrop Responses inthe Pennar Basin, Andhra Pradesh........................... 46 Box 3.3 Andhra PradeshDrought Adaptation Initiative: PuttingAdaptation into Practice........................ 49 Box 3.4 Andhra Pradesh Sectoral Programs: Comprehensive Base to BuildAdaptation Approaches.......50 Box 3.5 Weather-Indexed Insurance for Apculture inIndia..................................................................... 52 Box 3.6 Livestock Systems ......................................................................................................................... 54 Box 4.1 Rainfall Distribution inMaharashtra ............................................................................................. 57 Box 4.2 Unsustainable Groundwater Development PosesProblems for Ahmednagar and Nashik............59 Box4.3 Short-Term CopingMeasures and Responsesto DroughtinMaharashtra.................................... 60 V Box 4.4 Sugarcane Yield and Climate Change inAhmednagar: EPIC Model Projections ........................ 65 Box 4.5 Role o f Community Institutions and Participatory Water Resource Management inDrought Adaptation .......................................................................................................................................... 69 Box4.6 Snapshot o f Sectoral ProgramsinMaharashtra............................................................................. 71 Box 5.1 Flood Coping Strategies in Study Area ......................................................................................... 77 Box 5.2 FloodManagement in~ s s......................................................................................................... a 82 Box 5.3 PolandFloodEmergency Project (1997-2005): Good Practice inNonstructural Measures for FloodManagement............................................................................................................................. 84 Box AI Stochastic Weather Generator ....................................................................................................... 97 Box A2 RegionalClimate Models vs.Global Climate Models .................................................................. 99 Figures Figure 1.1Elements o f Vulnerability to Climate inRuralAreas o f India................................................... 20 Figure 1.2 Study Framework ................................................................................................................. 24 Figure 2.1 All-India Average Surface Temperature 1948-1998 ................................................................. 25 Figure 2.2 Variation o f All-India Monsoon Rainfall 1813-2003 ................................................................ 26 Figure 2.3 I M S SystemArchitecture ........................................................................................................... 32 Figure3.1 LocationMap o f Study Villages inChittoor District................................................................. 38 Figure 3.2 LocationMap o f Study Villages inAnantapur District ............................................................. 38 Figure3.3 HouseholdPercentage with Irngation Access (by Source & Landholding) .............................. 39 Figure 3.4 Impact o fDrought on Income.................................................................................................... 40 Figure 3.5 Determinants of Income Volatility ............................................................................................ 41 Figure 3.6 Spatial Distribution o f Average Annual Rainfall inthe Pennar Basin (Baseline and Climate Change Scenarios) .............................................................................................................................. 44 Figure 3.7 Yields of Rice, Groundnut, and Jowar under Different Climate Scenarios, Anantapur ............45 Figure 3.8a Per Hectare Average Profits inBaseline Scenario, Talupula Block, Anantapur...................... 47 Figure 3.8b Area Allocation inBaseline Scenario, Talupula Block, Anantapur......................................... 47 Figure 3.9a Per Hectare Profits inClimate Change Scenarios, Anantapur ................................................. 47 Figure 3.9b Total Profits inClimate Change Scenarios, Anantapur ........................................................... 47 Figure 3.10 Acreage Comparison between Projections o fthe Farm Economic Model and Data fiom HouseholdSurveys............................................................................................................................. 48 Figure 3.11Yield-Rainfall CorrelationinAnantapur inA2 Scenario ........................................................ 53 Figure 3.12 Cropping Pattern inAnantapur................................................................................................ 54 Figure 3.13 ProposedCrop Diversification................................................................................................. 54 Figure 4.1 LocationMap of Study Villages inAhmednagar District ......................................................... 58 Figure 4.2 LocationMap of Study Villages inNashik District................................................................... 58 vi Figure 4.3a Percentage o f Households with IrrigationAccess: Normal Year............................................. 59 Figure 4.3b Percentage o fHouseholds with IrrigationAccess: Drought Year ........................................... 59 Figure 4.4 Drought Impact on Income ........................................................................................................ 60 Figure 4.5 Nonincome Impacts on Households .......................................................................................... 60 Figure 4.6 Spatial Distribution o f Average Annual Rainfall inthe GodavariBasin (Baseline and Climate Change Scenarios) .............................................................................................................................. 63 Figure4.7a Bajra: Average Yields, Nashik................................................................................................. 64 Figure 4.7b Bajra: Distribution o f Yields. Nashik ...................................................................................... 64 Figure 4.8a Sugarcane: Average Yields ...................................................................................................... 65 Figure 4.8b Sugarcane: Distribution o f Yields............................................................................................ 65 Figure4.9 Sugarcane Productivity Trend inAhmednagar (Tonnes per Hectare) ....................................... 68 Figure 5.1 Location o f Study Villages inJagatsinghpur District ................................................................ 75 Figure 5.2 Location o f Study Villages inPuri District................................................................................ 75 Figure 5.3 Average Percentage o f Sources o f Income inTotal Monthly Income in Sampled Households (by Landholding Size) ........................................................................................................................ 75 Figure 5.4 Changes inAgricultural and NonagriculturalIncomes as a Result of a FloodEvent................76 Figure 5.5 Spatial Distribution o f Average Annual Rainfall inLower MahanadiBasin (Baseline and Climate Change Scenarios) ................................................................................................................ 79 Figure 5.6 Exceedance Probability Curves for Annual Peak Flows at Naraj Gauge Station ...................... 79 Figure 5.7a Yield Changes inA2 and B 2 by Crop, PuriDistrict ................................................................ 80 Figure 5.7b Distribution o f Yields under Climate Change. Puri District .................................................... 80 Figure 5.8 Impact of Climate Change on Farmer Profits Derived from Selected Crops. Puri District.......81 Figure A.1Flow o f Data and Information inthe IntegratedModeling System .......................................... 94 Figure A.2 Interface for Installationo f the Application.............................................................................. 95 Figure A.3 I M S Toolbar inArc View ......................................................................................................... 95 FigureA.4 Crop Yields ShownUsingBar Charts ...................................................................................... 96 FigureA.5 Spatial Representationof Average Crop Yield inthe Chittoor District.................................... 96 Figure B.1Spatial Representationo f Average Crop Yield inthe Chittoor District.................................. 104 Figure E.1Maharashtra Drought Adaptation Pilot and Convergence with Ongoing Programs ...............119 Figure G.l Distribution o fGroundnut Yields inAndra Pradesh............................................................... 131 Figure G.2 Area Alloction and Water Supply, Medium Farm, Andra Pradesh ........................................ 132 Figure G.3 Profits and Levels o f Water Shortage inA2 Scenario inAndra Pradesh................................ 132 Figure G.4 Millet FarmProfits and Shortage o f Water inMaharashtra.................................................... 132 Figure H.l Andra Pradesh Study Area...................................................................................................... 136 Figure H.2 Maharashtra Study Area.......................................................................................................... 136 FigureH.3 Orissa Study Area ................................................................................................................... 137 vii Figure H.4 Average Annual Rainfall inFour Districts inAndra Pradesh................................................. 138 Figure H.5Average Rainfall inKharif SeasoninFour Districts inAndra Pradesh.................................. 138 Figure H.6 Average Maximum Temperature inKharif Season inFour Districts inAndra Pradesh ........139 Figure H.7Average Kharif Season Rainfall inFive Districts inMaharashtra.......................................... 140 Figure H.8Average Annual Rainfall in Six Coastal Districts inOrissa ................................................... 141 Figure H.9 Average Rainfall inKharif Seasonin Six Coastal Districts inOrissa .................................... 141 Figure H.10 Average MaximumTemperature inKharif Season inSix Coastal Districts inOrissa .........142 Tables Table 1.1 Selected Indicators o f Changes inthe Earth's Climate inthe 20th and early 21st Century........17 Table 2.1 Resultso f Ricardian Assessments o f Climate Change Impacts on Crops inIndia ..................... 28 Table 2.2 Agronomic Assessments o f Climate Change Impacts on Agriculture inIndia (by Source and Region) ............................................................................................................................................... 29 Table 3.1 Socioeconomic Profile o f Sampled Householdsby Landholding Size ....................................... 39 Table 3.2 Average Share o f Sources o f Income inTotal Income inSampledHouseholds, 2003-04 .........39 Table 3.3 Percentage of Households with h g a t i o n Access ...................................................................... 42 Table 4.1 Socioeconomic Profile o f Surveyed Households inMaharashtra............................................... 58 Table 4.2 Outlays and Expenditures for Drought Reliefand Sectoral Programs,....................................... 67 Table 5.1 Floods and Resulting Damage inOrissa 2001-2007 .................................................................. 74 Table 5.2 Socioeconomic Characteristics o f Puri and Jagatsinghpur Districts Compared to Orissa State .74 Table 5.3 Average Seasonal Crop Productioninthe Surveyed Households............................................... 76 Table 5.4a Orissa: Outlays and Expenditures for Sectoral Programs 2002-2007 (Rsbillion) ...................83 Table 5.4b Onssa: Allocation and Expenditures for Drought and Flood Relief 2002-2006 (Rs billion) ..83 Table 6.1 Summary Recommendations for Adaptation .............................................................................. 92 Table B.1Definitions o fVulnerability...................................................................................................... 100 Table E.1Potential Impacts and Responses for Addressing Vulnerability to Climatet Variability and Climate Change inMaharashtra ....................................................................................................... 124 Table G.l Critical Precentage Changes for Diversificiation out o f Groundut inAndra Pradesh..............129 Table G.2 RiskAversion and Cropping Mix inAndra Pradesh................................................................ 130 Table G.3 Competitive Crop Prices,Fertilization Cost andUser Charge o fWater inMaharashtra..........130 Table G.4 RiskAversion and CroppingMix inMaharashtra ................................................................... 130 Table G.5 Farmers' Knowledge o f Climate Events and Cropping Mix .................................................... 131 Table G.6 Andra Pradesh- OLS Regressions ............................................................................................ 133 Table G.7 Maharashtra- OLS Regressions ................................................................................................ 134 Table G.8 Descriptive Statistics for Andra Pradesh.................................................................................. 134 ... Vlll Table G.9Correlation Matrix for Andra Pradesh...................................................................................... 134 Table G.10 Descriptive Statistics for Maharashtra.................................................................................... 135 Table G.11Correlation Matrix for Maharashtra ...................................................................................... 135 Table G.12DescriptionofVariables......................................................................................................... 135 ix Acknowledgments This report i s the product o f a collaborative effort between the World Bank and the Government o f India under the overall leadership of the Ministry o f Environment and Forests (MoEF). The report has also received significant support from several ministries and agencies of the Government o f India; and the state governments o f Andhra Pradesh, Maharashtra, and Orissa. Special gratitude i s extended to the following government officials: Dr.Prodipto Ghosh, Former Secretary, J.M. Mauskar, Additional Secretary, Rajshree Ray, Deputy Secretary, MoEF, Mr.J.C. Saharia, Former Secretary and Dr. K.S. Vatsa, Former Secretary, both from the Department o f Rural Development and Water Conservation, Government of Maharashtra (GoM); and Mr. Aurobindo Behera, Principal Secretary, Department o f Water Resources, Government o f Orissa (GOO). Significant contributions from various representatives from Non-Government Organizations (NGOs), central and state sectoral agencies during meetings and workshops held at various stages o f the study and their technical support to the study consultants through data and information, are gratefully acknowledged. This report has received funding from the U.K. Department for International Development (DFID), although the views expressed within do not necessarily reflect official policy. We would like to especially thank the Technical Advisory Group (TAG) members that participated inthe consultative process and guided the design o f the study: The TAG members include Dr R. K. Pachauri, The Energy and Resources Institute, Dr. A. Patwardhan, Technology Information Forecasting and Assessment Council (TIFAC), Delhi; Dr. P. K. Aggarwal, Indian Agricultural Research Institute, Delhi; Dr. K. S. Kavikumar, Madras School o f Economics, Chennai; Dr. Rupa Kohli, Indian Institute for Tropical Meteorology, Pune; Dr. Santosh Kumar, National Institute o f Disaster Management, Delhi; Prof. A. K. Gossain. Indian Institute o f Technology, Delhi; and Dr. Ajay Mathur, Bureau o f Energy Efficiency, Ministry of Power, Delhi.Our gratitude is extended to Dr.DavidRadcliffe, Senior Rural Livelihoods Advisor from DFID, Government o fU.K.,for his continuous supportto this activity. The contributions and written comments received from a number o f experts deserve special mention and are gratefully acknowledged, especially Dr. K. S. Kavikumar from the Madras School o f Economics for detailed comments and advice. Gratitude is also extended for comments, data and inputs from: Dr.Charles Batchelor, Dr. Stephen Foster, Dr. V. P. Dimriand Dr. D. Muralidharan from the National Geophysical Research Institute, Hyderabad; Mr. A. K. Jain and Mr. Pradeep Raj from the Groundwater Department, Government of Andhra Pradesh; Dr. Himanshu Kulkamifrom the Advanced Centre for Water Resources and Development and Management,Pune; Dr. S. P. Bagde, Groundwater SurveyandDevelopment Agency, GoM; and finally, Dr.S. Umarikar, Department ofWater Supply and Sanitation, GoM. The World Bank team was initially led by Bilal Rahill and in the subsequent phase by Richard Damania with Priti Kumar as co-task-team-leader. The team included Suphachol Suphachalasai, Carla Vale and Kumudni Choudhary. Other Bank staff who contributedto the production o f this report were Sanjay Pahuja, Ranjan Samantaray, Bela Varma, Rachel Susan Palmer and Vinod Ghosh. Technical background reports were produced by a team o f consultants from the Risk Management Solutions India (RMSI) Pvt. Ltd. led by Satya Priya and Murthy Bachu; The Energy and Resources Institute (TERI) led by Preety Bhandari and Suruchi Bhadwal with X support from Sreeja Nair, Souvik Bhattacharya, Meena Sehgal and Kadambari Anantram; and PragmatixResearchand Advisory Services Pvt. Ltd.ledbyA. J. James. Peer reviewers are Gretel Gambarelli, Chandrasekar Govindarajalu and Willem Janssen of the World Bank. Helpful comments and contributions were received from World Bank staff representing several sectors and regions, including Ernest0 Sanchez-Triana, Michel Pommier, Siet Meijer, Charles Cornier, Malcolm Jansen, Martien Van Nieuwkoop and Winston Yu. Special thanks are due to Kseniya Lvovsky and Karin Kemper for detailed and insightful comments on the draft report and for their support. The support and contribution o f Robert Watson for his guidance on scoping out the study; and IanNoble for guidingthe overall climate modeling exercise are also acknowledged. Finally, we would like to express our sincere appreciation to Michael Carter, Former Country Director for India, Jeffrey Racki, Former Acting Sector Director, South Asia Environment and Social DevelopmentUnit, and Fayez Omar, Senior Manager, IndiaProgram, for their overall guidance and support to this activity. Thanks are due to John Dawsonand Allison Russell for their meticulous contributions inediting the report and appendices, respectively. The opinions presented here and any errors are the sole responsibility of the authors and should not be attributed to the individuals or institutions acknowledged above. xi Glossary aquaculture. Breeding and rearing fish, shellfish, etc. or growing plants for food in special ponds. adaptation. A process by which strategies to moderate, cope with, and take advantage o f the consequences o f climatic events are enhanced, developed, and implemented. Types o f adaptation include anticipatory and reactive adaptation, private and public adaptation, and autonomous and plannedadaptation. adaptive measures or responses. Actions taken that result in building the capacity o f communities and boosting their long-term resilience to climatic shocks or stress. See also coping measures. adaptation strategy. A broad plano f action that i s implementedthrough policies and measures. Strategies can be comprehensive, focusing on national, cross-sectoral scales; or targeted, focusing on specific sectors, regions, or measures. adaptability. The degree to which adjustments are possible inpractices, processes, or structures o f systems to projectedor actual changes o f climate (IPCC 1995, Working Group 11). adaptive capacity. The ability o f a system to adjust to climate change (including climate variability and extremes), to moderate potential damages, to take advantage o f opportunities, and to cope with the consequences. Adaptive capacity i s limitedby existing information, technology, andthe resources o fthe systemunder consideration (IPCC 2001, Working Group 11). adaptive potential. A theoretical upper boundary o f responses based on global expertise and anticipated developments within the planninghorizon o fthe assessment(UNDP 2004b). basin. The drainage area o f a stream, river, or lake. biophysical vulnerability. Sensitivity o f a natural system to an exposure to a hazard. block or mandal. An administrative subdivision o f a district, which inturn i s a subdivision o f a state. C4 plant. A landplant that uses a so-called C4 fixation methodto transform carbon dioxide into sugar. Chemically speaking, the method allows for binding the gaseous molecules to dissolved compounds inside the plant for sugar production through photosynthesis. C4 fixation i s an improvement over the simpler andmore ancient C3 carbon fixation strategy usedby most plants. The intermediate compounds o fthe process contain four carbon atoms, hence the name C4. climate change. Following the IntergovernmentalPanelon Climate Change (IPCC), any change in climate over time, whether due to natural variability or as a result of human activity. This usage differs from that in the United Nations Framework Convention on Climate Change (UNFCCC), which defines climate change as "a change o f climate which is attributed directly or indirectly to human activity that alters the composition o f the global atmosphere and which i s in addition to natural climate variability observed over comparable time periods." See also climate variability. climate prediction or climate forecast. The result o f an attempt to produce a most likely description or estimate o fthe actual evolution o fthe climate inthe future. xii climate projection. A forecast o f the response o f the climate system to emission or concentration scenarios of greenhouse gases and aerosols, or radiative forcing scenarios, often based upon simulations by climate models. Climate projections differ from climate predictions in that they depend upon the emission, concentration, or radiative forcing scenario used, which are based on assumptions concerning, for example, hture socioeconomic and technological developments that may or may not be realized and are therefore subject to substantial uncertainty. climate variability. The variation in the mean state and other statistics (such as standard deviations and the occurrence of extremes) o f the climate on all temporal and spatial scales beyond that o f individual weather events. Variability may result from natural internal processes within the climate system (internal variability) or from variations in natural or human-induced external forcing (external variability) (IPCC 2001). coping measures or responses. Actions taken inthe short term by households and communities that result in diminishingthe impacts o f climatic shocks or stress on them temporarily and which helpto tide them over the stressperiod. drought. It i s defined in many ways and includes the phenomenon that results when precipitation is significantly below normal recorded levels, causing serious hydrological imbalances that adversely affect land resource production systems. In the case o f India, the definition o fwhat is considered "precipitation below normal" varies from agency to agency. event or impact year. The year inwhich an event such as a drought or flood takes place. emission scenario. A plausible representation o f the future development o f emissions of greenhouse gases and aerosols based on a coherent and internally consistent set o f assumptions about driving forces (such as demographic and socioeconomic development and technological change) and their key relationships. In 1992, the IPCC presented a set of emission scenarios that were used as a basis for the climate projections inthe Second Assessment Report (IPCC 1995). These emission scenarios are referred to as the IS92 scenarios. For the Third Assessment Report (IPCC 2001) new emission scenarios, namely the SRES scenarios (Special Report on Emission Scenarios of the IPCC), were published. These are known as the Al, A2, B1, and B 2 scenarios and were also utilized inthe preparation of the Fourth Assessment Report (IPCC 2007a, 2007b, 2007c; final report inpreparation). evapotranspiration. The combined process of evaporation from the Earth's surface and transpiration from vegetation. extreme weather event. An event that i s rare within its statistical reference distribution in a particular place. Definitions o f "rare" vary, but an extreme weather event would normally be as rare as or rarer than the 10thor 90thpercentile. Bydefinition, the characteristics o fwhat i s called extreme weather may vary from place to place. An extreme climate event occurs when the same considerations apply to an average o f a number o f weather events over a certain period o f time (e.g. rainfall over a season). flood. A phenomenon that occurs when an increase in precipitation, which i s above normal recorded levels in a specific timeframe, leads to the volume of water within a body o f water, such as a river or lake, to surpass the total holding capacity o f that body. As a result, some o f the water flows sit outside their normal perimeter, potentially causing serious damage and adversely affectingpeople and land resource production systems. xiii impact of climate. Consequenceso f climate change on natural and human systems that usually affect people and communities in a negative way (though they could also have beneficial effects). (climate) impact assessment. The practice o f identifying and evaluating the detrimental and beneficial consequenceso f climate change on natural andhumansystems. intensity.A measure o fthephysical strengtho f a damage-causing event, such as a flood. irrigation.A method o f purposely providing land with water, other than rainwater, by artificial means. normalyear. Any year inwhich an extreme event, such as drought or flood, does not impact a geographic area. policies and measures. Means o f addressing the need o f climate adaptation in distinct but sometimes overlapping ways. Policies typically refer to the courses o f action that governments can adopt to change economic and other behaviors through, for example, such instruments as taxation, command-control regulations, market mechanisms, incentives, and information gathering or dissemination. Measures are usually specific actions amenable to implementation, such as reengineering irrigation systems, planting different crops, or initiating a new industry. Manyprojects can also betermedmeasures. resilience. The ability to exploit opportunities, and to resist as well as recover from negative shocks. It i s also defined as the amount o f change a system can undergo without changing its state. risk (climate related). The result o f the interaction of physically defined hazards with the properties o f the exposed systems interms o f their sensitivity or (social) vulnerability. Risk can also be considered as the combination o f an event, its likelihood, and its consequences; that is, risk equals the probability o f a climate hazard multipliedby a given system's vulnerability (Lim et al. 2005). runoff. Surface flow occurring when the precipitationrate exceeds the infiltration rate o fthe soil or other surface material. sensitivity. The degree to which a system will respond, either adversely or beneficially, to a givenchange inclimate (or other external pressures). socioeconomic vulnerability. An aggregate measure o f human welfare that integrates environmental, social, economic, and political exposure to a range o f harmful perturbations. Can also be definedas the sensitivity o fthe human system to an exposure to a hazard. stakeholders. Any persons with interest in a particular decision, either as individuals or as representatives o f a group. This includes people who influence a decision or can influence it, as well as those who are affected by it. thermal expansion. Inconnection with sea level rise, this refers to the increase involume (and decrease in density) that results from the warming o f water. A warming o f the ocean leads to an expansion o fthe ocean volume andhence an increase insea level. vulnerability. See appendix B xiv Executive Summary A. Backgroundto the Study A.1 ClimateVariabilityand Change 1. With alpine conditions, arid deserts andtropical regions, India's climate is as varied as its landscape. The summer monsoon marks the most important event in the economic calendar o f rural India. Over 70% o f the annual precipitation falls between the months o f June and September and a good monsoon heralds a bountiful harvest and financial security. But when monsoons fail, or are excessive, suffering and economic loss can be widespread. Climate variability has been the source o f both misery and prosperity for much o frural India. India is already experiencing the effects o f climate variability. It i s at risk of considerably deeper impacts if climate projections are indicative of what may actually happen. 2. Recognizing the significance of climate variability on growth and development, the Government of India has established a range o f programs, policies and institutions to moderate the impacts o f climate-related risks. These long-standing programs have done much to unleash the development potential o f agriculture andhave helpedbuildresilience to climate shocks. India's disaster management programs rank among the most comprehensive in the world and have achieved considerable success in countering the most severe effects o f extreme events. When floods or drought descend, an elaborate relief machinery springs into operation, with rapidly arranged protective policies that include employment schemes, cash and food disbursements, and emergency health care. Fiscal Burden of Climate Extremes 3. The extensive relief systems have come at a substantial price on the public purse. Several state governments spend significantly more on relief and damages than on core rural development programs. In the state o f Maharashtra, a single drought (2003) and flood (2005) absorbed more of the budget (Rs 175 billion), than the entire planned expenditure (Rs 152 billion) on irrigation, agriculture, and rural development from 2002-2007. Climate change i s expected to increase the frequency o f extreme events in ways that are outside the realm o f experience, so relief measures and their financing may not be sustainable, particularly if droughts and floods concomitantly become more severe. This i s already beginning to compromise the effectiveness o f many development programs. Though relief can be strengthened and will continue to remain a vital part o f the defense against climate extremes, in the long term there i s a clear fiscal and development imperative to strengthen climate resilience o f the rural economy by addressing the root causes o fvulnerability. Vulnerability of Agriculture 4. With an emerging gap between the languishing perfonnance o f agriculture and the accelerating growth o f industry, the Government o f India has assigned the highest priority to supporting development in the agriculture sector in the Eleventh Five Year Plan period (2007-2012), with targeted growth o f 4%. However, delivering on the promise o f faster agricultural growth will be difficult, given the multiple constraints facing the sector including, fragmented landholdings, inadequate market access and 1 rapidly depleting natural resources. This will be made more challenging by the impacts and consequences o f ongoing and future climate change. For the poor and marginal' farmers clustered along the poverty line, even small climatic shocks could impose large and irreversible losses, triggering poverty and destitution. Reaching development targets will therefore require priority investments inbuildingthe climate resilience o fvulnerable rural communities with a portfolio o f adaptation options that can address climate risks.2 GeographicDiversity: A Further Challengefor Adaptation 5. India's immense geographic diversity adds to the complexity o f developing an adaptation strategy. Projections indicate that climate variations in India will be varied and heterogeneous, with some regions experiencing more intense precipitation and increased flood risks, while others encounter sparser rainfall and prolonged droughts. The impacts will vary across sectors, locations and populations. The implication for a country so diverse is that broad generalizations on ways to promote adaptation to climate change will be misleading. Consequently, there can be no one-size-fits-all approach to developing a climate risk management strategy: approaches will need to be tailored to fit local vulnerabilities and conditions. All o fthis renders adaptation policy makingcomplex and difficult. A.11 Objectivesofthe Study 6. It is inthis context that the Government o f India has acknowledged the need to develop adaptation strategies to deal with the possible human toll and economic costs o f climate variability and change. The aim o f this study i s to assist the government inthis endeavor by focusing on selected priorities. The overarching objective o f this report is to promote the mainstreaming and integration o fclimate related risks inIndia's development policies and processes, where this is appropriate. The objectives and scope of work were developed in close consultation with the Ministry o f Environment and Forests as the primary counterpart, a cross-section o f concerned ministries and departments in the central government and in three selected states, and scientific experts from academic, policy and research institutions. In the states, the Department o f Water Resources, Government o f Orissa, and the Department o f Rural Development and Water Conservation, Government o f Maharashtra, supported these assessments, reflecting a multisectoral interest inand demandfor adaptation solutions. 7. The focus o f this report i s on vulnerabilities in natural resources and rural livelihoods, which stand at the front line o f climate change impact^.^ The approach was dictated by government priorities, which indicated the need to (a) assess climate risks to agriculture and livelihoods in areas facing elevated and increasing exposure to droughts and floods; (b) generate better information on current coping and climate risk management strategies in response to droughts and floods; (c) develop and demonstrate the use o f a climate 1Marginal farmers are defined as those who own less than 1acre o f land. * These concerns have been echoed in the Planning Commission's working documents for the Eleventh Five Year Plan. The study builds upon and takes forward many o fthe recommendations o f an earlier World Bank assessment o f the impacts o f droughts inthe state o fAndhra Pradesh (World Bank 20069. 2 modeling framework that could be used to identify future climate risks and (d) use the informationto assist indeveloping the key elements o f a forward-looking adaptation plan that can help improve climate resilience and adaptive capacity. A.111 The Approach 8. Responding to these needs, the assessment is focused on drought-prone regions o f Andhra Pradesh and Maharashtra, and select flood-prone districts in Orissa. The study adopts multiple approaches to meet its objectives: 0 First, it learns lessons from the past and present by gathering statistical information to understand how rural communities inthe study areas cope with and buildresilience to extreme climate events. 0 This is complemented by a review of governmentalprograms and institutions, which identifies policy and administrative gaps and strengths in addressing climate risks. 0 The impacts o f climate variability and change are projected to differ in kind and magnitude from current climate patterns. So lessons from the past may be o f limited relevance in guiding future policy. Looking forward, the study builds an integrated modeling system (IMS) to assess future climate risks and vulnerabilities inthe studyregions. e Finally, the report synthesizes the results and articulates a way forward for promoting adaptation and building climate resilience o f rural communities in the study areas. 9. The modeling framework used inthis study i s a complex but powerful tool that generates information on future climate scenarios and the likely impacts on agriculture. The system integrates a climate model with a hydrological water balance model. Together these feed information (on temperature, precipitation, and soil moisture) into an agronomic (crop growth) model that simulates the impacts on crop yields. A custom-built farm-level economic model assesses the financial consequences for farmers and determines cost- effective adaptation strategies. The development o f the farm economic model represents an innovation o f this study and provides a tool that can be used to assess the financial effects o fdifferent policies and scenarios. 10. Before delving into the main results, a number o f caveats and qualifications are inorder. First, neither the issues, nor the locations, examined inthis report are intendedto provide an exhaustive account o f adaptation to climate variability and change in a country as large and varied as India. With a focus on droughts and floods, the case studies are indicative o f "hotspots" and regions at the edge o f climate tolerance limits. These zones constitute about one-third o f the country comprising manyo f the "lagging regions" with a large population, who are disproportionatelypoor, and most at risk from climate change. There are other regions such as the fragile Himalayas, the biodiverse Western Ghats, the vast coastal areas, and the prolific agricultural lands o f the Gangetic plains that are not covered inthis study and needto be considered insubsequent work. 11. Second, as in all matters relating to climate change, there are long and uncertain time frames. Forecasting climate events, or the economy, even a few years into the future remains an imprecise and hazardous exercise. By its nature the analysis o f climate change 3 must look aheadmany decades. The modelprojections should beinterpreted with caution and viewed as indications o f the possible direction and magnitude o f changes rather than as precise forecasts. Despite these uncertainties, policy makers are compelled to respond to climate risks and make decisions on matters that will be affected by future climate events. Modeling exercises can help policy makers to generate informed decisions based on scientific assessments o frisks, outcomes, and policy impacts. Nevertheless, it needs to be emphasized that model simulations are not predictionsbut scenarios based on a host o f assumptions. B. Copingwith andAdaptingto Drought Conditions B.1 Consequencesof andResponsesto Drought 12. India has a long history o f addressing droughts, so the study begins by exploring how farmers in selected drought-prone areas o f Andhra Pradesh and Maharashtra are affected by and respond to droughts. The immediate consequence o f drought is a predictable and often precipitous decline in agricultural production and income. This ignites an ominous chain o f events - indebtedness, distress sales, asset depletion, and deteriorating health - all o f which perpetuate poverty and deprivation. However, in any community certain households are less affected by drought than others, irrespective o f landholding, wealth, or location. 13. What explains why some exhibit greater resilience to drought than others? The answer lies in the interplay o f key factors: the availability o f water, a critical farm input; economic incentives that shape the way farmers react to climate risks; and the opportunities created by policy and circumstance. Among these the following variables are found to have a disproportionate bearing on vulnerability to droughts: Reliable irrigation supplies, in particular groundwater, can provide insurance against income losses due to meager rainfall. Beyond this self-evident link, the assessment finds more nuanced impacts. The availability o f irrigation supplies tends to promote greater reliance on lucrative and water-intensive crops such as rice and sugarcane. If irrigation supplies are assured through a drought, then agricultural incomes are protected. Conversely, when water sources become depleted there is a more dramatic fall in incomes, with debilitating consequences. In broad terms, holding other factors constant, a household with access to irrigation duringa drought year inAndhra Pradesh earns about 50% more income. The clear implication is that judicious and sustainable water use can provide an . indispensable buffer against deficient rainfall. For this to occur there i s an overwhelmingneedto tackle the unrestrained competition for groundwater. Household indebtedness is another major consequence o f drought in communities facing heightened climate risks in areas with degraded and scarce natural resources. The assessment finds that once encumbered by high debt, households are locked into agriculture and remain more exposed to climate risks. Obvious remedies such as debt-forgiveness schemes may help to appease suffering, but they do not address one o f the root causes o f the problem - an overreliance on rainfall-dependent sources o f income. This suggests scope for 4 . introducing cost-effective policy instruments that simultaneously tackle the problem o f indebtedness and provide incentives for job mobility. Finally, the provision o f local publicgoods, notably infrastmcture and education, provide opportunities for income diversification, thereby limitingthe exposure to drought risks. Infrastructure and education in climate-vulnerable communities yields a double dividend - they generate well-recognized development benefits in the near term and simultaneously buildresilience to drought inthe longer term. In particular education exhibits increasing returns in building climate resilience, whereas infrastructure stimulates economic activity and enhances employment and business opportunity. The policy implication is that climate vulnerabilities need to be integrated into decisions that guide the location and design o f public investments and infrastmcture. B.11 FutureProspectsfor Drought-ProneRegionsunderClimateChange 14. What does future climate change hold for these regions? The modeling framework developed in this study i s used to generate projections of climate events based on two commonly used emission scenarios developed bythe IntergovernmentalPanelon Climate Change (IPCC) for the period 2070-2100.4 The assessment distinguishes between the influence of climate on agriculture, from other possible shifts in the economy (such as prices, economic structure and te~hnology).~ projections find considerable variability The o f impacts across regions and crops. Indry areas most crops respondfavorably to higher precipitation and elevated levels o f carbon dioxide (termed carbon fertilization). But these can be offset by higher temperatures. The net effect i s determined by the magnitude of changes andbaseline conditions. Drought-Prone Districts of Andhra Pradesh 15. Inthearidstudyregions ofAndhraPradeshthe climateprojectionsindicatesubstantially higher temperatures (2.3"C -3.4"C, on average) and a modest but more erratic increase in rainfall (of about 4% to 8% at the basin level). With high prevailing baseline temperatures these changes generate deteriorating agroclimatic conditions, with declining yields for the major crops (rice, groundnut, and jowar). Though all yields decline, conditions are more favorable to groundnut, which is already prevalent in the area, reflecting farm-level adjustments to arid conditions. Despite groundnut's suitability to these harsh conditions there are well-recognized risks that prolonged monocropping brings:pests, disease, and fertility loss. Projections suggest that declining yields o fmajor dryland crops are mirrored in lower agriculture incomes. Inthe harsher climate change scenarios, farm incomes could decline substantially (by over 20%), suggesting that agriculture as currently practiced may not be capable o f sustaining large populations on 4The first (termed A2) depicts a gloomy and pessimistic world where high greenhouse gas emissions result in severe climate change. The other (termed B2) emphasizes sustainability, global cooperation, and lower greenhouse gas emissions. Damages due to changes in climate would call for different remedies from impacts due to variations in economic conditions. Hence the study presents results for key climate change scenarios, holding economic and technical factors constant. It then tests the limits and sensitivity o f the projections by allowing for wide changes in economic and technical variables. The report focuses on results that appear robust to substantial variations inparameters. 5 small rain-fed farms. Recognizing these limits the Government o f Andhra Pradesh has initiated numerous programs to encourage agricultural and occupational diversification anda forward-looking drought adaptation programsupported bythe World Bank. Drought-Prone Districtsof Maharashtra 16. The drought-prone belt in Maharashtra (specifically Nashik and Ahmednagar districts) offers a striking contrast. The climate projections suggest a significant, though more variable, increase inrainfall (approximately 20% to 30% at the basin level) accompanied by higher temperatures o f about 2.4"C to 3.8"C, on average. As a result the yield o f several dryland crops (including the millet varieties o f jowar and bajra) exhibit small improvements and provide a measure o f relief to rain-fed farmers with a boost o f about 8% to 15% in incomes. Prospects for other crops are less certain. Sugarcane i s widely grown on irrigated farms in arid regions o f the state. Under the climate change scenarios sugarcane yields are expected to decline considerably (by nearly 30%) as a result o f increased heat stress causedbythe warmer climate. 17. Negative trends insugarcane production are already visible across the state. Sugarcane i s generously subsidized and has been implicated in the overabstraction o f groundwater. There is mounting evidence o f a significant reduction inthe output and yields, due inpart to increasing levels o f environmental degradation. Climate change pressures would reinforce the manycurrent benefits from encouraging a shift from sugarcane to less water intensive crops. C. Elementsof a Strategy for Adaptation to Drought Conditions 18. I s there a need for additional policy and public investments to promote adaptation to climate change? Vulnerabilities to climate extremes and change are often related to poverty. So it may be argued that as India grows more prosperous, it will inevitably build greater resilience to climate risks. The myriad government programs that deal with education, infrastructure, and job creation also serve a complementary objective o f reducing community exposure to climate risks. In this context, adaptation policy can be viewed as an adjunct to good development policies that promote equitable growth. All o f this might suggest that adaptation to climate change requires no additional policypriority or interventions. 19. There are, however, highrisks associated with complacency that could magnify the costs o f climate variability and change. The projections in this report suggest a considerable and mounting human toll from climate change and highlight the need and urgency for mitigating the avoidable costs, particularly among the vulnerable sections o f society. Incomes on the small rain-fed farms in Andhra Pradesh could decline by 5% under modest climate change and by over 20% under harsher conditions, bringing farmers closer to, and in many cases under, the poverty line. The escalating fiscal strain o f the current drought relief system reinforces the need to tackle the root causes o f drought vulnerability as a priority inthe development process. 20. This report suggests that building greater climate resilience requires a combination o f measures packaged with the right incentives and implemented at multiple levels o f government (local, state, and national). Reactive or singular approaches to droughts, such as relief and emergency assistance or debt relief alone for that matter, are essential to 6 appease suffering, but could generate perverse incentives that perpetuate climate risks and impede appropriate adaptation. Consequently these should be complemented with a combination o f other initiatives that promote longer-term climate resilience. C.1 PolicyApproaches Policiesto BuildAdaptive CapacityandResilienceto ClimateRisks 21. A strategy to build climate resilience and ignite growth needs to take account o f a region's comparative advantage, resource constraints, and the impending changes brought about by climate. The exact policies and interventions will differ by location and circumstance, with emphasis given to four overarching strategies: (a) the need for policy and investment decisions to be based on sound scientific knowledge o f risks, calling for the use o f diagnostic risk assessment tools to generate policy-relevant information; (b) innovations andreforms inagriculture and water management that promote more climate resilient cropping systems; (c) cost-effective andefficient management o f climate risks to promote income diversification through economic instruments; and (d) institutional structures to facilitate these changes. These strategies are formulated on the basis o f the study findings in combination with a review o f the literature and a consultative process with government officials andNGOs, andare discussed below. StrengtheningClimateRiskInformationandTools to MatchNeeds 22. Climate change will have heterogeneous and spatially variable impacts. The first step in building adaptation policy is to identify vulnerabilities and risks to determine priorities for investment and policy. A recent Bank assessment has also emphasized the need for such informed decision making(World Bank 20060. 23. There are two immediate areas where climate risk information i s required: 0 First, climate change could have important ramifications for assets with long lead times and long design lives. The location, construction andrefurbishmento f these will need to incorporate climate risk information. This is particularly important for arid regions where changing water flows or rainfall patterns could require modification or relocation o fvital irrigation infrastructure. 0 Second, the assessment finds that local public goods (notably education and infrastructure) play a powerful role inenhancing climate resilience. Consequently climate risks need to be integrated into policy decisions in these sectors, calling for the use o f diagnostic risk assessment tools to determine how much to invest andwhere to invest. 24. The capacity to generate this information rests with national research centers, while the needanddemand for the information lie with affected communities andthe government. This argues in favor o f a recent suggestion by the Planning Commission to build a climate information system at the national level to disseminate information for planning andmanagement to end users (PlanningCommission2007). This initiative will also need to be supported by improved meteorological information at the subbasin and local levels (blocks and subblocks) to improve forecasting and monitoring capabilities. 7 Innovationand ReformsinAgriculture Water Management 25. Indrought-prone areaswater scarcity isthelimitingfactor ofproductionandplays akey role in shaping the fortunes o f agriculture. Much has been done across the country to address water shortages, with a particular focus on supply-side remedies, including large infrastructure, watershed, rainwater harvesting and water conservation programs, and a host o f community initiatives. However, water management still remains a formidable challenge. The climate change projections indicate that even when farmers have largely adapted to arid cropping patterns, increased demand and consequent water stress could severely jeopardize livelihoods and diminish agricultural productivity. There is an overwhelming case for more aggressively pursuing water conservation in semi-aridand arid regions. Greater attention must be given to hybrid approaches that emphasize the efficiency o f groundwater use and increase the effectiveness o f watershed activities to conserve soil moisture and harvest water. Such measures are not a substitute for pricing policies and water policy reform, which would need to focus on the demand side, for example by strengthening incentives and controlling groundwater demand at the wider geographic scale necessary for effective management. However, they provide interim and feasible measures for reducing vulnerabilities (DFID2005a). Research and Extension 26. The study makes a strong case for a shift in agricultural systems in order to overcome future climate change pressures. It i s clear that small and medium farmers in dryland areas will need greater support with knowledge and policy assistance to make this transition work on a large scale. Muchis already occurring across the country andthere i s researchin dryland farming for rice, horticulture and numerous other crops. Strengthened support for agricultural research and extension i s essential to promote more sustainable modes o f dryland farming. This could include looking at opportunities in farm services associated with low costs o f production and intensifying ago-forestry and livestock- based production systems suitable to dryland areas, among others. "Smart Subsidies" and Incentives 27. In the medium to near term, alternative mechanisms are needed to deliver support to farmers more effectively, with the resulting savings being used to increase public investment in ways that reduce the exposure to climate risks. This would also require complementary measures that address the farm-level incentives (including subsidies and regulations) that have implicitly encouraged the production o f water-intensive crops (such as sugarcane) inaridregions. Experience elsewhere suggests that the use o f interim smart subsidies may offer apragmatic way to shift incentives and, thus, croppingpatterns to modes that are better suited to agroclimatic conditions. Smart subsidies recognize that there are costs to altering croppingpatterns andprovide incentives to change the crop mix by shifting support from environmentally degrading activities to more benign forms o f production. As an example, moving price support from water-intensive agriculture to drylandcrops could helpcounteract the current incentives to cultivate water-thirsty crops. 8 FinancialandEconomic Instrumentsto PromoteIncomeDiversification 28. In semi-arid and arid areas where the natural productivity o f agriculture is low and threatened by droughts, income diversification remains the most obvious and effective way o freducing exposure to climate risks. This bringsnew andunfamiliar transition risks to farmers that can be tackled through a variety o f financial incentives that facilitate the shift to nonfarm activities andpromotejob diversification. Debt Relief Coupled with Other Instruments 29. Farmer indebtedness is among the major impediments to occupational mobility. An important priority and challenge for policy i s to find cost-effective ways o f reaching poorer farmers to help reduce their risk exposure. Coupling debt relief with new risk mitigation instruments is an obvious way to prevent a debt-induced poverty trap. Two innovations merit furtherpolicy consideration and scrutiny: 0 The relief o f old debt could be coupled with the provision o f capital for a new business. This would simultaneously reduce indebtedness and lower the transaction costs o f occupational shifts by providing new opportunities. The myriad ongoingmicro-credit schemesprovide a vehicle to pilot such schemes. 0 A variant o fthis approach would have debt relief coupledwith insurance to cover the initial risks o f shifting from farming to other businesses and provide protection against new and unfamiliar sources o frisk.6 InstitutionalChange: ConvergenceandSynergy between Programs 30. There are a largenumber o fcentral- andstate-sponsored programs for addressing drought risks that are implemented under different guidelines and by various implementing and coordinating agencies. Synergies between these programs could be enhanced through an integrated approach to coordinate priorities and fill gaps inthese programs. This remains a challenging task, as it requires considerable institutional dexterity to synchronize diverse programs managed at different levels o f government within a common framework. There are several institutions at the central level, including the National RainfedArea Authority and National Disaster Management Authority, that are capable o f assuming the role o f an apex coordinating agency whose convening power could be harnessed to coordinate the different planning and implementation processes. Furthermore, enlarging existing schemes, such as the state Employment Guarantee Scheme (EGS) and the National Rural Employment Guarantee Scheme (NREGS) to include resilience-building activities would be a fruitfbl policy measure to encourage sustainable farming systems in rain-fed areas. Finally states could build in a adaptatiodclimate change dimension into the district agricultural plans7 which would go a long way in mainstreaming the climate risk management agenda as well as creating Variants o fbothapproaches are suggestedinWorld Bank 2006f. The National Development Council Resolution o fMay 2007 under the chairmanship o f the Prime Minister o fIndia gives highpriority to incentivising states to develop comprehensive district agricultural plans that will include livestock, fishing, minor irrigation, rural development works and other schemes for water harvesting and conservation. 9 enabling conditions for translating the recommendations o f the report into actions on the ground. D. ClimateChangeandVulnerability to Floods D.1 Consequencesof andResponsesto Floods 31. Floods are a natural feature o f India's river basins. They replenish groundwater, deliver topsoil and nutrients to support agriculture in otherwise infertile regions, and sustain valuable ecosystems. Excessive flooding poses risks to human life and is a major contributor to the poverty and vulnerability o f marginalized communities. It is estimated that the flood-affected area has more than doubled in size from about 5% (19 million hectares) to about 12% (40 million hectares) o f India's geographic area in the past five decades. This has occurred despite generous and rising government spending on a multitude o f flood protectionprograms. 32. Orissa i s among the most flood-affected states in the country. Frequently it has coped with simultaneous droughts in one part o f the state and extensive floods in another, as well as with cyclones and other natural calamities. Two flood-prone districts - Puri and Jagatsinghpur - are the focus o f this study. Households in the study areas endure harsh conditions. Villagers are accustomed to moving homes and losing crops and property because o f flood. Livelihoods and occupations have responded and adjusted to the predictable risks o f flood damage. Rice, which i s among the most flood-resistant o f crops, dominates agriculture. There i s also an emphasis on cultivation in the dry rabi months, when flooding can often improve yields by delivering nutrients and soil moisture. 33. The communities have diversified into a range o f more flood-resilient activities, such as aquaculture, fishing, dairy, and petty business, though the scale o f these activities i s still minor. Despite significant levels o f adaptation, floods continue to disrupt and devastate communities, with the impact bearingdisproportionately upon the poorer segments o f the community (the small farmers and the landless). Even when the poor diversify into nonagricultural activities, they remain more vulnerable to floods. While the nonagricultural income o f the large landholders falls by a meager 5% in a flood, that o f the landless declines by about 14%, reflecting their fragile economic status as unskilled, casual workers. D.11 FutureProspectsfor Flood-ProneAreas under ClimateChange 34. Adding to these already high risks, the climate projections suggest that temperatures, precipitation, and flooding are likely to increase, with adverse impacts on crop yields and farm incomes. Among the more substantial effects is a spatial shift in the pattern o f rainfall towards the already flood-prone coastal areas. Climate change i s also projectedto bring a dramatic increase in the incidence o f flooding. As an example o f the implied magnitudes, the probability that the discharge might exceed 25,000 cubic meters per second (cumecs) (at the measuring station at Naraj on the Mahanadi River in Orissa), i s currently low - about 2%. But under climate change, this is projected to rise dramatically to over 10%. This suggests a clear need for improved and accurate forecasting tools to guide the appropriate location and design for flood protection infrastructure and other high-value assets. 10 35. Inthe studydistricts o fPunand Jagatsinghpur, the assessmentfinds that rice yields could decline by 5% to 12% and profits by 6% to 8% under climate change. With the dominance o f rice and high levels o f preadaptation to floods, there is little that can be done to build flood resilience through adjustments in cropping patterns and farming practices. There i s a need to further strengthen current flood protection initiatives and develop a proactive, comprehensive, and anticipatory flood management strategy. 36. Orissa has established a wide-ranging flood management policy with an emphasis on relief and protection. When floods strike, an elaborate relief machinery provides employment, cash, food, health care, and shelter. There is also a comprehensive action plan that envisions a host o f structural measures such as a cascade o f reservoirs, dams, raising and strengthening o f embankments, and interbasin transfer o f water. Despite the substantial fiscal burden o f the flood relief and protection system floods continue to encumber livelihoods and impede development. With the prospect o f much more severe and intense flooding underclimate change, it is necessary to ask how the system couldbe strengthened to buildgreater flood resilience incommunities. 37. There is no single remedy for mitigating flood damage and site-specific measures are requiredto address particular vulnerabilities. Experience suggests the need to integrate hardandsoft engineering approachesthrough three components: 0 Advanced systems for the detection and forecasting o f floods; 0 Anticipatory and proactive actions designed to minimize flood risks and build capacity to withstand flood events; 0 Reactive actions that deal with the aftermath o f floods and include compensation andrelief. E. Elementsof a Strategyfor BuildingFloodResilience E.1 Strengthening Systems for Detection and ForecastingFloods 38. With the projected future changes inthe spatial distribution, intensity, and frequency o f floods more advanced forecasting and risk diagnostic tools will be needed to guide the location o f high-value investments and the engineering design o f flood protection structures. The system's effectiveness could be enhanced by combining data collection, telemetry, flood forecasting, and flood warnin! elements into one integrated flood management and information system for a basin. Flood inundation mapping i s another important planning tool that i s needed to guide zoning and investment decisions, but its use across India is limited. There is growing recognition in Government o f India that generating such information should be a highpolicy priority. 8As an example the HirakudDam, which is the maincontrol structure onthe Mahanadi, was originally designedfor a flood o f 42,500 cumecs, whereas more recent calculations indicate that the maximum probable flood i s 69,500 cumecs. Floods need to be partially regulated by advance reservoir depletion, which may have impacts on water availability for irrigation, and this inturn calls for a basinwide flood forecasting and management system. 11 E.11 StrengtheningAnticipatory Measures 39. Although technology can help detect and even forecast floods in a timely way, the information needs to be integrated into planningand policy for longer-termmeasures that reduce (a) the magnitude o f the flood; and (b) vulnerability to a flood o f any given magnitude. TheAssault on Floods: Importance of Structural Protection 40. Structural defenses are an indispensable tool in controlling flood damage, but with prohibitive costs and design limits they cannot offer full protection. Economic considerations argue for an emphasis towards the protection o f the higher-value assets (for example urban areas, infrastructure), with greater importance given to building adaptability and flood resilience elsewhere. It i s unlikely that floods can be totally subdued, so careful monitoring and planning o f new settlements in these flood-prone areas must remain a priority for government authorities. This has been recognized in a number o f reportsg. There is a need to strengthen incentives and institutional systems for strategic asset management so that infrastructure investments are rendered more sustainable. This must be accompanied by changes in budget priorities to enhance community resilience to floods. TheAccommodation of Floods: Importance of Nonstructural Resilience Building 41. Building flood resilience in agriculture. Flood-resilient agriculture provides a way to insulate incomes against flood damage. Numerous pilots have been attempted with more rainfall-tolerant or short-duration varieties o f certain crops to minimize flood-related losses. Though economically viable solutions remain elusive, these initiatives have potential and warrant continuing support and dissemination. A further shift in agriculture to the dry (rabi) season could be promoted by increasing access to irrigation in the dry months. 42. Incomediversification.Income diversificationprovides a robust way o fmitigating flood risks. The economic instruments that are relevant for encouraging income diversification for drought management - such as credit and insurance schemes linked to job diversification - are equally pertinent in the context o f floods. The spread o f self-help groups inOrissa provides apotential community base for launching such schemes. 43. Adaptingto floods. There are already numerous and successful pilots inOrissa that aim to promote flood-based livelihoods. This i s the quintessential form o f flood adaptation. With the escalating demand for fish in India, aquaculture has considerable promise for unleashing rural growth. To further develop this potential there i s a need to address supply chain obstacles to improve the marketingo f ahighlyperishable commodity. 44. Primacy of planning and zoning. The combined pressures o f rapid population growth, landscarcity, andintensifying flood risks call for strengthened andmore careful planning and flood zoning. Landuse planningand water management needto be combined into a synthesized plan with coordination between various departments and levels o f These include a World Bank report on the Northeast (World Bank 2007a) and numerous government documents including: PlanningCommission2002; Planning Commission2001; and Shukla 1997. 12 government. A greater challenge i s the implementation o f a plan that would affect many interests and would need processes that involve public participation and stakeholder engagement. F. ConclusionsandRecommendations: 45. With an ambitious growth target o f 8% to 10%for the medium term, the Government o f India recognizes that accelerating the productivity and sustainability o f the agricultural sector will be a prerequisite to achieving its poverty reduction and development goals. The challenges are substantial and will call not only for the familiar investments in agriculture (such as price stability, connectivity, marketability and irrigation), but also for addressing the new and unprecedented risks emerging from climate change. This report demonstrates that climate change will continue to affect the lives andproduction systems o f the millions in India who reside in high-risk rural areas, with a mounting human toll that falls disproportionately upon the poor. Consequently there is an urgent need for action now to avoid higher future costs and missed opportunities associated with a development path that compromises on climate risk management. Fostering a shared vision o f the nature o f climate change and the implications for the country's development prospects will be critical incatalyzing a policy commitment and also helpingto integrate climate risks indevelopment programs over the coming decades. 46. Fortunately, many o f the policy actions required to build resilience to the impending changes in climate are wholly consistent with, and supportive of, current development objectives. Adaptation actions and investments provide a cost-effective way o f addressing future climate risks. India has considerable technical and scientific expertise to understand, analyze andact upon climate risks. There are many encouraging initiatives and policy reforms that are moving in the right direction. These provide an ideal foundation for developing a comprehensive strategy for promoting adaptation to climate change andbuildingsystemic resilience invulnerable communities. The table on the next page summarizes the policy actions andinterventions suggestedbythe study. 13 SummaryRecommendationsfor Adaptation" ExpectedOutcomes Strengthening publicly accessible climate information Baseline information for the integration systems/mechanisms and related management tools to o f climate risks into policy, planning and matchneeds investment decisions Establish a climate information management system at central level for developing climate diagnostic and risk assessment tools with feedback mechanisms to end users. This would include: enhanced data collection systems at local level, hydrogeological data collection, information for groundwater management, and systems for improved detection and forecasting o f floods. Build climate risk assessment as a requirement for all relevant highvalue and long-lived infrastructure projects. Fostering climate-resilient reforms in agriculture and Promote climate resilience o f agriculture water resourcemanagement and reduce risks Promote agricultural research and extension services towards systems and cultivars better suited to local climate and its variability Targeted implementation and development o f basin level water resource management integrating groundwater resources and applying instruments that deliver economic, social and legal incentives to increase water productivity, using participatory approaches where appropriate Supporting the management of climate risks through Promote income andjob diversification to economic mechanisms and instruments that promote reduce climate vulnerability efficiency Explore new innovative financial instruments to promote income diversification, such as 9debt relief instruments coupled with credit for job diversification >debt relief coupled with insurance for new business >risks community-based risk financing schemes Introduce interim smart farm subsidies to encourage switch to more suitable and climate resilient cropping practices ''A number o f recent projects supported by the World Bank have incorporated some elements o f these recommendations in operations. Examples include: the Hydrology I(1995 - 2003), which focused on nine states including Andhra Pradesh, Maharashtra and Orissa), Hydrology I1(2004 - 2010), which currently covers five new states. These include enhanced information systems as project components. The National Agriculture Innovation Project (2006 - 2012) and the National Agriculture Competitiveness Project (under preparation at the time o f writing) recognize and buildinelements o f climate risk. 14 ISummaryRecommendations for Adaptation" I ExpectedOutcomes Improving institutional capacities and linkages among Promote policy synergy, identify needs sectoral programs and fill gaps. Establish capacities and strengthen role o f central bodies inorder to enhancecoordination and operational linkages between departments at all levels o f government. This could include establishing convergence committees for management of drought and flood. Integrate measures targeted towards management o f future climate risks in the planning process including at the local level indistrict agriculture plans." 47. Moving ahead the first priority is to implement and mainstream the actions proposed in the table. Inaddition to this climate change will also have wider impacts that go beyond the flood- and drought- affected areas that are the focus o f this report. Consequently future work will be needed to fill knowledge and policy gaps. Most notably further analytical work i s needed in three priority areas o f highdevelopment significance. First, the effects o f climate change on the rice and grain productionregions o f Indianeedto be assessed in considerably greater detail to determine impacts on food security goals and growth targets. Second, glacial melt remains the most dramatic threat to water supplies, food production and life-sustaining ecosystems inthe country. Furtherwork on the likely consequences remains an urgent priority. Similarly, the threats from sea level rise on coastal communities and cities is another important issue. On the policy front, agricultural trade distortions in the developed world increase climate risks and vulnerability in developing countries, suggesting the need for integrating adaptation and climate change issues inglobal trade negotiations. " GoI's National Development Council ResolutionofMay 2007 gives highpriority to building comprehensive district agricultural plans. 15 1. Introduction,Context,and Objectives 1.1 Introduction 1. Adapting to climate variability and change has long beenpart o f the human development agenda. Countries and dynasties have risen and fallen because o f climate-related issues, particularly those associated with the scarcity o fnatural resources.12However, the current pace o f change in the world's climate is unprecedented in recent human history and experiences accumulated to date are unlikely to be sufficient to provide a clear way forward. The challenges will be particularly acute in developing countries, where large and still growing populations live in high-risk areas or rely on rain-fed lands for their livelihoods. 2. Recognizing the significance o f climate variability for growth and development, the Government o f India has established a range o f programs, policies, and institutions that have reduced the exposure o f rural populations to climate risks. These long-standing programs have done much to unleash the development potential o f agriculture. India has achieved its food security goals, produced a food surplus and developed muchprosperous farming communities. Policies have also helped to build resilience to extreme climate events over the years: when floods or droughts descend, the numerous relief programs provide a safety net for the poor and assure a more rapid recovery. India's disaster management programs rank among the most comprehensive and effective in the world. However, accelerating climate change i s expected to increase the frequency o f extreme climate events inways that are outside the realm o f experience. Therefore, there is a need to address the costs associated with climate change, emphasizing complementary policies and actions that can integrate climate resilience and enhance the sustainability o f rural livelihoods. 3. Itis inthis context that the Government o fIndiahas acknowledged the need to develop a strengthenedinformation base to assist in planning and mitigating the consequences o f increased climate variability and change. India's Ministry o f Environment and Forests i s preparingits second formal communication on climate change, reflecting its commitment to the UnitedNations Framework Convention on Climate Change (UNFCCC).'3 Further, in articulating its Eleventh Five Year Plan, the Planning Commission of India has recognized the need for proactive adaptation, particularly inthe agricultural sector, which stands to be most affected by climate change (Planning Commission 2007). The objective o f this study i s to assist the Government o f India in these endeavors. The study focuses on three particularly vulnerable states - Andhra Pradesh, Maharashtra, and Orissa. The main audience for the study i s therefore the Government o f India and the three state governments involved inthe process. 4. The focus o f this report is on vulnerable rural communities that face current and hture risks from extreme and frequent droughts and floods. The study aims to develop the elements of an adaptation strategy that would assist ina better integration o f climate risk 12See Diamond 2005. l3This is the SecondNational Communication on Climate Change to the UNFCC. 16 management into India's development efforts. Scaling up the study approach and refining the methodological framework presented inthis report, would allow India to assess, and where necessary act upon, the threats and opportunities that result from both existing and future climate variability in vulnerable areas. As a prelude to these issues, the report begins with a brief description o f the current scientific consensus on climate change and the challenges indeveloping an adaptation strategy. 1.2 Context:ClimateVariability, a DevelopmentChallenge 1.2.1 ClimateChangeandthe DevelopmentImpacts 5. The Earth's climate has changed and will continue to change in the foreseeable future, regardless o f potential mitigation actions, with consequent impacts and implications for development and growth. On a global scale, the Earth's climate has warmed, precipitation patterns have altered, sea levels have risen, and most nonpolar mountain glaciers are in retreat. Table 1.1 highlights the trends in global changes in the last decades. These observed changes in climate have, in large measure, been attributed to human activities that have increased atmospheric concentrations o f greenhouse gases through the use o f fossil fuels andlarge-scale changes inlanduse. 6. In the 199Os, climate-related disasters negatively affected about 2 billion people in developing nations, representing about 40% o f the total population in the affected c~untries.'~Looking forward, climate model projections indicate that there will be more frequent warm spells, heavier rainfall insome areas and more frequent droughts inothers, tropical cyclones will gain in intensity, and extreme high tides will be more common. Climate models project that the next two decades will see a warming o f about 0.2"C per decade. Table 1.1 Selected Indicators of Changes in the Earth's Climate in the 20th and early 21st Century Indicator Observed change(s) (typically after 1960) Global mean surface Eleven o f the last 12 years (1995-2006) rank among the warmest years in the temperature record o f global surface temperature. Increasedby an average o f 0.74"C (0.56OC to 0.92OC) duringthe 1906-2005 period. Hot daysheat index Increased in number and frequency over most land areas along with number of hot and warmer nights. Cold daydfrost days Decreased innumber and intensity over most land areas. Heavy precipitation Increased over most land areas. Significantly increased in eastern parts o f North events and SouthAmerica, northern Europe, and northernand central Asia.' Frequency and severity Increased summer drylng and associated incidence o f drought in some areas. In o f droughts parts o fAsia and Africa, the frequency and intensity o f droughts have increased. Global ocean average Increased to depths o f at least 3,000 m (ocean has been absorbing about 80% of temperature the heat addedto the climate system). Global mean sea level Rose at an average rate o f 1.8 mm (1.3 mm to 2.3 mm) per year between 1961 ~~~ 14EM-DAT:The InternationalDisasterDataBase, http://www.em-dat.net/. 17 and 2003. The rate was faster between 1993 and 2003 at about 3.1 &year. The highest contribution to sea level rise has been from retreat o f glaciers and melting o f icecaps. Mountain glaciers1 Declined, on average, inboth hemispheres. snow cover a. This i s a long-term trend observed from 1900-2005. Source: IPCC 2007a. 7. Globally, a significant amount o f modeling and research has been done by experts on the impacts of climate variability and change, with an emerging consensus that climate change will have a negative impact on development patterns and growth potential. The most recent findings of Working Group I1o f the Intergovemmental Panel on Climate Change (IPCC)" reinforce these broad conclusions and express "high confidence" (implying a probability of 90%) that the predictions o f the major climate change models are consistent with observed warming at regional and global scales. Box 1.1provides an overview o fthe key findings o fthe IPCC's Working Group I1report (IPCC 2007b). Box 1.1WorkingGroupI1of IPCC: Summaryof RuralImpacts The report concludes with high confidence (90% probability) that "climate change is projected to impinge on the sustainable development o f most developing countries o f Asia, as it compounds the pressures placed o n natural resources and the environment that are associated with rapid urbanization, industrialization, and economic development". SouthAsia Agriculture/foodsupply Freshwaterresources (concluded with medium to high (concluded with medium to high (concluded with either high or very confidence)a confidence)a highconfidence)a Crop yields could decrease up to At mid- to high latitudes, and By mid-century, while average 30% incentral and South Asia by where local average temperature river runoff and water the mid-21st century. increases anywhere from 1°C to availability at high latitudes and Endemic morbidity and mortality 3"C, crop productivity i s in some wet tropical areas are are expected to rise due to projected to increase slightly projected to increase by 10-40%, diarrheal disease primarily depending on the crop. Past that at mid-latitudes and in the dry associated with floods and point, crops will see declines in tropics they will decrease by 10- droughts. Increases in coastal yield insome regions. 30%. Some o f these dry regions water temperature would At lower latitudes, especially are already water-stressed areas. exacerbate the abundance and seasonally dry and tropical The extent of drought-affected toxicity o f cholera inSouth Asia. regions, crop productivity is areas will likely increase. The largest number o f people projected to decrease for even Heavy precipitation events, affected by sea level rise will be small local temperature increases which are very likely to increase in the heavily populated large (1-2"C), which would increase risko fhunger. infrequency, will augment flood 15 The Intergovemmental Panel on Climate Change (IPCC) was established by the World Meteorological Organization and the United Nations Environment Programme in 1988 to assess scientific, technical, and socioeconomic information relevant for the understanding o f climate change, its potential impacts, and options for adaptation and mitigation. It comprises three Working Groups and a Task Force, which meet regularly to review the globally published scientific/technical literature on climate change and to issue official assessments onthe situation. 18 deltas o f Asia (and Africa). Globally, the potential for food risk. Glacier melt in the Himalayas is production i s projected to go up . In the course o f the century, projected to increase flooding as long as local average water supplies stored in glaciers and rock avalanches from temperature rise does not exceed and snow coverage are projected destabilized slopes, and affect 3°C; above this, it is projected to to decline. This will reduce water water resources within the next decrease. availability in regions supplied two to three decades. This will Increases in the frequency o f by meltwater from major be followed by decreased river droughts and floods are projected . mountain ranges, which are flows as glaciers recede. to affect local production home to more than 1 billion Freshwater availability, negatively, especially in people. particularly in large river basins, subsistence sectors at low Sea level rise is expected to i s projected to decrease, which, latitudes. exacerbate inundation, storm along with population growth surge, erosion, and other coastal and increasing demand arising hazards, thus threatening vital from higher standards o f living, infrastructure, settlements, and could adversely affect more than facilities that support the a billionpeople by 2050. livelihood o f island communities. a. Very high confidence implies that the statement has at least a 9 out o f 10 chance o f being correct; highconfidence implies an 8 out o f 10 chance; and medium confidence a 5 out o f 10 chance. 1.2.2 Local Variability: A Challengefor AdaptationinIndia 8. For India, climate change poses particular challenges due to its vast geographic diversity. Climate projections for India suggest that impacts are likely to be varied and heterogeneous, with some regions experiencing more intense rainfall and flood risks, while others encounter sparser rainfall and prolonged droughts (see chapter 2 for details).I6 Even though a vast body o f knowledge exists on the impacts o f climate variability and change, little i s known about the impacts at the meso (districthubdistrict) and micro (revenue village) levels or about specific adaptation strategies at the local levels that can diminishthe risks associated with climate variability and change. 9. There is growing recognition that the impact o f climate change is highly disaggregated andwill vary across sectors, locations, andpopulations. InIndia, the climate vulnerability o f agriculture and livelihoods depends on a suite o f interconnected factors that vary locally, including (a) physical factors such as differences in soil conditions, water supplies and infrastructure; (b) economic factors such as local prices and technology that determine cropping patterns and farm management strategies; and (c) policies (for example taxes, price support, subsidies) and institutions (figure 1.1). The implication for a country as vast and varied as India is that broad generalizations on ways to promote adaptation will be misleading ifthey neglect local conditions. Consequently, there can be no "one-size-fits-all" approach to developing a climate risk management strategy; approaches will need to betailored to fit local conditions. l6The climate projections are summarized inIndia's official InitialNational Communication on Climate Change to the UNFCCC. 19 Figure 1.1Elementsof Vulnerability to Climatein Rural Areas of India makesit subjectto high growththoughat a IndiaIS denved 61% oflandareais Currentincentives 1.2.3 Preparingfor an Adaptation Strategy 10. Adaptation to climate change i s analogous to many other forms o f risk management. It requires an assessment o f possible threats and opportunities arising from climate variability and, where necessary, incorporating these into policy through the appropriate institutional mechanisms. This calls for local-level information on current and hture climate vulnerabilities to identify risks and potential impacts. However, the provision o f better information isjust the first step inpromoting more climate-resilient outcomes. For an adaptation strategy to be effective, it must result inclimate factors being integrated as a normal part o f policy making and risk management. This requires systems to translate scientific results into practical policy information and an institutional framework that allows for the integration and mainstreaming o f climate risks into policy making where this i s relevant. This report strengthens the knowledge base to facilitate the development o f a strategy to build climate resilience in agriculture and, more broadly, to support India's development pathway. 1.3 Objectives, Process,andApproach 1.3.1 Objectives 11. The overarching objective o f this study is to mainstream and integrate climate risk management into development policy by enhancing the understanding o f climate and 20 climate-related issues in the agricultural sector. The scope o f the study and choice of sector was dictated by government priorities, which indicated the need to assess the climate risks in regions where vulnerability was high and potentially increasing. The recommended focus was on areas facing greater exposure to droughts and floods. The study responds to the following specific needs that were determined inconsultationwith the Government o fIndia: 0 To generate better information and understanding on the current coping and climate risk management strategies inresponse to droughts and floods; 0 To understand government responses and their effectiveness in supporting both coping and adaptation measures; 0 To develop and demonstrate the use o f a climate modeling framework to identify future climate risks inselected areas o f India; 0 To use this information to assist in identifying the key elements o f an adaptation plan that can help improve climate resilience and adaptive capacity in selected areas o f India; 0 To further raise awareness o f the problem and effective solutions among all actors andstakeholders. 1.3.2 Approach 12. To achieve these objectives the study has adopted multiple approaches: (a) First, it gathers information from selected areas o f India on current coping and adaptation measures to learn lessons from the past and present to understand (i) how rural communities cope with and build resilience to extreme climate events; and (ii) identifythe key determinants o f vulnerabilities to climate risks among to farmers. (b) This i s complemented by a review o f governmental programs and institutions and climate-related risk management strategies in the case study states. The review identifies policy and administrative gaps and strengths in addressing climate risks and suggests ways to complement existing approaches to deal more effectively with climate risks. (c) Looking forward, the study builds a modeling framework to assess future climate risks and vulnerabilities inthe study regions. (d) Finally, the study synthesizes the results o f the modeling framework and the other components to suggest a framework for policy options and actions to address adaptation inIndia. 13. The assessment i s focused on regions in the states o f Andhra Pradesh, Maharashtra, and Orissa, where a countrywide vulnerability assessment found climate risks o f droughts or floods to be highand increasing (appendix A). The case study areas are typical o f many parts o f Indian agriculture at the edge of climate tolerance limits, but cannot be representative o f the vast geographic and climatic variability across the entire country. For example, other important areas o f vulnerability include the fragile mountain 21 ecosystems in the Himalayan ranges and the Western Ghats, coastal zones, and urban areas o f India. 1.3.3 Process 14. The primary counterpart in the study has been the Ministry o f Environment and Forests o f the Government o f India. The scope and objectives were developed in close consultation with the Ministry and a wider cross-section o f other concerned ministries and departments in both the central government and in selected states. This included consultation in Andhra Pradesh, Maharashtra, and Orissa with relevant departments. A panel o f scientific experts from academic, policy, and research institutions in India provided guidance on the study approach and acted as a quality filter to review the technical results. This was complemented by extensive discussions and workshops at the state level with various government agencies, culminating in an international conference on adaptation to climate variability and change heldinNew Delhi inDecember 2006 (see appendix C for details). 15. A key feature o f this study has been extensive consultations in the states on the methodology andemergingfindings. The focus states for deepening the engagement were Maharashtra and Orissa, as the Bank's involvement inAndhra Pradesh was recognized to be on strong grounds through a separate pilot initiative on drought adaptation. As a result o f the consultations, the state governments o f Maharashtra and Orissa requested rapid assessments o f government programs focused on climate-related issues in general, and drought and flood relief management in particular, which were supported by the study. The Department o fWater Resources, Government o f Orissa, and the Department o fRural Development and Water Conservation, Government o f Maharashtra, provided hll cooperation and support to these assessments. Further, the study developed a conceptual framework for a drought adaptation program in collaboration with senior officials from the Department o f Rural Development and Water Conservation as an example o f an operational program on adaptation. Appendix F presents the conceptual framework for a drought adaptation initiative. 1.3.4 OperationalContext 16. The report is part o f a wider program o f engagement on climate change beingundertaken by the World Bank. Recognizing the development significance o f promoting adaptation to climate change, a significant portfolio o f activities are planned and under way, especially in Latin America, Africa and China. A bulk o f the portfolio focuses on technical assistance and knowledge management activities. In Africa, pilot projects to mainstream adaptation into the Bank operations are being developed in Burkina Faso, Kenya, and Tanzania, while in China an adaptation project associated with the irrigation sector is under preparation. In Latin America, a pilot project in coastal areas o f some islands in the Caribbean is being implemented and another pilot, which aims to address glacier retreat in the Andean region, is under preparation. This study is most closely related to and extends the modeling framework and approach developed in an earlier study on drought adaptation strategies for Andhra Pradesh, India (World Bank 20060. The Andhra Pradesh study has led to a pilot drought adaptation initiative in Andhra Pradesh, which i s testing innovative approaches related to farming systems and 22 management o f natural resources to promote greater resilience in arid conditions (see chapter 3 for details). 1.3.5 Report Structure 17. This report is arranged in six chapters. Figure 1.2 summarizes the study framework and organization o f the report. Chapter 2 provides a snapshot o f India's current climate and a review of climate modeling projections o f future changes. It also describes the study methodology and design framework used for the analysis. Chapters 3 and 4 present detailed study results for droughts in Andhra Pradesh and Maharashtra, respectively, while chapter 5 presents study results for two flood-prone districts in Orissa. The final chapter summarizes the key findings o f this work and proposes some opportunities for actionto strengthenadaptation measures. 23 2. India's Climate: Background, Trends, and Projection Tools 1. Has India's climate changed? What do climate projections suggest about future shifts in the country's climate and what impacts might these have on agriculture? This chapter attempts to answer these hndamental questions. It begins with an overview o f India's climate, which establishes the baseline against which fbture changes are compared. It proceeds to highlight the possible impacts o f climate change on the agricultural sector, based on the scientific literature. This provides the context for describing the methodological approach developed in this study and its contribution to a greater understandingof climate change. The study methodology is presented, supplemented by a descriptionoftechnical issuesinappendices A and G. 2.1 India's CurrentClimate, ItsVariability, andTrends 2. Temperature changes. Much o f India i s warming. The mean annual surface-air temperature has risen by an average o f 0.4"C in the last 50 years (Rupa Kumar et al. 2002). Figure 2.1 shows the observed changes in mean temperature, which exhibit considerable annual variation, with an upwardtrend. Figure 2.1 All-India Average SurfaceTemperature 1948-1998 "c 1 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 1948 1958 1968 1978 1988 1998 Source: Data fromIndianInstitute o f Tropical Meteorology, http://www.tropmet.res.in/ 3. Variable rainfall patterns. The most important climatic feature o f the Indian subcontinent i s the summer monsoon. As much as 70% o f the annual aggregate precipitation i s received in a short period from June to September during the southwest monsoon. Fragments o f the southeastern states receive rainfall during the winter months. Meteorological records confirm that the monsoon exhibits considerable random (and unexplained) variation, butnevertheless has a relatively stable core (figure 2.2). 25 Figure 2.2 Variation of All-India Monsoon Rainfall 1813-2003 --II I Source: Data up to 2000 from Indian Institute of Tropical Meteorology, http://www.tropmet.res.id. Subsequent data from newspaper reports o f rainfall. 4. Higher frequency of droughts. Because o f the dominance o f the monsoon, India's climate and weather exhibit the heaviest seasonal concentration o f precipitation in the world. Almost 20% o f India's total land area i s drought prone. The frequency o f droughts has beenincreasing over time: there were six droughts between 1900 and 1950 compared to 12 inthe following 50 years, and 3 droughts have already occurred since the beginning o f the 21st ~entury.'~ Escalating levels o f ecological degradation, resulting from such factors as deforestation, receding water tables and overgrazing have increased the vulnerability o f ecosystems to drought. 5. Increased frequency of floods. The Ganges-Brahmaputra and Indus river systems are highly prone to flooding. The magnitude o f flooding has increased in recent decades, from approximately 19 million hectares affected 50 years ago to 40 million hectares in 2003, about 12% o f India's geographic area. Floods have occurred almost every year since 1980, and their extent substantially increased in 2003 due to widespread rains, which affected even some o f the most drought-prone areas. Inrecent years an increase in populationinvulnerable areas, inadequate drainage and deforestation have all contributed to the rise inflood damage. 2.2 FutureClimate ChangeProjections 6. What might the future hold for India's climate? Projecting future climate i s a complex and uncertain exercise. The global models that generate projections have technical limitations, such as the level o f resolution and data inputs. The inherently uncertain nature o f socioeconomic scenarios and responses adds to the challenges o f projecting future greenhouse gas emissions and the associated climate impacts. Nevertheless, projections are needed to assess possible future hazards, risks, and opportunities, without which it would be impossibleto develop a coherent climate risk management program. 7. The most widely quoted projections for climate change in India are those derived from the Hadley Centre regional climate model (HadRM2), which provides simulated l7 EM-DAT:The OFDNCRED InternationalDisaster Database, http://www.em-dat.net, see also Winrock 2005. 26 outcomes for a range o f scenarios. Projections are available for the period 2041-2060 and suggest: ' 0 An increase in average surface temperatures across all seasons, with an increase o f2°C to 4°C south o f latitude 25", and inexcess o f 4°C inthe northernregion; 0 More variable precipitation during the monsoon season, with a possible decrease towards the west and an increase over the IndianOcean andthe Western Ghats; 0 An overall increase inrainfall intensity by 1-4 millimeters per day, accompanied by an increase in the highest one-day rainfall. However, given the country's varied geography, some parts o f northwest India could also witness a decrease in extreme rainfall; 0 Glacial retreat causedbywarming, though the extent remains uncertain; 0 A rise insea level, which would threaten economic assets, coastal cities, and large coast-dwelling populations. 2.3 Methodology of the Study 2.3.1 Approachesto MeasuringClimateChangeImpactson CropYields 8. Given the scale and magnitude o f these changes, there are concerns that climate change could compromise agricultural productivity in India, threaten food security goals and undermine development efforts. The unease is not without reason: a large and growing rural population live inhigh-riskareas, agricultural output and water supplies depend on the monsoons, and much farming i s on rain-fed lands. Consequently there have been numerous attempts to assess possible impacts o f climate variability and change on agriculturalproductivity. 9. Measuring the impacts o f climate change on agriculture is a complex and challenging exercise as there is much uncertainty inmoving from climate change to physical impacts on crops and market responses. Two approaches have evolved to determine crop responses to climate change: statistical methods (termed the Ricardian approach) and agronomic crop growth models. Statistical assessments 10. The statistical approach examines how farmers in a certain climate scenario perform relative to similar farmers indifferent climate scenarios. This generates estimates o f farm performance across different climate zones that can be used to infer the consequences o f climate change. Table 2.1 summarizes the results from the statistical studies o f the impacts o f climate change in India. These suggest that a temperature increase o f 2°C would generate a modest loss of between 3% and 9% o f current agricultural income. But for a 3°C rise intemperature, the studies predict a wide range o f losses from 3% to 26% o f income. These estimates are broad andapproximate averages, based on the assumption that temperature and rainfall impacts are uniformly distributed across the country. Furthermore, while statistical approaches capture current and observed farm-level ** Notethat the dates do not correspond to particular calendar years, but represent probable climate events that are predictedto emerge within that period. 27 responses to different climate regimes, they cannot draw inferences on future events and responses. Table 2.1 Results of RicardianAssessments of ClimateChange Impactson Crops in India Temperature % change in net agricultural change revenue per hectare Source 2°C -3 to -6 Sanghi, Mendelsohn, and Dinar 1998 2°C -7 to -9 Kumar and Parikh 1998 2°C -8 Kumar and Parikh 2001 3.5"C -20 to -26 Kumar and Parikh 1998 3.5"C -3 to -8 Sanghi, Mendelsohn, and Dinar 1998 Agronomic models 11. Agronomic approaches use a crop model calibrated from controlled experiments that are designed to simulate climate events and different management regimes. Typically these models assume that farmers are "myopic", in that they do not respond to predictably changing conditions, nor do they learn from past experiences. As a result, estimates o fthe costs o f adapting to climate change are exaggerated. This i s a weakness in past studies that have used agronomic models to estimate agricultural impacts o f climate change in India. Table 2.2 provides a summary o f the key studies. Despite a wide range o f simulated impacts the results exhibit a consistent pattern o f responses. The assessments show that: 0 Crop yields-are influenced by the interplay o f three key climate parameters: (a) the level o f carbon dioxide (termed carbon fertilization); (b) the temperature change; and (c) the level and distribution o f precipitation. For most crops, elevated levels o f carbon dioxide and higher precipitation rates (except where rainfall is excessive) promote crop growth. Since current temperatures throughout much o f India are high, these beneficial effects are offset by further warming. 0 The overall impact o f climate change on crop yields depends on the baseline conditions o f these parameters and the balance o f these conflicting forces. Inarid locations where crops already suffer heat stress, a small increase in average temperatures can lead to a dramatic decline in yields. The same temperature change ina cooler climate zone couldproduce an increase inyields. 12. The clear implication is that broad generalizations o f crop responses to climate change will be misleading if they do not take account o f location-specific baseline climate and soil conditions. 28 Table 2.2 Agronomic Assessments of Climate Change Impacts on Agriculture in India (by Source andRegion) I Crop I Yield change (%) I Scenario 1 Model Rice 0 +2OC; doubling C02 CERES-Rice -20 +2"C; doubling C02;water shortage Wheat 0 +3"C: doubling CO, CERES-Wheat ILa1et aL 1999:Madhya Pradesh I Soybean I From-4 to 0 I +3"C; doubling CO,; -10% daily rainfall I CROPGRO / Saseendran et aL 2000: Kerala I Rice -6 +1.5"C CERES-Rice +12 I +1.5"C; +2 &day rainfall; 460 ppm C02 Agganval and Mall 2002:parts of northern, eastern,southern, and westernIndia" From+3.5 to +4.3 (2010) Optimistic IPCC scenarios: +0.loC, 416 ppm C02;+0.4"C, From+13.8 to +22.3 (2070) 755 ppm C02.Both at current crop managementlevelb CERES-Rice From+I .3 to +1.9 (2010) Pessimistic IPCC scenarios: +0.3"C, 397 ppm C02;+2"C, From+3.6 to +9 (2070) 605 mmCO,. Both at current croD managementlevel 1.11.. From+5.1 to +7.4 (2010) From+16.6 to +25.7 (2070) Optimistic Ipcc ORYZAIN From+2.5 to +4.1 (2010) Pessimistic IPCC scenarios IKalra From+6.1 to +16.8 (2070) et al. 2007: DEFRA study -5 to -8 +l"C; no change inC02 Rice -10 to-16 +2"C; no change inC02 CERES-Rice -21 to -30 +4"C Maize -10 to-30 +1"C to +4"C; 350ppm C02 CERES-Maize Jowar -7 +1"C CERES- -12 II+2"CC Sorghum I WorldBank 2006f Groundnut +2 Jowar +3 Max. temp.+2"C; min.temp. +4"C; annual rainy days -5%; I Sunflower I +IO ,550ppmC02 1 - 1IMaize +3 EPIC Rice 1I-8 II Groundnut I 0 Max. temp. +2"C; min.temp. +4"C; annual rainy days -5%; Jowar I o 550 ppm CO,; cumulative monsoonrainfall (Jun-Sept) -10% Sunflower +9 Maize 0 a. Margin of error can be as much as 32%, depending on the uncertainty in climate change scenario and other factors. Sensitivity analyses were runfor increasesintemperature level ofnutrients fed to the crops, and variations inC02levels. These showed that, as long as temperature remains unchanged and C02 levels increase, yields will increase; however, with temperature increases, this C02effect is nullified for increasesintemperatureas low as 0.9"C. b. "Current crop managementlevel" assumes no change incurrentnutrientapplication and irrigation. c. Further increases in temperature resulted always in lower yields irrespective of increases in C02. The beneficial effect of additional C02upto 700 ppmwas nullified by an increaseof only 0.9OC. 29 2.3.2 Methodology for Climate Impact Analysis 13. The approach developed in this study incorporates the lessons from these studies and advances the approach to assessing climate change impacts. The agronomic models for India provide a plausible description o f crop responses to climate change. However, they fail to capture adaptive responses to climate change at the farm level. Experience shows that farmers can and do adapt to climate variations, which allows them to moderate some o f the adverse effects. Consequently, crop losses from past studies that have used agronomic models are likely to be overestimated. The framework developed inthis study addresses this problem by linking a farm economic model to the agronomic model to estimate economically prudent responses o f farmers to changing climate conditions. Recognizing the country's enormous heterogeneity and the varied climate impacts, it focuses on selected vulnerable areas. 14. Study areas. Since droughts and floods are predicted to increase in frequency and intensity, the focus is on districts indrought-affected areas o fthe Pennar basin inAndhra Pradesh, the Godavari basin in Maharashtra, and the flood-prone area o f the lower Mahanadi basin in Orissa. The choice o f districts and blocks was determined through a risk mapping exercise (see appendix B), which identified areas where agriculture is already stressed and where current trends indicate worsening conditions. The selection o f villages was based on a purposive sampling approach aimed at capturing the impacts o f irrigation coverage and infrastructure on adaptive capacity". To assure a representative coverage o f the population, household surveys were conducted on samples o f households stratified across landholding categories. This approach differs from the Ricardian studies that compare agricultural productivity across a wide spectrum o f climate zones. The focus here i s on understanding adaptive responses in areas already stressed by droughts or floods. Therefore, the results complement other approaches. 15. Learning from the past and present. The study begins by exploring current coping practices and identifying factors that influence the vulnerability o f farmers to climate risks in the selected study areas. It uses data from household surveys to explore how farmers respond to current climate risks by comparing outcomes in a "normal" year with those in a drought or flood year. The statistical exercise i s useful in illuminating the factors that help communities insulate livelihoods from climate risks and the effectiveness o f different coping mechanisms. Complementing the statistical assessment i s a review o f governmental programs and institutions for managing climate risks. The review identifies the strengths and gaps in addressing climate risks and suggests ways to enhance the synergies and effectiveness o f government programs. 16. Looking to the future. With its focus on the past, the above approach i s limiting, since many of the predicted impacts o f climate change are expected to differ in kind and magnitude from current climate patterns. Accordingly, an integrated modeling system (IMS) was developed to generate climate scenarios and assess the impacts on agriculture. The results o f the I M S exercise, in conjunction with the above assessments, assist in the development o f a framework of appropriate adaptation responses and policies to build climate resilience. The study design has intended to take into account two contrasting districts per state and selected villages accordingly fromthese districts. 30 17. Details of modeling using the IMS. The I M S is based on the two most widely used IPCC scenarios inIndia, termed A2 and B 2 (box 2.1). The former describes a pessimistic world with higher levels o f greenhouse gases and more aggressive climate impacts, while the latter depicts a more benign scenario with milder climate change. The system architecture i s presented in figure 2.3. The system comprises the following linked subcomponents: 0 A regionalclimate model (HadRM3) providesresults for a spatial resolution o f 50 kmx 50 km and generates climate projections from 2070 to 2100 at the regional level; 0 A stochastic weather generator projects these climate impacts to the local (blocklmandal) level; 0 A hydrological water resource model (SWAT 2000) evaluates the effects of climate change on water resources; 0 An agronomic model (EPIC) uses these estimates o f temperature, precipitation, andsoilmoisture andpredicts crop yields; 0 A custom-built farm-level economic model assesses the financial impacts o f I climate change on farmers and determines effective adaptation strategies. Box 2.1 Emission Scenarios Selected for the Study In order to predict future climate change, a projection of how anthropogenic emissions of greenhouse gases (and other constituents) will change in the future is needed. A range o f emission scenarios has been developed by the IPCC). Inthis study, A2 and B2 were selected to runprojections''. The conditions assumed in the scenarios are as follows: (see the table below for a summaryo fthe main assumptions ineach scenario): A2 describes a very heterogeneous world. The underlying theme is self-reliance and preservation of local identities. Fertility patterns across regions converge very slowly, which results in continuously increasing population. Economic development is primarilyregionally oriented andper capita economic growth andtechnological change are more fragmented and slower than inother scenarios. B2 describes a world inwhich the emphasis is on local solutions to economic, social, and environmental sustainability. It is a world with a continuously increasing global population at a rate lower than A2, intermediate levels of economic development, and less rapid and more diverse technological change than in the B 1 and A1 scenarios. While the scenario is also oriented toward environmental protection and social equity, it focuses on local and regional levels. MainAssumptions of Climate Change Scenarios A2 and B2 Population World GDP Annual C02 Cumulative C02 (billion) (lO"12 1990 (gigatonnedyear) (gigatonnesof USD/year) carbon) Scenario 2050 2100 2050 2100 2050 2100 (1990 -2100) A2 11.3 15.1 82 243 17.4 29.1 1,862 B2 9.3 10.4 110 235 11 13.3 1,164 I Source: IPCC 1998. I 18. The climate projections are from the third-generation Hadley Centre model (HadRM3), produced at a 50-kilometer resolution, and are the most recent simulations available for India. The stochastic weather generator filters these coarse projections down to the block level, while preserving the statistical properties o f the regional data. The water resource Being the first of its kindstudy for India for which a methodologyhad to b e developed, it was decided to focus o n the two future scenarios prescribedby IPCC. 31 model (SWAT 2000) i s widely used in India and simulates the hydrological cycle at a daily time step. It projects runoff and seepage at the basin and subbasin levels. These modules feed data input to the agronomic model (EPIC). Operational links are established through which surface water flows, derived using the SWAT model, are used as inputs for soil moisture inthe EPIC crop simulation model. 19. From the wide suite o f agronomic models that have been developed, the EPIC model was chosen because o f its previous use in one o f the study areas (Andhra Pradesh) and presumed suitability for capturing climatic impacts on crop yields in Indian conditions. EPIC i s a field-level simulation model that operates on a daily time step to simulate crop growth and yield.21 For any given set o f management inputs (for example fertilizer, irrigation, tillage, date o f sowing) and climate events, EPIC predicts yields. 20. EPIC outputs interact with the economic model. The development o f a farm-level economic model represents a major innovation o f this study. As in all economic assessments, farmers are presumed rational, favoring more lucrative and safer options over less profitable and riskier alternatives. Box 2.2 provides a brief description o f the structure o fthe economic model and fbll technical details are inappendix H. Figure2.3 I M S SystemArchitecture 2' The crop growth model uses radiation-use efficiency in calculating photosynthetic production o f biomass. The potential is adjusted daily for stress from water, temperature, nutrients (nitrogen and phosphorous), aeration, and radiation. Crop yields are estimated usingthe harvest index concept. Harvest index increases as a nonlinear function o f heat units from zero at the planting stage to the maximum value at maturity. The harvest index may be reduced by high temperature, low solar radiation, or water stress duringcritical crop stages. 32 Box 2.2 Structureof the FarmEconomicModel The economic model replicates the sequence o f farming events incorporated inEPIC. It is based on the following structure: At the start o fthe season a farmer determines the cropping mix and the area devoted to each crop. This decision i s basedon the expectedprofits from each crop, which depends on prices, inputcosts, andyields. The universe o f available crops was determined by resource availability and the data requirements to calibrate the agronomic models. It includes the major crops currently sown in each area.Fanners operate under a series of technical and economic constraints (such as water availability and prices). Once planting decisions have been made, seeding takes place and over time the weather pattern unfolds. As inall economic assessments it i s presumed that farmers maximize their expected payoffs from cultivation, subject to constraints and attitudes to risk (see below). Farmer's respond to the actual weather by adopting management techniques that maximize their payoffs. For instance, indry years it may be necessary to irrigate some crops (rice) more intensively and reduce water allocations for other crops. If this occurs, it will also be necessary to adjust fertilization rates. The EPIC module predicts yields under different management regimes, while the corresponding economic module computes the associated payoffs. Since not all farmers are identical, the analysis distinguishes between three types o f cultivators, depending on the availability o f land. Subsistence farmers are classified as those with landholdings up to 2 hectares. They are driven by subsistence needs and the imperative to survive takes precedence over commercial considerations. This is modeled as a safety-first constraint, where the primary objective i s to earn a threshold amount o f Rs 12,000, which i s the subsistence threshold defined in India's National Sample Survey. Medium farmers have holdings between 2 and 3.5 hectares and large farmers have holdings in excess o f 4.5 hectares. The farmers attempt to maximize the commercialpayoffs fiom farming. There are further complexities inassessinghow individuals mightrespondto climate change. Since planting decisions are taken before actual weather i s observed, expectations and attitudes to risk are important in determining management techniques on the farm. Large farmers have greater assets that can act as a poverty buffer, so may have less to lose if they gamble on the promise o f good rains. Conversely, the poor with fewer assets are more likely to be risk averse, preferring a more secure but lower income, to a risher but higher rate o f return. Hence, attitudes to risk will vary with asset holdings, wealth, and preferences. The model allows for different risk-taking behavior by incorporating a risk aversion parameter usingthe mean-variance method. This implies that more risk-averse farmers incur a greater penalty for choosing options with higher losses (down-side risks), so they select safer strategies even ifthis means sacrificing some income ina good year. The model i s solved by backward induction. The second stage o f cropping techniques i s solves first, yielding optimum payoffs for each crop. Given this information, crop choices and planting areas are determined. Simulations are also performed for different climate expectations formation methods including rational expectations (based on the actual distribution o f yields) and adaptive expectations (based on learning). 33 21. Qualifications. The I M S system provides a powerful tool for developing alternative climate and economic scenarios that generate information on physical responses o f crops and the economic consequences at the farm level. But as with all modeling exercises there are well-known limitations and caveats that need to be recognized when the informationi s used to guide decisions. 22. Scenarios are not forecasts. The model outputs are outcomes o f assumed scenarios rather than precise predictions o f the future. Unforeseen changes may occur in input prices, technology, land, and labor markets. To predict changes in crop prices requires a global model o f demand and supply for each crop and information on the likely trade interventions that might apply in the future. Unfortunately, global crop models generate poor forecasts, so there i s no robust and reliable way o f determining crop price changes. All of this suggests considerable uncertainty in developing projections and the need for caution when generating andinterpreting them22. 23. However, uncertainty need not render the assessments useless and can be addressed in various ways. Two approaches were followed inthis study. First, sensitivity analysis was performed to test the robustness o f various results. This was done for most economic variables (prices, risk aversion parameters, and expectations mechanisms) by testing the limits beyond which the results change significantly. A second more important approach was to create the model in a user-friendly (Excel) interface that would allow other analysts to develop and simulate different economic scenarios. 24. Assumptions about water. The hydrological model used in the study includes a simulation o f the groundwater component inthe hydrological cycle. However, the model makes the assumption that the vegetation cover (including agricultural crops inthe winter (rabi) and summer (kharif) seasons) inregions that are not covered by surface irrigation is maintained, and implicitly relies on groundwater storage. Groundwater abstraction is not explicitly modeled. Therefore, the model i s likely to underestimate the negative impact o f decreasing groundwater availability on the agricultural crop cover, especially in areas where groundwater abstraction exceeds the average annual infiltration. 25. Modeling necessarily entails simplifications that are imposed either by data availability or technical limitations. Inthis context, assumptions relating to the treatment o f irrigation in the model warrant discussion. Water availability for crops in EPIC (the agronomic model) comes from three sources: rainfall as determined by the weather generator; runoff and percolation as simulated by the hydrological model; and an exogenous (predetermined) supply o f irrigation water. The agronomic model does not distinguish between the sources o f irrigation water, such as groundwater or surface water. In the context o f projecting plant growth this is, arguably, a plausible simplification, since crop yields depend on the level o f soil moisture rather than the source from which moisture i s obtained. But in the context o f economics and long-term sustainability, the sources o f irrigation do matter. Overabstraction o f groundwater could jeopardize water availability and thus lead to different cropping patterns and crop yields. These links have not been developed in the model due to the high complexity and scale o f the issues. Accommodating these concerns is beyond the scope o f current modeling capacity and must await further understanding. 22 At a global and national scale climate change will affect production levels and prices. Determining these impacts is beyond the scope of this report andthis is left to future work. 34 26. Assumptions about crop mix: The agronomic analysis is based on the assumption that the crop-mix remains the same and only their proportions may vary under different climate scenarios. This i s a limitation caused by the difficulty and inordinate expense of parameterizing the agronomic model and shared in much o f the literature. It was not possible within the resource envelope to incorporatemore crops inthe analysis and hence the focus was on crops that dominate inthe chosen areas. 27. Needfor informationon climaterisks.As argued inthis chapter, the modelprojections should be interpreted with caution andviewed as indications o f the possible direction and magnitude o f changes, rather than as precise forecasts. Despite these uncertainties, policy makers are compelled to respond to climate risks and must make routine decisions on matters that will be affected by fbture climate events, such as investment in infrastructure. Modeling exercises provide the only known way to generate some o f the information that i s needed to assess risks and policy impacts, and thereby improve decision making. 35 3. Climate Variability and Change: A Case Study in Drought- ProneAndhra Pradesh 3.1 IntroductionandBackground 1. Situated in southern India, Andhra Pradesh i s the fifth largest state inthe country, both in terms o f geography and population. With much o f the state lying in the arid parts o f the Indianpeninsula, the specter o f drought is a recurring threat to rural livelihoods. Surface water resources are limited, with modest scope for fkrther increase in supplies. High levels o f groundwater abstraction, coupled with low and variable rainfall, have led to sharp declines in the water table in most parts o f the state. Consequently, water availability has emerged as one o f the binding constraints on farming systems and associated livelihoods inthe drought-prone areas. 2. The government o f Andhra Pradesh has an impressive array o f programs that tackle drought exposure and deficient rainfall in the state. These range from short-term relief schemes to strategic initiatives that include watershed area programs, crop diversification incentives, and water conservation schemes. Yet the human impacts o f drought continue to be devastating for affected rural communities, suggestingthat the low-hanging fruits o f drought policy have been harvested, leavin more complex and intractable problems that take time to address (World Bank 20060.2 Climate change has raised concerns that the f situation could worsen ifdroughts become more frequent or intense across the state. Thus the need for an assessment o f current and future climate risks and vulnerability in the rural sector has become more urgent. 3. This chapter takes a closer look at how affected communities respond and cope with drought in selected areas o f the state. It uses household surveys to investigate the impacts o f drought, coping mechanisms and to identify the key factors that promote greater drought resilience among households. The chapter then explores the prospective trends in agriculture and incomes under projected future climate change scenarios. It uses the integrated modeling system (IMS, described in chapter 2) to explore how the yields o f key crops and incomes might change under future climate patterns and uncertain water supplies. Finally, it synthesizes the findings to recommend some strategic priorities to address drought adaptation that complement and support the state's development objectives. 3.2 Characteristicsof StudyArea 3.2.1 Climateand Geography 4. The focus o f the study i s on two districts o f Andhra Pradesh - Anantapur and Chittoor. Both districts are arid, especially susceptible to drought, and found to be highly vulnerable to further climate variability because o f the limited adaptive capacity o f most o f its residents, due to low incomes and restricted alternative employment opportunities. 23An overview ofprograms ispresentedlater inthis chapter. 36 box 3.1 Groundwater CrisisLooms Large over Anantapur and Chittoor Districts Iepleting groundwater resources are threatening the livelihoods o f farming communities in inantapur and Chittoor, calling for greater discipline in water management and land use lractices. In the past decade, poor rainfall has lead to higher dependency on groundwater for nigation. According to statistics from the Andhra Pradesh Groundwater Department, a clear .eclining trend in water table depths is observed in both districts (see figures below). Rates o f lotential evaporation and transpiration are almost thrice the normal rainfall, with average roundwater recharge o f 60 millimeters in Anantapur and 80-90 millimeters in Chittoor per mum. On average, only 5-7% o f the rainfall contributes to annual aquifer replenishment, ccording to studies carried out by the National Geophysical ResearchInstitute, Andhra Pradesh. itwo-storied granite-based aquifer system(the shallow aquifer isinweatheredmantleandthe .eeperaquifer infracture zone with closely netted sheetjoints at different depth levels) persists in he area, with a significantly low storage potential. Good monsoons usually replenish these quifers though much below optimum levels, and they sustain for a year or a season. As a result, ven today in some areas, the shallow dug wells in these areas are performing well with the upport o f basin development activities. However, the performance o f the deeper aquifer system 3 more complex as it entirely depends on its interconnectivitywith the shallow aquifer. Duringa rought year or successive drought years, a rapid decline o f groundwater levels and well yields ~ccurs.The usual reaction o f farmers to wells drying up i s to make deeper wells in unweathered ones (bedrock), which are unproductive or, if productive, empty the little static resource within hat season or even much earlier. Thus, indiscriminate construction o f a series o f deeper wells has lecome a key factor leading to declining groundwater levels. Ionthly Average Groundwater Level and fonthly Average Groundwater Level and lainfall inChittoor District to March 2007 .ainfall inAnantapur District to March 2007 *t I I I, I A 1 /k ource: Groundwater Department, Andhra Pradesh. 5. 'he districts are located inthe Rayalseema region of the state, inthe Pennar basin, which xtends over a vast flat area o f 55,123 square kilometers. Cropped areas occupy about ,i50,000 hectares o f land in Anantapur district, most of which has been classified as highly drought vulnerable over the last decade. About 366,000 hectares in Chittoor district are cropped and nearly half o f the total area i s highly vulnerable to drought. Though both districts are drought prone, Chittoor i s less arid, with an average rainfall o f 700-1,000 millimeters per year, compared to an average o f 500-750 millimeters per year 37 for Anantapur, which i s the driest district in the state. The geohydrology o f these arid zones is characterized by granites that have relatively low groundwater recharge rates, estimated at 5-7% o frainfall. Box 3.1 summarizes the status o f groundwater resources in these districts. 3.2.2 Anatomy of the Sample Villages 6. Three villages from each district were chosen for the analysis (figures 3.1 and 3.2). On average, villages in Chittoor are more prosperous, with an average monthly income o f Rs 2,686, while inAnantapur the average monthly income i s Rs 1,73 1.24 Table 3.1 provides a snapshot o f the socioeconomic status o f households. Predictably the large landholders earn higher incomes and also own more productive and consumptive assets. Educational status and accessto health facilities also exhibit some variation with landholding size. 7. High dependence on agriculture. Agriculture is the primary source o f income in the sampled villages in both districts, with little evidence o f income diversification (table 3.2). Large and medium landholders are most heavily dependent on agriculture and obtain approximately 86% o f their incomes from farming. It i s only the small farmers and landless, with few assets, who are forced into wage employment in both the agricultural and nonagricultural sectors. Diversification into other forms o f business or agricultural activities, such as dairying, i s minimal, and remittances are low even among the poorest andmost drought-vulnerable groups. Figure 3.1 LocationMap of Study Villages Figure 3.2 Location Map of Study Villages inChittoor District inAnantapur District G2;9s--= * s , u o . , y ~ s @ usrR)i7cAmu Sm'EtU.(w -eQ!J?.xw%es SRDVWrRKl 8. Limited and unequalaccess to irrigation.Despite low and variable rainfall, over 80% o f sampled households indicate that they have no access to irrigation facilities. Unlike many other parts o fthe country, conjunctive uses o f ground and surface water are rare in these villages and most groundwater i s on farms not covered by surface water.25There are considerable inequities inirrigation supplies: larger landholders have greater access to tubewells, whereas canal irrigation i s concentrated among the medium landholders, and the bulk o fmarginal farmers engage inrain-fed agriculture (figure 3.3). 24The difference is statistically significant using the Welch t-statistic at the 99% level of confidence. 25World Bank 2006freports a similar finding. 38 Table 3.1 SocioeconomicProfile of Sampled Households by Landholding Size Landholdingsize Ownership Ownership of Mean normal (% of total of utility productive monthly Years of Access to health households) assets (YO) assets (YO) income (Rs) schooling facility (YO) Large (14%) 48 59 4,550 10 23 Medium(48%) 33 22 2,428 10 27 Marginal (8%) 24 7 1,479 9 19 Landless (30%) 16 0 1,092 7 10 Source: World Bank calculations based on TEN survey data (2003-04). Table 3.2 Average Share of Sources of Income in Total Income in Sampled Households, 2003-04 Agriculturalincome Nonagriculturalincome Agric. Nonagric. Petty Cultiv. labor Total labor business Dairy Remits. Total Large 82 4 86 12 1 1 1 14" Medium 63 24 87 10 1 1 1 13 arginal 39 34 73 18 5 1 3 27 ss 0 74 74 20 4 0 2 26 a0 I 60 40 20 0 Large Medium Marginal Source: World Bank calculations based on TERI survey data (2003-04). 39 Figure3.4 ImpactofDroughton Income26 (e) 5000 4500 ~?II Mean normalmonthly income 4000 IMeandroughtrrunthlyincome 3500 3000 2500 2000 1500 1000 500 0 Large Medium Marginal Landless I 3.3 Impact of Drought 3.3.1 Effectof Drought on Households 9. The survey compares outcomes in a normal year with a drought year.27When drought descends, all households are heavily impacted and face a dramatic decline in their incomes (figure 3.4). Large landholders, being more dependent on agriculture, experience the sharpest (70%) declines intheir incomes. 10. Vulnerability to climate shocks i s multifaceted and goes beyond income volatility. The ability to stabilize incomes is one, albeit important, dimension o f vulnerability that i s emphasized ineconomic assessments. But it fails to capture many o f the broader impacts o f climate shocks on welfare, including poverty, social disruption, decline in health, dislocation o f local markets, and the disruption o f public services. Households clustered along the poverty line are most susceptible to hardship and destitution from climate change. Figure 3.4 shows that the incomes o f smaller farmers plunge closer to the poverty line during drought. Incontrast the large farmers, despite suffering a greater relative drop intheir incomes, are stillbetter off due to their higher income levels. This calls for more targeted efforts to build adaptive resilience among smaller farmers. Apart from declining incomes, villages reported that droughts increased health expenditures (12% o f the sampledpopulation), led to disruption o f education (lo%), forced migration to towns in search of alternative sources o f employment (17%), and led to distressed sales o f land, cattle, andjewelry (34%). 3.3.2 Copingwith Drought 11. Household responses to drought have been largely reactive and do little to build long- term drought resilience. Credit remains the most common coping response to drought (68% of sampled households). Large landholders borrow from formal sources (such as 26The poverty line inruralAndhra Pradeshis Rs.263 (Government ofIndia2001). Source: World Bank calculations on the TEN survey data, drought year (2002-03) and normal year (2003-04) 27The survey was for the recent drought year in 2002. 40 banks), while the landless and small farmers borrow from moneylenders at inflated interest rates. Occupational shifts are another typical response to drought (28% o f sampled households). Poorer households (small and landless) move into unskilled, casual employment (construction, mining, and quarrying), while the large and medium landholders shift to temporary salaried employment (clerical jobs and small business). Few households appear to have altered cropping patterns inresponse to drought and there i s a dominance o f groundnut across the districts, reflecting a high degree o f adjustment to aridagroclimatic conditions. Vulnerability to DroughtandIts Determinants In sum, when drought hits, households inthese villages suffer a precipitous decline in income. But a closer look at the impacts shows that some are better able to smooth income fluctuations than others, irrespectiveo f landholdings or wealth. What explains the greater resilience o f some households to drought over others? To unravel the determinants o f vulnerability and identify connections between factors, a detailed statistical assessmentwas performedusing regression techniques. Appendix Greports the technical results. Here the focus i s on discussing the results and presenting the implications. Definitions o f vulnerability abound and encompass a variety o f dimensions such as social norms, customs, and intrinsic abilities - factors that are difficult to measure.28This study adopts a more pragmatic approach and defines vulnerability in economic terms, as a measure o f income volatility that i s captured through a statistic termed the coefficient o f variation. Figure 3.5 summarizes the linkages and pathways through which vulnerability i s either amplified or diminished. The analysis identifies links with economic factors (debt and occupational mobility), the resourcebase and its management (crop mix and water availability), and certain initial conditions (landholdings and location). A brief overview o f factors with important policy implications follows. Figure3.5 DeterminantsofIncomeVolatility Indebtedness infrastructure 41 Water Resources and Use 0 The assessment finds that a greater dependence on water-intensive crops increases vulnerability. This reflects a familiar trade-off confronting farmers in areas with unreliable rainfall: gambling on a good monsoon, with water-intensive crops, will pay dividends ifthere i s adequate rainfall. But when drought hits, this also exposes the farmer to greater risks and a deeper fall inincome. Consequently income volatility i s higher among those farmers who rely on water-intensive crops. 0 Irrigation access has a large impact on income levels in these arid villages. Consistent with other evidence, the sources o f irrigation are also important. Households with access to groundwater typically have higher incomes.*' But with anemic recharge rates and high levels o f abstraction, droughts deplete groundwater reserves, leaving households exposed to the vagaries o f rainfall (box 3.1 and table 3.3). Consequently, existing irrigation facilities provide little buffer against drought. Inpolicy terms there needs to be strengthened water management that brings abstraction in line with recharge. This would have to include consideration o f changes in cropping patterns, irrigation techniques and job diversification. Table 3.3 Percentageof Householdswith IrrigationAccess Normal year Drought year Tubewell Canal Tubewell Canal Large 22 11 5 1 Medium 7 16 0 0 Marginal 9 9 0 0 Landless 0 1 0 0 Source; World Bank calculations based on TEN survey data, drought year (2002-03) and normal year (2003-04). Economic Factors 0 Households that are heavily indebted also experience a greater drop in drought year incomes. Evidently, high levels o f debt lock households into agriculture and this lack of flexibility inhibits the poor from seeking remedies that lower their exposure to drought risk. This needs to be tackled through a range o f economic instruments that address the root causes o f indebtedness, including climate-related risks. Appendix A presents a summary of the numerous definitions and approaches that have been used to assess vulnerability. 29 Inanormal year (controlling for other factors), households with accessto groundwater earnnearly 50%morethan those without any irrigation access and those with access to canal irrigation earn 15% more than those without any irrigation. 42 0 The ability to diversify income out of agriculture i s an effective strategy for insulating incomes against drought. The analysis finds that there are two key factors that promote income diversification: infrastructureand education. o Infrastructure stimulates broader regional development, thus permitting greater income diversification, which reduces income volatility. o Education builds human capital and hence provides additional earning capacity to households, as well as the ability to cope with drought. The implication is that these interventions bring additional cobenefits by building climate resilience. But these investments have a long lead time, so there is a need to develop tools to identify risks and create systems to assure proactive interventions invulnerable areas3' 3.4 FutureProspectsunder Climate Change 14. The assessment so far has focused on the past and has identified the determinants o f vulnerability and the common coping responses to drought. Looking forward, many o f the predicted effects o f climate change are expected to differ inkind and magnitudefrom current climate patterns. So history may offer only limited guidance on how to mitigate or prepare for these future risks. Accordingly this section uses the I M S to explore future climate scenarios and its impact on agricultural outputs and incomes. It begins with a briefdescriptiono f climate projections and the resulting crop responses inthe study area. It then uses the economic model to ask how these changes might influence farming behavior, cropping patterns, and farm incomes. The following section explores possible strategic interventions that could help ameliorate some o fthe adverse consequences. 3.4.1 Projectionsof Climate Change 15. Results are presented for two climate scenarios: A2 and B2 and these are compared to a baseline that describes 30 weather events from 1961 to 1990.31 The system generates predictions for the Pennar basin in Andhra Pradesh, which includes the districts o f Anantapur and Chittoor. Lookingto the future the climate model projects: e An average increase in annual precipitation o f about 8% (to approximately 709 millimeters) in the A2 scenario and 4% (to about 683 millimeters) in the B2 scenario. 0 This is accompanied by increases in both the annual and kharif minimum and maximum temperatures. The average increase in the former could range from 2.5"C (B2) to 3.4"C (M),while the latter could range from 2.3 "C (B2) to 3.1"C (W. 0 There is also greater variability o f rainfall, both within and between years, and this is reflected in more erratic runoff with significantly reduced flows in dry years. Importantly, this greater variability implies a greater frequency o f low rainfall years that would be categorized as drought events under current criteria. 30Other controls include landholding size, village variables, etc. as detailed inthe tables inAppendix G. 31A 2 and B 2 are the IPCC scenarios. Projections are from 2071 to 2100. A 2 defines a world of high and rising greenhouse gas levels with correspondingly more severe climate change. The B 2 world i s one with a greater emphasis on sustainability and involves lower emissions (see box 2.2, chapter 2). 43 0 The pattern o f rainfall i s also expected to change: the traditionally wetter months o f June and July are expected to receive less rainfall, while precipitation would increase by a smaller amount in the drier months of May, September, and October. 0 A change inthe spatial distribution o frainfall is also projected. Figure 3.6 depicts projected changes inaverage annual precipitation inthe basin. It shows that at the basin level the A2 scenario projects rainfall increases in segments of the basin across 50% o f the land area.32In B2 the changes are more concentrated in the southwest, the north and northwest, and the southeast. The areas that experience a decline inrainfall are inthe north, northwest, and a small portion o fthe southeast. The district-level maps in appendix H show that spatial variation also occurs withindistricts. Figure 3.6 Spatial Distribution of Average Annual Rainfall in the Pennar Basin (Baseline and Climate Change Scenarios) - RIVER - STATE EOUNDARY 32InA2, some parts ofthe basintowards the northwest andthe southeastern boundary, as well as some segments in the central part and the western outer boundary, will see an increase inrainfall. Only one small segment inthe north will experience a decline. 44 3.4.2 CropResponsesto ClimateChange 16. With projections ofthis magnitude, there will inevitably be variations incrop yields. The focus i s on the impact on three major crops: rice, groundnut, and jowar, which collectively account for approximately 70% o f total arable output inthese districts. Jowar and groundnut are dryland crops that are well suited to arid conditions and prolonged heat. Though rice is a water-intensive crop, field surveys indicate that farmers devote a small portion o f their land (about 0.5 hectares) to rice in order to meet basic household consumptionneeds and fodder requirement^.^^ 17. Deteriorating agroclimatic conditions under climate change are reflected in lower yields. The model finds that despite slightly higher rainfall the average yields o f all three crops decline. The reduction is more pronounced under the severe conditions o f A2. Figure 3.7 presents results for one block, Talupula in the district o f Anantapur, with similar trends elsewhere. Groundnut yields fall dramatically inA2 by 28% and more modestlyinB 2 by 6%. Jowar exhibits greater resilience with a decline o f 4% in A2 and 2% in B2, while average rice yields decline by 10% and 4% in A2 and B 2 re~pectively.~~ crop The responses reflect the complex interplay of three key climate parameters: (a) changes in the level and distribution o f rainfall; (b) higher temperatures; and (c) the elevated levels o f carbon dioxide. Box 3.2 uses groundnut yields in Anantapur district, Pennar basin, to illustrate these interactions. Figure 3.7 Yields of Rice, Groundnut, and Jowar under Different Climate Scenarios,Anantapur IBaseline 2 I 5 1 0 5 3.4.3 EconomicAssessment 18. How might these changes influence farm incomes, crop choices, and farm management regimes? To answer these questions the economic model compares the baseline results (under prevailing conditions), with the A2 and B 2 scenarios. In the baseline, groundnut dominates interms o f average profitability per hectare, followed byjowar andrice (figure 3.8a).35 The economic model estimates the resulting cropping mix, taking account o f a 33 This is imposed as a constraint inthe model, details are inAppendix G. 34 Average yields mask changes inthe distribution o f crop responses. These are reported inappendix G. Specifically, under climate change the peaks o f the distributions are lower, while the tails are thicker. This implies that adverse climate events that (a) generate low yields (bad outcome) become more frequent; (b) the intermediate outcomes are less common; and (c) there i s a small increase inthe frequency o fbeneficial outcomes. 35Profits from each crop reflect: (a) the yield, (b) the price of the crop and (c) the production costs. Consequently there is no simple one-to-one mapping between yields andper hectare profits. A crop with higher yields may be less profitable if it either sells for less, or costs more to produce. 45 host of factors including farm size, the volatility of returns, expectations of climate events, and water availability. Prices are based on those that currently prevail and, where relevant, include the Minimum Support Proce (MSP). These prices are varied in the simulations. Figure 3.8b shows the predicted cropping pattern in Talupula (a similar pattern holds elsewhere and is not reported). 19. The model finds that in the baseline, irrespective of farm size, the bulk of the cropped area i s devoted to groundnut. This confirms empirical data that also shows that monocropping o f groundnut i s common practice in many mandals in Anantapur. Not surprisingly farm management methods have evolved to cope with these harsh arid conditions. However, studies on cropping systems also identify risks with monocropping associated with pest attacks and fertility loss. Box3.2 IllustrationofTrends inCropResponsesinthe PennarBasin,AndhraPradesh Crop yields are sensitive to a host o f weather- and climate-related factors including temperature, the level and distribution o f rainfall, and COz concentration. Higher global levels o f C02concentrations are associated with more aggressive temperature increases. As an illustration, the figure below tracks the combined impacts o f temperature and carbon fertilization on groundnut yields - holding rainfall constant at baseline levels - in the Pennar basin. The simulations are based on higher levels o f C02 accompanying the higher temperatures. The combination o f temperature and C02 values are drawn from the HadRM model projections. Groundnut yield increases with temperature from the baseline level up to +l"C, after which the yield starts declining. Underlying thispattern is the assumedrelationship between the level o f C02concentration and the temperature and also their interaction impact: C02fertilization has a positive impact, while a rise in temperature has a negative impact on yields. Groundnut yield also depends on the level and pattern (distribution) of rainfall. If the pattern o f rainfall i s unchanged then higher precipitation will shift the yield curve upward, while less rainfall will shift it downward. Moreover, the change o f distribution o f rainfall will create a horizontal shift in yield projection. All in all, this illustrates that yield projections are highly sensitive to the set o f assumptions underpinning different climate scenarios. These results help explain the differences in yield projections between a previous study (World Bank 20060 and this study. The earlier study examined the consequences o f less aggressive climate scenarios in an earlier period, which placed projections along the rising portion o f the yield curve. The projections in this study are based on warming o f 2-3"C, which i s associated with declining yields. Yield-Temperature Response:EPICModelPrediction 180 - 1 70 - 160 - 150 - 140 - 1 30- 120 7 46 Figure 3.8a Per Hectare Average Profits in Baseline Figure 3.8b Area Allocation in Baseline Scenario, Scenario, Talupula Block, Anantapur T i u p u l a Block, Anantapur ----I 7000 6000 5000 4000 3000 2000 1000 0 Rce Groundnut Jowar SmallFarm WdiumFarm Large Farm 3.4.4 ProjectedConsequences of ClimateChange 20. With all else equal, under climate change, groundnut still remains by far the most lucrative (profitable) crop (Figure 3.9a). Consequently there is no change inthe planting mix (when choosing between the three dominant crops) that would generate more lucrative returns underthe climate change scenarios. 21. With declining average yields the overall profitability o f agriculture declines. The greatest reduction i s projected for A2 (approximately 20%) and reflects the impact on yields. There i s a more modest and tolerable decline o f about 5% in B2. With returns from farming falling (figure 3.9b), the income from agriculture may not be sufficient to sustain the marginal farmers who comprise a large segment o f the rural population clustered along the poverty threshold. This finding confirms the conclusion o f an earlier World Bank report, which highlights the distributional impacts o f drought on the poor and marginal farmers (World Bank 20060. Figure 3.9a Per Hectare Profits in Climate Change Figure 3.9b Total Profits in Climate Change Scenarios, Anantapur Scenarios, Anantapur (Thousand Rs) 35,- ___ 7000 . 6000 5000 4000 I 3000 2OW 1000 0 I Rice Groundnut .bw ai I Smll Farm MedwmFarm LargeFarm I 47 3.4.5 Extensions, Validation, and Sensitivity of Results 22. It is important to emphasize that these results are based on a host o f assumptions about prices, market structure, and crop availability. Inthe future, more profitable possibilities for agriculture may emerge, including the creation o f drought-resistant crops, alternative crop mixes, and new farm management techniques. The objective o f this assessment is not to predict the future, but to identify and isolate the potentia1 effects o f climate impacts under particular scenarios. The framework presented here is flexible and can be used to explore a range o fpolicy issues and scenarios. 23. Validation of results. An important test o f any analytical tool lies in the ability to capture reality. Figure 3.10 compares the predicted outcome in the baseline scenario against the observed crop mix. It demonstrates a remarkable consistency between the model projections and the actual cropping mix and provides some confidence in the model structure and calibration. Figure 3.10 Acreage Comparison between Projections of the Farm Economic Model and Data from Household Surveys 90 80 70 60 50 40 30 20 10 1 0 Rice I Groundnut Rice I Groundnut Anantapur Chitt oor Area C r a m 24. Sensitivity of results. The conclusions presented here simply illustrate L e possible magnitudes o f climate impacts under prevailing conditions and assumed scenarios. Predicting futureprices and technologies several decades ahead i s an impossible exercise. Nevertheless it i s still necessary to determine the robustness and sensitivity o f results to key assumptions. Appendix G summarizes a host o f simulations that test the limits beyond which the key predicted results change. These show that the predicted cropping patterns are robust to significant changes in several parameters, including a variation in the risk aversion parameter inthe range o f plausible values (from 0-3) and a 20% to 25% change inthe price o f rice. Results are also presented for the effects o f increasing levels o f water stress on farms. These show that with less imgation water available at the farm level there i s a predictable decline in cropped area and lower incomes, but this i s accompanied by a gradual shift in cropping patterns (towards jowar, which is more heat resilient). Simulations were also conducted for hypothetical changes in water charges. Unsurprisingly since farmers have pre-adapted to arid conditions, higher water charges simplylowers profitsbutleadto no change inthe cropping mix. 48 3.5 PullingTogether the Pieces:PolicyImplications 25. I s there a need for additional policy to promote adaptation? The ability to cope with adverse climate events depends on the level of wealth and the economic and social infrastructure of communities. So as Andhra Pradesh grows more prosperous, it will generate more job opportunities and inevitably build greater immunity to climate fluctuations. The myriad government programs that deal with education, infrastructure, andjob creation also serve a complementary objective ofbuildingclimate risk resilience. In this context, adaptation policy can be viewed as an adjunct to good development policies that promote equitable growth. All of this might suggest that adaptation to climate change requires no additional policy priority or interventions. 26. Inspite ofthis, there are highrisks associated with complacencythat could magnifythe costs of climate change. Though the projections in this report are broad, they suggest a considerable and mountinghuman toll from climate change and highlight the need and urgency for mitigatingthe avoidable costs, particularly among the vulnerable sections of society. Box 3.3 Andhra Pradesh Drought Adaptation Initiative: Putting Adaptation into Practice A Drought Adaptation Initiativepilot isbeing implementedinselected villages o fMahabubnagar and Anantapur districts by the Society for the Elimination o f Rural Poverty and the Watershed Support Services and Activities Network, in collaboration with district collectors in the pilot districts and under the oversight o f the state Department o f Rural Development. A state government interdepartmental steering committee and a convergence committee comprising the commissioner for rural development and National Rural Employment Guarantee Scheme representatives have been set up to oversee project implementation. The initiative will be implemented over a period o f three years with technical assistance from the World Bank and seeks to (a) identify gaps and missinglinks inthe ongoing drought-related programs and activities in Andhra Pradesh; (b) facilitate institutional integration at state, district, and community levels for delivering drought-related assistance; (c) design and test the innovative methods and instruments for helping selected communities to adapt to drought, targeting different groups within these communities; and (d) improve awareness o f drought adaptation options and approaches and disseminate the results o f the pilot efforts to build support and demand for wider replication. The Drought Adaptation Initiative pilot focuses its resources on four areas o f interventions: (a) management o f common natural resources, dealing with pro-poor water resource (particular groundwater) and common land management; (b) production systems, focusing on diversification and intensification in agnculture, livestock, and horticulture, with technology innovation; (c) economic instruments and marketing, with a focus on improved access to markets, credit, and insurance for new and innovative activities specifically designed for drought adaptation; and (d) institutional support and capacity building, with a focus on institutional strengthening o f farmers and other villager organizations, including such community-based organizations as self-help groups, watershed committees, and credit committees. Pending successful outcomes, the pilot i s expected to build support and demand for wider replication in Andhra Pradesh and provide lessons to other semiarid states. 27. The Government of Andhra Pradesh assigns a high priority and commitment to strengthening development outcomes, as indicated by its support to robust relief machinery and recent strategies to buildlong-term resilience to climate risks among rural 49 communities. Under the oversight o f the Department o f Rural Development the state i s implementing an innovative drought adaptation pilot initiative indryland areas (box 3.3). In addition, there are a number o f sectoral programs in irrigation, agriculture, water conservation, rural development, and forestry that provide a comprehensive and varied platform for strengthening adaptation outcomes on the basis o f more effective synergy and coordination (box 3.4). This foundation can be considerably strengthened through additional and complementary initiatives and approaches that would strengthen and harmonize many o fthe existing initiatives. 28. The focus would be on four priority areas: (a) diagnostic risk assessmenttools to generate information for integrating climate risks inpolicy; (b) management o f natural resources - water and agriculture; (c) management o f climate risks through economic instruments; and (d) institutional changes to manage climate risks. Box 3.4 Andhra Pradesh Sectoral Programs: Comprehensive Base to Build AdaptationApproaches Andhra Pradesh has several long-standing governmental programs are currently in operation that address parts o f a comprehensive drought adaptation strategy. These government programs are briefly reviewedbelow. 0 Water resources programs. Apart from a number o f medium irrigation projects that are planned or under way, numerous community tanks (small surface water reservoirs) are being revived and expanded to improve tank system-based livelihoods through the Department o f Irrigation. In addition, the state launched a micro-irrigation scheme to promote the use o f drips, sprinklers, andrainguns on farmer fields. 0 Watershed programs. Centrally sponsored watershed programs in drought-prone areas are supplemented by state-sponsored programs, which are sometimes donor assisted. State government watershed programs include the Andhra Pradesh Rural Livelihoods Program, assisted by the Department for International Development, Government o f U.K.; the Integrated Watershed Development Program; and the Employment Guarantee Scheme, which uses relief-time labor to construct soil and water conservation structures. 0 Agriculture and forestry. Special programs on crop diversification and sustainable dryland farming combined with incentives (for example seed subsidies) are underway. The state i s considering incentive mechanisms for farmers that adopt the system o f rice intensification practices to encourage wider adoption o f this water-saving technology, especially in the rabi season. The vana samrakshana samitis in Andhra Pradesh, as part o f the Government o f India's forestry programs, promote community participation inprotecting existing forests and new plantations, supportedby bans on open grazing and promotion o f stall feeding. 0 Rural livelihoods. Swarnajayanti Gram Swarozgar Yojana, the centrally sponsored self- employment program for nonfarm-based livelihoods, i s complemented by state government programs that support rural livelihoods, including the Andhra Pradesh Rural Livelihoods Programand the Indira Kranti Patham. Both are statewide poverty reduction projects aimed at buildingstrong institutions for the poor and enhancing their livelihood opportunities through community investment funds. Through these programs, the feasibility o f strengthening the employment components in the National Rural Employment Guarantee Scheme through the inclusion o f innovative activities (for example fodder production, farm composting, and tank desiltation) inrain-fed farming systems i s being explored. 3.5.1 Informationon Vulnerabilities andRisks 29. The impacts of climate change are expected to be heterogeneous and spatially variable, suggesting the need for tools to identify local vulnerabilities, potential impacts, and risks. A recent Bank assessment has also emphasized the need for such assessments (World 50 Bank 20060. While there are many uncertainties in projecting future changes, the diagnostic information can provide an indication o f their direction and magnitude. The development o f such information systems is a first step and common challenge for generating public policy innew anduncertain fields. This study suggests two immediate andurgentareaswhere diagnostic information is required: e Tools are needed to identify areas facing increased risks and so build synergies with existing programs and target more effective assistance to vulnerable communities, most notably with regard to investment in local public goods that buildclimate resilience. e The economic life o f many long-term infrastructure investments will be affected by climate risks that need to be assessed using diagnostic tools. This has important ramifications for assets whose economic lives could be affected by climate change. The location, construction, and refurbishment o f these will need to incorporate potential climate-related risks. This is particularly important for Andhra Pradesh, where investments in irrigation can contribute to protecting agriculture against drought. 30. To utilize this information effectively it must be integrated into routine policy decisions, when appropriate. This is a particular challenge that calls for substantial institutional and capacity strengthening, issues that are addressedlater inthis report. 3.5.2 FinancialandEconomicInstrumentsto PromoteDroughtResilience Debt Relief to Facilitate Income Diversification 31. Inareaswhere natural productivity is low and agriculture is at the edge o fclimate limits, income diversification remains the most obvious and effective way o f reducing the level o f exposure to climate risks. This study has identified indebtedness as an impediment to occupational mobility. Small and poorer farmers who are least able to cope with adverse climate shocks respond to the lack o f formal credit by turning to moneylenders. High levels o f debt tend to lock households into agriculture and inhibit occupational flexibility. With limited assets and accessto start-up capital these groups confront the most systemic climate risks. So an important priority and challenge for policy i s to find innovative and cost-effective ways o freachingthese poorer farmers to help reduce their risk exposure. 32. Coupling debt reliefwith new risk mitigation instruments i s an obvious way to prevent a debt-induced poverty trap. To prevent debt-induced poverty and occupational lock-in, policy must tackle the root cause o f the problem - an overreliance on rainfall-dependent sources o f income. To reduce exposure to climate risks, debt relief or subsidized credit could be linked to other incentives that promote occupational mobility and lower dependence on agriculture. Two innovations merit further policy consideration and scrutiny: e The relief o f old debt could be coupled with the provision o f capital for a new business. This would simultaneously reduce indebtedness and lower the transaction costs o f occupational shifts byproviding new opportunities. 51 0 A variant of this approach would have debt reliefcoupledwith insurance to cover the initial risks of shifting from farming to other businesses and provide protection against new and unfamiliar sources of risk.36 Box 3.5 Weather-IndexedInsurancefor Agriculture inIndia The insurance company ICICILombard, incollaborationwith the Hyderabad-based microfinance institutionBASIX, piloted a rainfall-indexed insurance to protect farmers from drought duringthe groundnut and castor growing season. This was the first weather insurance initiative in the developing world. It was sold to 230 farmers, mostly small, in Mahabubnagar district, Andhra Pradesh, in2003. In2004, the programwas significantly modified interms o f geography, product design, and scope, and was further improved in 2005 by adding new features recommended by farmers. Within three years, the small pilot has graduated into a large-scale operation in which 7,685 policies were sold in 36 locations in six states. Similar products are also being offered by the Agncultural Insurance Company o f India, and the scheme has achieved wide acceptance among the farmers. Weather-indexed insurance i s less susceptible to the problems intrinsic to traditional multiperil crop insurance. The publicly available weather indicators are easily measuredand transparent and the automatic trigger and low-cost weather-monitoring stations reduce insurer's administrative costs, which inturn makes products more affordable to farmers. Moreover, the exogenous nature o f the weather indicators helpsprevent both adverse selection and moral hazards. A major challenge indesigning weather-indexed insurance is minimizing basis risk: the potential mismatch between payouts and actual losses. Since indemnities are triggered by weather variables, policyholdersmay experience yield loss in specific locations and not receive payments. Some farmers may be paid without losses. The effectiveness depends on how well farm yield losses are capturedby the index used. Weather insurance contracts essentially trade o f f basis risks for transaction costs, and the insurance will not be attractive ifthe basis risk becomes too high. A low correlation between yield and rainfall projectedby the EPIC agronomic model for the study districts suggests that the implementation o f rainfall index insurance may encounter future difficulties. Sources: World Bank 2005b; Skees, Hazell, andMiranda 1999; Hess 2003. Rainfall Insurance: A Challenge 33. Crop insurance schemes are frequently promoted as a cost-effective way o f reducing and pooling climate risks and, when linked to particular cropping systems, can be used to promote adaptation. A variety of crop insurance programs are available across the country, but there i s an emerging consensus that these have failed to adequately protect many vulnerable sections o f society. This has led to the creation of a new generation of weather-indexed insurance products (box 3.5). A common feature o f these instruments i s that payment to the farmer is triggered when rainfall falls below some prespecified limit. A clear advantage is that weather indicators are observable, verifiable, and transparent, makingfalse claims less likely. 34. The framework developed in this study can b e used to assess the feasibility and potential impacts of various insurance schemes on cropping patterns and incomes. A prior World Bank study identified the high and rising costs of drought insurance as a major 36Variants ofbothapproaches are suggestedinWorld Bank 2006f 52 impediment to scaling up (World Bank 20060. This study suggests a further challenge that could limit the effectiveness o f such schemes. 35. Weather-indexed insurance will only help mitigate risks and smooth incomes ifthere is a stable correlation between the selected weather indicator and crop yields. Figure 3.11 plots rainfall against crop yields in the A2 scenario. It shows that there i s no simple conelation between these variables.37 Consequently rainfall-based insurance would trigger payments in years when incomes are high and withhold payment in years when yields and incomes decline. While this conclusion is preliminary, it cautions against the use o f insurance to pool climate risks. A greater concern is that even when climate risk insurance helps appease suffering, it may prolong agricultural dependence in situations that call for diversification into less water-reliant forms o f economic activity. Figure3.11 Yield-RainfallCorrelationinAnantapur inA2 Scenario /Yield(tlha) 1.6 1.4 1.2 1 0.8 0.6 0.4 y=O.OOO4x+O.6616 0.2 R2= 0.0726 0 4 I 100 200 300 400 500 600 700 800 900 1000 Kharlf Rainfall(mm) 3.5.3 ManagementofNaturalResources andAgriculture 36. Water conservationand management.The government o f Andhra Pradesh recognizes that there i s an overwhelming case for more aggressively pursuing water conservation across the state, as evidenced by its commitment and support to the Andhra Pradesh Water and LandTrees Act. The projections indicate that even when farmers have largely adapted to arid cropping patterns, increased demand and consequent water stress could severely jeopardize livelihoods and render agriculture less viable in these regions. But there are no easy solutions. Community-based strategies to address groundwater overconsumption face significant challenges as customary rights have at times created strong disincentives for sustainable collective management. But this does not mean that these strategies are without benefits and potential. Greater attention should be given to processes that increase the efficiency o f groundwater use; support the adaptation o f households, communities, and regions to less water-intensive forms of livelihood; and increase the effectiveness o f watershed activities to conserve soil moisture and harvest rainwater. Such adaptive measures are not a substitute for the much needed water policy 37 Iti s recognized that simple plots and linear regressions are inadequate for capturing these complex relationships. The figure is therefore illustrative. A more extensive statistical search failed to unravel significant and robust relationships between any o f the climate parameters (including second and third moments o f the distribution) and crop yields. Some o f these results are reported inR M S I 2006b. 53 reform that would likely enable the control o f groundwater demand at the wider geographic scale necessary for effective management. However, they provide feasible interimmeasures for reducingvulnerabilities. 37. Agriculture diversification. There is already much evidence that farmers in Anantapur and Chittoor have adjusted to arid conditions by growing groundnut. Though groundnut i s well suited to arid conditions, increased rnonocropping and declining diversity can increase vulnerability to disease andpests, unreliable rainfall, and price fluctuations. This suggests the need for extension o f support systems in rain-fed areas that promote alternative and equally profitable dryland cropping systems. Recognizing these risks, the government o f Andhra Pradesh has initiated a pilot scheme to encourage diversification into less water intensive crops, such as maize, oilseeds, soybean, Bengal gram, and pulses. Figure 3.12 and 3.13 illustrate the current cultivation patterns and the government's diversificationtarget for Anantapur. 38. Water-intensive rice, currently grown inthese areas, serves an essential subsistence need for food and fodder. Research and extension on various forms o f dry paddy irrigation are warranted. Emerging methods o f rice cultivation, such as the system o f rice intensification (SRI) currently being piloted in the state, are innovative and hold promise for scaling up. Indigenous inputs and community-managed support systems for services and inputs appropriate for dryland farming also offer further potential for promoting climate resilience in agriculture. Ongoing nonpesticide management practices supported by the government and the maintenance of community seed buffers for diverse crops are additional strategies that would reduce the cost o f cultivation and consequently also the debt burden o f farmers. In addition, the importance o f livestock to rural livelihoods in marginal cropping systems is increasingly being recognized as a short term response, but it comes withrisks (box 3.6). Figure3.12 CroppingPatterninAnantapur Figure3.13 ProposedCropDiversification I I I I 100% Anantapur:Totalarea proposedfor crop 90% diversificationduringkharifand rabi2005106 80% (in hectares) 70% 60% 5% 50% 40% 30% - 20% 95% 10% 0% B1Groundnut. Kharif2005/06 IPaddy Rabi2005/06 - 1 199697 1997-98 m a 9 9 1999-w 2000.01 b Source: Department ofAgriculture. Source: Departmentof Agriculture Box 3.6 Livestock Systems Ownershipof livestock, especially inarid, semiarid and other noncongenial rain-fed settings, is a critical component of livelihood security. Being more drought resistantthan crops, livestock can provide a safety net against drought, spreading the risks and providing a more even stream of income to eliminate seasonal hunger. But there i s mounting evidence that increasedreliance on livestock dryland pastures could be counterproductive if it leads to further overgrazingand land 54 degradation. International experience points out that on arid lands rainfall fluctuations occur (a) from year to year; and (b) in cycles o f dry years followed by wetter years. This makes it difficult to devise strategies and grazing management plans to cope with such variability. The light stoclung required to match average rainfall can reduce the risk o f forage deficit and financial loss due to death and starvation o f animals in low rainfall years. But this implies lower incomes in good years, although conserving forage may produce healthier animals that command higher prices. Instead, livestock policies in arid areas should facilitate rapid destocking in bad years through opportunistic herding strategies that rapidly adjust grazing pressures to ecological conditions, instead o f assuming that a single stoclung rate will be appropriate for all years. But the larger message i s that sustainability o f livestock management must be enhanced in volatile environments by developing support systems for water-resilient livestock systems and paying close attention to ecosystem productivity and sustainability. Inareas such as Anantapur, farmers tend to have a large number of small ruminants, which, though drought resilient, can be more damaging to pastures. However, the market for this produce i s growing. Sheep and goats comprise roughly 70% o f the livestock population in these areas, with the remainder as milch animals. There is much global evidence o f the pasture damage that can occur with an overreliance on goats. The strategy used by the smallholder mixed (crop- livestock) communities i s to purchase animals during the rainy season, when fodder is available, and sell them during the summer season when there i s a shortage o f fodder. But the greater incidence o f drought in recent years has contributed to a sharp decline in livestock populations. There is a need for biomass intensification targeting small ruminants in these rain-fed areas for more secure and productive livestock systems. Other solutions include promoting the production o f fodder-yielding crops, the development o f fodder banks, and the chopping o f fodder by farmers under rain-fed or irrigated conditions to overcome the shortage o f green fodder duringrabi, when the rainy season ends. Agriculture-embedded livestock systems have a strategic advantage and yield multiple benefits. 3.5.4 InstitutionalNeedsandPriorities 39. There are a large number o fcentral- and state-sponsored programs for addressing drought inAndhraPradeshthat are being implementedunder different guidelines andby various implementing and coordinating agencies. These include the watershed and river basin programs under the Drought-Prone Areas Programme, rural livelihood projects, and farmer schools under the Department o f Agriculture. Since natural boundaries seldom coincide with administrative borders, synergies are lost and these fragmented and uncoordinated approaches have rendered strategic management of river basins and natural resources difficult. Programmatic approaches to drought adaptation are needed to coordinate priorities and fill gaps inthese programs. This remains a challenging task as it requires integration o f diverse programs within a common framework, not only with respect to financial allocation but also with institutions that operate at different levels of government. While the foundation o f empowering community-based institutions (for example mandal samakhayas) has been promoted by a number o f rural livelihood programs in the state, further building their adaptive capacity to manage climate risks would only strengthen them in the right direction. The state's current efforts to explore opportunities to strengthen the labor dimensions in the National Rural Employment Guarantee Scheme and its linkages with other programs holds great potential to catalyze sustainable development inrain-fed areas and set an example for other semiarid states o f India. 55 4. Climate Variability and Change: A Case Study in Drought- ProneMaharashtra 4.1 IntroductionandBackground 1. About a quarter o f India's drought-prone districts are in Maharashtra, with 73% o f its geographic area classified as semiarid. The state i s the second most populous in India with 98 millionpeople. It has a large urbanpopulation, a literacy rate o f 77% (compared to the national average o f 65%), and an economy that contributes 20% o f the country's manufacturing sector output and 13% o f its gross domestic product. Industryand services are well developed and the state remains the financial hub o f the country. But paradoxically, almost 47% o f its population lives below the poverty line. Acute poverty has largely retreated to the rural areas and reflects the low productivity o f the rural economy, on which 56 millionpeople still depend for employment and income. 2. Several factors account for the languishing state o f agriculture in Maharashtra: heavy monocropping in some areas, limited value addition to support agribusinesses, a degrading resource base, excessive withdrawal o f groundwater, and unfavorable market conditions. Furthermore, irrigation, which covers only 16% o f the total agricultural area, i s accessible mainly to larger farmers that have access to power and i s widely used for the cultivation o f sugarcane, a water-intensive cash crop (World Bank 2002a). 3. The drought proneness o f the state i s a critical additional stress factor that adversely affects productivity, livelihoods, and the rural economy. Ironically, the cultivated areas lie predominantly in drought-affected districts (Ahmednagar, Solapur, Nashik, Pune, Sangli, Satara, Aurangabad, Beed, Osmanabad, Dhule, Jalgaon, and Buldhana), which account for 60% o f the net sown area. These areas lie in the rain shadow region east o f the Sahayadri mountain ranges in Maharashtra and the adjacent Marathwada region. Aridity appears to be encroaching upon adjacent areas: districts that previously had moderately assured rainfall, such a Vidarbha, have been afflicted by declining and unpredictable rainfall with debilitating impacts on the local economy (box 4.1). Maharashtra experienced severe and successive years o f drought in 1970-1974 and 2000-2004.38 The state Employment Guarantee Scheme (EGS), a relief andrehabilitation program o f state support, was introduced in 1972 inresponse to a devastating drought. 4. Large tracts o f rain-fed agricultural land in the state have become unremunerative. The agrarian crisis has become acute, with signs o f a breakdown o f coping mechanisms among vulnerable groups whose exposure to drought appear to be increasing. For these reasons Maharashtra represents an important case for assessing the coping capacities o f communities and the underlyingvulnerabilities associated with droughts. 38These periods refer to the fiscal years 197011, 197112, 197213, and 197314, and 200011, 200112, 200213, and 200314,respectively. 56 IBox 4.1 Rainfall Distribution in Maharashtra I Maharashtra is divided into 35 districts split among five regions: Vidarbha (in the northeast), Marathwada (in the south-central region), Khandesh (in the northwest), the rain shadow region (extending from the northeast to the southeast between the coatal districts and Marathwada), and Konkan/Western Ghats (in the southwest, on the coast). There is a wide variation in the distribution o f rainfall across the state, with the coastal belt, the Konkan region, receivingmore than 2,000 millimetersannually, with the second highest rainfall being recorded in the Vidarbha region. Overall, rainfall in Maharashtra increases steadily towards the east and average rainfall in the easternmost districts i s about 1,400 millimeters.The rain shadow and Marathwada regions are the drought-prone areas o f the state, with an annual average rainfall of less than 600 millimeters. These regions are generally characterized by extreme aridity, hot climate, and acute deficiency in water availability. More recently areas in Vidarbha, which usually have reliable rainfall, have experienced variable and reduced precipitation (Planning Commission 4.2 Characteristicsof Study Area 4.2.1 Climate and Geography 5. The two drought-prone districts o f Nashik and Ahmednagar, located in the Godavari basin, were chosen for the case study based on a vulnerability profiling exercise (see appendix B for details). Five villages from the drought-prone belts o f the Ahmednagar andNashik districts were identified for the study (figures 4.1 and 4.2), and a total o f 420 households were selected within these villages. Though both districts are located in drought-prone zones, they are substantially different: Nashik i s located closer to the higher rainfall Western Ghat region and i s the more fertile and moist area. It receives an average annual rainfall o f about 1,000 millimeters, and only 25% o f its area is fully affected by drought. Incontrast, the district of Ahmednagar i s arid and lies inthe scarce rainfall zone. The district is hot and dry with an average annual rainfall o f 579 millimeters, the lowest in the state. All 14 blocks o f the district are partially or fully affected by drought. Around 25% o fthe cultivated area is irrigated. The major food crops include wheat, millets (jowar and bajra), and pulses. Commercial crops grown in certain areas include sugarcane, groundnut, and cotton. 4.2.2 Anatomy of the Sample Villages 6. In both districts agriculture is the primary source of income. Reflecting agroclimatic conditions, there are great differences in average incomes between the districts: sampled villages in Nashik are much more prosperous with an average monthly income o f Rs 22,500, whereas in the arid Ahmednagar villages the average monthly income is Rs 3,455. Large and medium farmer households are more dependent on agricultural sources o f income, draw much higher incomes, and own more assets than their marginal and landless counterparts (table 4.1). The standard o f education and accessibility to health facilities are similar among most households, except for many of the landless households. There is no significant livestock ownership except in one village in Ahmednagar (Karegaon), where all the households owned livestock. 7. Access to irrigation is o f considerable importance to agriculturalists in drought-affected areas. In the sampled households, the quantity and quality o f irrigation varies with 57 landholdings. Access to irrigation inthe sampled villages in Maharashtra is considerably greater than in Andhra Pradesh. More than 40% of the households in Maharashtra use either tubewells or canal irrigation, compared to 18% in Andhra Pradesh. However, the irrigation pattern in both states i s similar, with sharp inequities in the distribution of irrigation facilities. Access to both tubewells and the canal system i s strongly correlated with the size o f landholdings (figure 4.3a). There are also consistent patterns in the quality anddurability of irrigation supplies: ina drought year the irrigation sources o f the marginal landholders are more rapidly depleted than for other landholding categories (figure 4.3b). Box 4.2 describes problems associated with overabstraction o f groundwater inAhmednagar andNashik. Figure 4.1 Location Map of Study Villages in Figure 4.2 Location Map of Study Villages in ashik District Table 4.1 Socioeconomic Profile of Surveyed Households in Maharashtra Land category Ownership Ownership Share of Mean Access to (% of total of utility of productive agricultural normal Years of health households) assets (%) assets (%) income (%) income (Rs) Schooling facility (%) Large (28%) 25 15 86 17,453 6,636 62 Medium(47%) 21 8 78 10,03 1 11 82 Marginal (6%) 24 3 66 4,611 11 88 Landless(19%) 4 0 67 4,319 9 63 58 Figure 4.3a Percentage of Households with Figure 4.3b Percentage of Households with IrrigationAccess: NormalYear Irrigation Access: Drought Year Modivm Msrginal I Note: None of the landless households have irrigation access. Conjunctive use of tubewell and canal irrigation, which accounts for under 1%of the households, i s included. Box 4.2 Unsustainable Groundwater Development Poses Problems for Ahmednagar and Nashik Groundwater overdraft inhard rock areas i s often self-limiting. Once the weathered layer near the surface is dewatered, abstraction rapidly declines due to falling water levels or water quality problems. This i s the case in large parts o f western and central Maharashtra, where groundwater abstraction for sugarcane cultivation has reached unsustainable levels. The most important aquifers in Maharashtra are the Deccan basalts, where groundwater occurs within the shallow weathered and fractured zones extending to depths of 15-20 meters. Deeper aquifers are known to exist within the basaltic lava sequence, but are o f limited capacity. The average water table depths in the shallow aquifer range &om 5 to 10 meters below ground level duringthe post-monsoon period and from 15 to 20 meters below ground level during the pre-monsoon period. Recharge rates are low (8-14% o frainfall). According to state groundwater experts, overuse o f groundwater following a poor rainfall season can result in a groundwater drought in the same year, but it more typically happens in subsequent year(s). Much depends on (a) the levels o f abstraction and (b) the levels o f recharge (i.e. rainfall, run-off and percolation) during the drought and preceding periods. Thus groundwater does in many drought events provide the desired buffer to insulate incomes for a limited period of time, but this is not assured. During a normal monsoon season the shallow aquifers recover through recharge, but increased abstraction and greater incidence o f drought have led to cumulative damage (see figures below). Duringapoor rainfallyear, insufficient recharge and overabstraction leads to aprogressive decline in water levels. The effect is worsened when pumping o f groundwater takes place during the kharif (monsoon) season which is atypical during a good monsoon. The excessive pumping leads to a recession o f the water table and consequently wells dry up during the pre-monsoon period (March or April) o f the next calendar year. The behavior o f the aquifer during drought conditions depends on a number o f site-specific factors, such as the intensity o f drought, extent o f groundwater abstraction, storm pattern and location o fthe village inthe watershed. Thus, proper enforcement o f regulation, planned development o f groundwater combined with suitable agricultural practices can greatly reduce the threat to groundwater security inthese districts. Sources: Directorate of Groundwater Surveys and Development Agency and Department o f Water Supply and Sanitation, Government o fMaharashtra. 59 4.3 ImpactofDrought 4.3.1 Effect of Drought on Households 8. Droughts have a devastating impact on households inthese districts. Decliningincome, in turn, has repercussions on other aspects o f household developmental status. Large landholders, with greater dependence on agricultural income, register the largest decline in income (62%). But with greater wealth and assets they also have greater capacity to withstand fluctuations in income (figure 4.4). Non-income indicators show that the poor marginal and landless farmers are worst affected by deficient rainfall (figure 4.5). As many as 33% o f the landless households report that drought caused disruption in schooling for children. Drought i s also associated with deteriorating health and decreased food consumption, with the landless poor being most severely impacted. A number o f short-term strategies have been developed by rural households in Maharashtra to cope with drought. Box 4.3 describes the main strategies used, as indicatedbythe survey. Figure4.4 Drought Impact on Income Figure 4.5 Nonincome Impacts on Households P) 183 000 - _._. 16 Ow 5Mean NormalIncome P 14 ow 0 Mean Drought Income I25S 12 ow 10wo 20 8 w o 15 6 WO 10 4000 2 OD0 5 0 0 Medium /Box 4.3 Short-Term Coping Measures and Responsesto Drought inMaharashtra 1 When drought descends, households seek alternative sources o f income: increased reliance on wage labor, petty business, dairy, and remittance flows. Adjoining villages with good irrigation infrastructure and cities and towns in the immediate vicinity provide alternative options for income generation, particularly for the marginal and landless categories. Opportunities during drought conditions are limited and impacts are significant for farmer households. On average, 89% o f the surveyed households responded that they embrace some short-term measures to cope withthe crisis o ffalling income: 0 About 50% o f households borrow to cover immediate needs; o f these 53% borrow from formal sources and 47% from moneylenders, often at highrates o f interest. The bulk o f the marginal and landless resort to informal sources o fborrowing. 0 50% o f households temporarily change agricultural practices, reducing fertilizer and seed inputsor growing alternative crops that require less water. 0 50% resort to crop insurance schemes. 0 Seasonal migration i s highinthe rain-fed villages, and among the marginal and landless. 4.3.2 Vulnerability and Income Volatility 9. In drought periods, some households are more vulnerable than others. The empirical analysis finds many shared elements o f drought vulnerability in two dissimilar, though arid, states - Andhra Pradesh and Maharashtra. This would suggest that the case studies 60 have identified some o f the key drivers o f vulnerability among drought-affected population^.^^ Three factors have a disproportionate bearing on vulnerability: e It is no surprise that income diversification turns out to be a key factor that helps to reduce income volatility among the households. The assessment finds that indebtedness locks household into agriculture and increases exposure to climate risks. e Conversely, education and infrastructure bring dual benefits: the usual development gains as well as fostering income diversification into nonagricultural sources, which inturnreduces exposure to drought risk. e Access to groundwater and other sources o f water in dry years can provide a buffer against drought and meager rainfall. The analysis shows that access to groundwater in normal years tends to promote reliance on water-intensive cropping systems and so perpetuates water-intensive agriculture (which i s also more lucrative). If water supplies are assured through a drought, then climate impacts are ameliorated and drought incomes are stabilized. But for most households groundwater supplies decline substantially during drought. The consequence for these households i s a dramatic reduction in incomes during drought years. This highlights the critical role that groundwater could play as a safety net for households in regions with scanty and depleting aquifers (see box 4.2 for a more detailed discussion). It i s important to recognize that these results are based on the findings o f a single and severe drought. Ingeneral, there will be considerable variability in the effect o f a drought on aquifer supplies, reflecting the interaction o frecharge levels (that depend on current and past rainfall, run-off and percolation) and abstraction rates that partly reflect precipitation levels and water needs (e.g., the crop mix). Consequently, deficit rainfall could produce a drought in the same year or with a lag. Inmany cases the impact o f a drought is felt in subsequent years. 4.4 FutureProspectsunder ClimateChange 10. Turning to the future, the climate change projections indicate that these districts will experience significantly different climate patterns. To assess the likely consequences for agriculture, this section begins with a brief overview o f climate change under the two P C C scenarios A2 and B2. Using the Integrated Modeling System (&IS), it then examines crop responses to climate change and briefly investigates the economic consequences. Inthe following section, the policy conclusions are summarized. 39 Income volatility i s measured through the coefficient o f variation. Regression analysis i s conducted using the ordinary least squares technique with standard errors corrected for heteroskedasticity. The regression includes 409 households and only statistically significant determinants are reported here. All factors reported are stable across specifications and are lughly significant. Full technical details are relegated to appendix G. Here, a broad overview is presented that emphasizes the mainpolicy implications. 61 4.4.1 Projectionsof ClimateChange 11. The I M S finds that the following changesmay occur across the Godavaribasin:40 0 An increase inprecipitation o f about 36% (to approximately 840 millimeters) in the A2 scenario and 24% (to about 770millimeters) inthe B2 scenario. 0 This i s accompanied by a projectedincrease inannual maximum temperatures, on average, o f 3.8"C inA2 and 2.4"C inB2. 0 Rainfall i s found to become more variable but the variation will be very similar in B2 compared to A2; the higher rainfall is expected to increase runoffby 12.5%in B2 and by 13.5% inA2. 0 The monthly rainfall pattern i s projected to switch somewhat: while a marginal drop would take place in May and July, an increase would be seen inthe already wet months o f August and September. 0 InA2 about 50% o f the basin's areawould see an increase inrainfall and inthe remaining 50% rainfall will stay within the current range. The increase would take place inthe mid-west, the north, and a small pocket inthe south. InB2, about 60% o f the basin area will stay within the same range o f rainfall while 40% will experience an increase (figure 4.6). At the district level, these changes translate into increases o frainfall for both Nashik and Ahmednagar (see appendix H) 40Projections are for a 30-year period from 2071 to 2100. As noted before, projections for earlier periods at the required level o f spatial disaggregationare unavailable. 62 Figure 4.6 Spatial Distribution of Average Annual Rainfall in the Godavari Basin (Baseline and Climate ChanEe Scenarios) 4.4.2 Crop Responsesto Climate Change 12. The analysis considers two common millet crops, bajra andjowar, that together account for over 60% of the total cultivated area in Nashik and 40% in Ahmednagar. The other main crop considered in the analysis is sugarcane, which accounts for about 10% o f cultivated land in Ahmednagar. While a variety o f newer crops have been introduced to these districts, they represent a small proportion o f the cropped area. The choice o f crops in this study is representative o f the majority o f farms and has also been guided by the need to examine how crops with different climatic preferences respond to projected climate change. Bajra andjowar are included because they characterize drought-resilient crops that are typically grown on rain-fed farms. Intimes o f stress, these crops are given a survival irrigation ifwater is available. Sugarcane i s included because o f its economic significance and because it is a water-intensive cash crop that i s grown exclusively on (larger) farms with substantial irrigation supplies. 13. Crop yields and cropping patterns inthe baseline scenario are compared with those inthe A2 and B2 scenarios. With the rise in temperature and rainfall, the agronomic model finds that average yields of bajra increase dramatically in the more arid district o f Ahmednagar (figure 4.7a). InNashik, bajra exhibits a modest increase inthe A2 scenario, 63 with a negligible impact inB2.41Jowar yields also increase by 8% under the A2 scenario, and by 6% under B2. 14. Accompanying the improving yields is a change inthe pattern (distribution) o f outcomes. The distribution o f yields across 30 climate events inthe baseline and inA2 is illustrated in figure 4.7b. Bajra yields are higher across the entire distribution in Nashik, which implies that beneficial climate outcomes occur with greater frequency, while adverse climate events, which generate low yields, become less common. Ahmednagar exhibits a similar pattern. Figure 4.7a Bajra: Average Yields, Figure4.7b Bajra: Distributionof Yields, Nashik Nashik lields ( t h 0.90 0 80 0 70 0 60 0 50 Nashik Ahmednagar 15. Sugarcane displays a starkly contrasting pattern o f responses. Sugarcane yields are expected to decline considerably (by nearly 30%, figure 4.8a). The decrease in yields i s attributed to increased moisture stress caused by the warmer climate, coupled with the low responsiveness o f sugarcane to carbon dioxide levels. Similar outcomes are reported for sugarcane under climate change scenarios in other regions, such as the Caribbean islands, Mauritius, and Australia.42 Box 4.4 provides a concise explanation o f the complex interplay o f changing climate patterns on sugarcane yields under simulated conditions. Looking more closely at the distribution o f yields, Figure 4.8b shows that there i s a uniform deterioration in outcomes. In particular, climate events that generate low yields are more frequent, whereas high-yield outcomes become less common. 4'The EPIC model predicts that bajra yield inthese regions increases with bothtemperature and rainfall. InB2 there is an almost imperceptible change of about I.% yields. inaverage 42 For a summary of studies in other countries see IPCC 2001, Working Group 11: Impacts, Adaptation and Vulnerability, section 17.2.8, food security. Available online at http:llgrida.nolclimatelipcc~tarllwg2/633.htm. 64 Box 4.4 Sugarcane Yield and Climate Change in Ahmednagar: EPIC Model Projections The response o f sugarcane yield to climate change i s rather complex. The impact takes place through several channels, including temperature, water stress, and the level and distribution o f rainfall. A C4 plant (see glossary) such as sugarcane does not benefit from C02 fertilization. Interestingly, the EPIC model predicts that increasing COz concentration from 420 ppm to 550 ppmwill cause approximately a 13%yielddrop inthe crop. The figure below shows that sugarcane yield responds negatively to hotter conditions. Sugarcane yield drops by 6% with a 1°C temperature rise. The damage i s even greater when the warming intensifies: sugarcane farms will experience 22% yield loss with a temperature rise o f 2"C, and as much as 40% loss with a rise o f 3°C. The level and distribution o f rainfall also affect yields. Higher precipitation will shift the yield curve upward, while less rainfall will shift it in the opposite direction (green arrows). Furthermore, the change o frainfall patternwill move the curve horizontally (orange arrows). Although increasedrainfall leads to a positive yield response, the EPIC model predicts that this i s more than offset by other forces that overall cause yield losses for sugarcane under the A2 and B2 scenarios. SugarcaneYield-TemperatureResponse 'lelda (tha) 70 35 4 0 1 2 3 Change In Temperature (degree mlcius: I I Source: Data from RMSI. Figure 4.8a Sugarcane: Average Yields Figure 4.8b Sugarcane: Distribution of Yields IF";"" 1 I Baseline A2 65 4.4.3 EconomicAssessment 16. The economic module estimates the expected profitability o f different crops and the resulting cropping mix. Starting first with the baseline, the model finds that under current climate conditions sugarcane i s vastly more profitable than either bajra or jowar. The per hectare profits o f sugarcane are approximately Rs 19,000 (based on current prices and inputcosts), while those from bajraare Rs 1,300 inAhmednagar andRs2,300 inNashik. Consequently, on farms where there i s adequate water, sugarcane i s projected as the dominant crop. Elsewhere, for millet growers the crop mix i s determined by water availability; the model finds that there i s a gradual shift from bajra towards jowar as farm water supply declines.43 4.4.4 ProjectedConsequencesof ClimateChange 17. Under the A2 scenario, the higher yields o f bajra and jowar translate into moderately higher profits o f 15% and 8% per hectare, respectively. Conversely, the profitability of sugarcane declines dramatically by nearly 30% per hectare in A2 and by 25% in B2. But at current prices sugarcane still remains considerably more profitable, so there i s little incentive to switch cropping patterns from sugarcane to either bajra or jowar. The returns per hectare from bajra would need to rise dramatically (eightfold, from Rs 2,337 to Rs 19,165 per hectare) for it to become a competitive alternative to sugarcane.44 This suggests that despite improving millet yields and declining sugarcane output, a shift to a more less water-intensive cropping patternmay not eventuate without a change incurrent economic policies. 18. As before, these results are based on the assumption that product prices and all other economic conditions remain at (or close to) prevailing levels. While this approach i s useful in isolating the impacts o f projected climate factors from other future drivers o f change (such as prices and market structure), there still remains a need to examine the robustness o f the projections. These are reported in appendix G. For instance, the appendix shows that ifwater charges were raised from a baseline o f Rs. 1.2/mm to about Rs. 40/mm the entire farm area allocated to sugarcane would shift to other less water- intensive crops. The change across this trajectory i s linear. 4.5 PullingTogetherthe Pieces: PolicyImplications 19. The fiscal burden. A review of existing government and civil society measures for drought adaptation suggests that Maharashtra's drought reliefmechanism has no parallels within India, and the government has developed a well-structured response to address coping and distress associated with drought years. Appendix D provides a summary o f these initiatives. The fiscal and administrativeburden on the state machinery i s very high. During the Tenth Five Year Plan (2002-2007), the Government o f Maharashtra had a planned sectoral allocation o f about Rs 152 billion. However, a single drought (2003/4) 43 The higher average profitabilityo f bajra depends on there being sufficient water for survival irrigation inhot and dry spells to counter the impacts o f heat stress. As the availability o f water declines there is a steady shift towards jowar, which i s found to be more resilient inthe simulations. Note that the price of sugarcane, when the study was undertaken, was Rs 387 per tonne, while that o f bajra was Rs 2,958 per tonne. Sugarcane's productivity was at 50 tonnes per hectare, w h l e that of bajra was 0.8 tonnes per hectare. 66 and flood (2005/6) cost the government Rs 175 billion (table 4.2). Ifother drought years are included the gap widens further, with expenditure on relief much greater than the amount allocated to development programs. The fiscal burdenunderscores the urgencyo f enhancing climate resilience and considering long-term adaptive measures that can be linkedto sectoral programs. Table 4.2 Outlays and Expenditures for Drought Relief and Sectoral Programs, Tenth Five Year Plan (2002-2007) Amount Drought and flood damages Amount total outlay (Rsbillion) and relief (Rsbillion) ~~~ ~ Agriculture and allied services 45.97 Drought damages and relief 57.21 (2003/4) Rural developmenta 61.99 Flood damages and relief 90.30 (200516) Irrigation and flood control works 44.89 PlannedEGS outlay (2202/7) 27.98 Total 152.75 175.49 a. Since the Employment Guarantee Scheme (EGS) also acts as a drought reliefprogram, the planned EGS outlay for 2003-04 has been taken out o f the rural development component o f the Tenth Five Year Plan outlay and added to the relief expenditure and i s an underestimate o f true expenditures which are higher. Source: Government of Maharashtra 2006. Annual Plan 2006-07: Maharashtra State Part 11,Planning Department, GovernmentofMaharashtra, Mumbai. 20. Policy implications. The assessment inthis chapter suggests many shared policy themes that would helpbuildclimate resilience inMaharashtra andAndhraPradesh. . Inboth states local public goods (infrastructure and schooling) not only bring well-recognized development gains, but also have an additional benefit o f reducing long-term exposure to climate risks. These risk mitigation benefits need to be factored into investment decisions and so require the development o f . diagnostic tools and institutional approaches to integrate climate vulnerabilities in policy. Likewise, indebtedness i s known to create a poverty trap, but it also emerges in Maharashtra as a constraint on income diversification that amplifies climate risk exposure. Consequently the policy framework suggested for Andhra Pradesh would apply, with appropriate modifications, to Maharashtra. Inparticular there i s considerable scope in Maharashtra for coupling debt relief with new business capital or insurance for the start-up risks o f a new business. 21. Inadditionto these there aretwo additional policy priorities that warrant special attention in Maharashtra - the reliance on sugarcane and the implications for water use and groundwater supplies. 4.5.1 PolicyAssistanceto Facilitate a Shift from Sugarcane Farming in Dryland Areas 22. Sugarcane i s generously subsidized and has done much to fuel rural prosperity among growers in Maharashtra. But its cultivation is also implicated in the overabstraction o f groundwater. Though sugarcane is highly water intensive, requiring about 2,500 millimeters o fwater per hectare, it is grown on a vast scale ina region that is arid andhas an average annual rainfall o f 600 millimeters. Strong political support has seen the extension o f irrigation command but these regions continue to need drinking water 67 tankers during droughts. The support policies for sugarcane are complex and include administered prices, a proposed export subsidy, and implicit support through negligible user charges for water abstraction. With much o f the sugarcane being cultivated on large irrigated farms, the subsidy i s regressive, accruing disproportionately to the larger farmers. Climate change projections suggest a sharp decline in future sugarcane yields, which will bring intense pressure on the government to increase or maintain current subsidies, with undesirable fiscal consequences. 23. There is evidence that natural resource degradation is beginning to undermine the short- term benefits o f unsustainable irrigation practices. Increased shortage o f water and soil degradation have led to a significant reduction in the output and yields o f the crop. Throughout the state sugarcane output has fallen sharply since 1999, despite a marginal increase in the area under production. Yields per hectare have also fallen, a trend clearly visible in drought-prone districts such as Ahmednagar (figure 4.9). Farmers with limited water supplies have experienced declines in yield from 80 tomes per hectare to 25-40 tomes per hectare. A handful o f forward-looking farmers and community-based organizations inwestern Maharashtra have begun to abandon cane cultivation in favor o f rain-fed traditional crops andintercroppingpractices for soil regeneration. 24. Looking forward, climate change would reinforce the many benefits from encouraging a shift from sugarcane to less water-intensive crop choices. But this will require concerted policy action. Targeted research and extension is needed to explore the possibility o f equally lucrative alternatives to sugarcane as well as to help farmers to minimize the risks o f changes in cropping patterns. However, the outcomes o f research are always uncertain, so there is no assurance that a profitable substitute to sugarcane can be found. The elimination or modification o f subsidies may be required to induce the required shift in cropping patterns. But as elsewhere this would inevitably provoke considerable resistance from the current beneficiaries. So the use o f interim smart subsidies may be needed to shift incentives and cropping patterns in ways that are better suited to the state's agroclimatic conditions. Experience elsewhere, including in the European Union, Australia, and New Zealand, has shown that gradual changes are more acceptable and reforms can be accelerated when accompanied by other forms o f support that target more benignactivities. Figure4.9 Sugarcane ProductivityTrend inAhmednagar (Tonnesper Hectare) (t/ha) 120, 1 110 100 90 80 70 60 50 68 4.5.2 ManagingGroundwaterResources 25. As in Andhra Pradesh, integrated water management remains an overarching priority both for current and fbture agricultural development. Despite the projected increase in rainfall and runoff it i s unlikely that water availability for agriculture will increase in the future with the growing demands of industry, demography, and expanding urban centers. Groundwater could play a major role in h e l i n g sustainable rural economic growth. But for this to occur, abstractions need to be brought in line with recharge. The analysis in this report has shown that contrary to intent, groundwater dependence has failed to provide a buffer to many farmers intimes o f deficit rainfall. Unrestrained competition for groundwater has promoted water-intensive agriculture, leading to overabstraction and increased vulnerability to drought risks. F o r agriculture to become sustainable and drought resistant there is an urgent need to promote judicious water management, emphasizing bothdemand- and supply-side options. Box 4.5 Role of Community Institutions and Participatory Water Resource ManagementinDroughtAdaptation Several landmark examples can be cited from Maharashtra o f community-managed water resource management initiatives that have resulted in significant local benefits to communities, including improvednatural resourcemanagement and livelihoods inlow rainfall environments. Located in Ahmednagar district, the village o f Hiwre Bazaar i s not covered by any major irrigation program; years ago, it was similar to thousands o f other villages in the same block without access to any irrigation. However, effective watershed development and management over the last 15 years have transformed the earlier conditions andreapedpositive developments in terms o f ecosystemrestoration (for example improved soil moisture content) and assured incomes from agriculture even duringdrought years. This has also reduced outward migration. The village has developed its own water regulations linked to its crop plans, which promote a mix o f vegetable and millet crops. Annual decisions on cropping intensity ensure efficient management o f the resource and its equitable distribution for crop growth. InKorhate village inNashik, water user associations W A S )administer water resource sharing for irrigation in major projects. Water allocations are entirely based on cropping patterns and associated volumetric allocations. The WUAs have been found to function effectively and to distribute water equitably, ensuring allocations to small and marginal farmers. Drip irrigation for horticulture crops i s promoted. The government o f Maharashtra further strengthened local bodies during2005, empoweringWUAs with full legal authority to manage water distribution, maintain irrigationchannels, and resolve conflicts. Initiated in the 1970s, the Pani Panchayat initiative in Pune district prioritized hnlung water in the village and restricted the cultivation o f water-intensive crops. There are currently 25 pani panchayat schemes in Maharashtra, based on either a groundwater or surface water communal source. Within a pani panchayat village, nearly a third o f the village land i s typically brought under the scheme, which i s managed under the principles o f delinking land and water rights and cultivation of only seasonal crops. Hydrological parameters, such as groundwater level or rainfall, are used to assess the amount o f water that can be used during the year for crop irrigation. These schemes have survived several droughts successfully, but more recently there i s evidence that some o f these initiatives may be endangered by gradual overabstraction inthe surrounding areas. Sources: TERI 2007; DFID2005a. 69 26. Solutions to the problem o f groundwater overdraft are difficult to design and implement, as there are significant technical and institutional challenges in the management o f groundwater at the higher aquifer or watershed levels where community-based management might occur. These include: 0 Monitoring data are lacking and there are limitations in technical ability to quantify hydrological parameters such as flows, recharge, and water balance, which impede effective groundwater management. A wide array o f watershed programs is in place across the country but their long-term impacts are poorly documented, including with respect to their ability to reverse the problem o f groundwater overdraft. 0 The creation o f institutions and organizations capable o f functioning at the level necessary for aquifer management remains a challenge, and financing models for these local management institutions still remain elusive. 0 Despite fairly robust groundwater regulations, system managers have limited power to enforce regulations or penalties for violating the rules. So incentives for enforcing sustainability remainweak. 27. Recognizing these problems, numerous attempts are being piloted in Maharashtra to promote livelihood-focused adaptive approaches that provide community incentives to directly manage resources (box 4.5). 4.5.3 StrengthenedIntegrationwith OngoingStatePrograms 28. There i s potential in Maharashtra to buildupon a wide range o f relevant programs (box 4.6). But adaptive remedies and programs must endeavor to strengthen convergence with these ongoing programs to focus the impact o f these programs on drought adaptation. For instance, the state has close to ten different watershed programs, each o f which are implementedunder a different set o f guidelines and by different agencies. There is scope to integrate these guidelines into a unified framework for implementation to enhance outcomes with respect to financial allocation and institutions at various levels. The implementation o f such programs can be encouraged by a consortium o f organizations that have established models for scaling up innovative experiences in the states. Further, there is scope for enhancing synergies with the state Employment Guarantee Scheme (EGS) and the National Rural Employment Guarantee Scheme (NREGS) by integrating into the schemes a range o f allied drought-proofing activities, including public works, the repair o f tanks and community water storage facilities, land development, soil and water conservation, crop planning, and agroforestry. 29. Credit availability and poor marketing systems are crucial impediments to progress in drought-affected areas o f Maharashtra. The small farmers remain dependent on moneylenders for routine credit required for inputs such as seeds, fertilizers, and pesticides. Efforts are needed to strengthen the availability o f credit from public and private microfinance institutions to ensure effective access to credit. It is recognized, however, that several administrative bottlenecks impede effective access, as demonstrated inthe low disbursal o f the drought relief credit package announced by the government o f Maharashtra for farmers in the Vidarbha region. But as suggested in the previous chapter, debt reliefdoes little to tackle the root causes o f the problem and needs to be accompanied by complementary mechanisms that provide incentives for income 70 diversification. So there remains scope to pilot schemes that lower the costs of job mobilityandincome diversification, especially amongthe poor andvulnerable. Box 4.6 Snapshot of Sectoral Programs in Maharashtra Water ResourcePrograms Investment inirrigation infrastructure i s large and growing, yet the irrigated area has stagnated at 16%. In addition, while irrigation potential has increased, actual reservoir storage has declined significantly, from 71% in2000/1 to 59% duringthe peak drought year o f 2003/4, primarily due to siltation, poor maintenance, and low rainfall. Fundingi s from a variety o f sources, including through state government funds and centrally sponsored schemes, such as the Accelerated Irrigation Benefits Programme, the Command Area Development Programme, and the Rural Infrastructure Development Fund. The ongoing Maharashtra Water Sector Improvement Project is assisted by the World Bank. WatershedPrograms There are close to ten central and state-sponsored watershed programs inthe state, many o f which operate in drought-prone areas. In addition, the newly formed Marathwada and Vidarbha Watershed Missions in Maharashtra aim to develop fallow and underdeveloped land and groundwater tables in rain-fed areas. In addition, a number o f nongovernmental organizations, including the Water Organizations Trust, the Bharatiya Agro Industries Foundation, Marathwada Sheti Sahyog Mandal, and Action for Agricultural Renewal in Maharashtra run their own watershed programs in limited pockets. The watershed programs include measures such as improved surface runoff collection structures, better groundwater recharge, drainage line treatment, and increases invegetation cover. Agriculture Developing drought-resistant varieties has not been a major initiative o f the Department o f Agriculture. However, dryland horticulture (hit crops requiring less water) has been promoted with some success by the Horticulture Development Programme and i s linked with the state Employment Guarantee Scheme. Since 2005/6 this state-sponsored scheme has been merged with the NationalHorticulture Mission, with fundingbeing shared with the central and state funds. The district-level agricultural technology management agencies responsible for decentralized program planningprovide a useful vehicle for linkages and promotion o f innovative activities. RuralLivelihoods Swarnajayanti Gram Swarozgar Yojana is the single unified centrally sponsored self-employment program for nonfarm-based livelihoods. It aims to establish a large number o f sustainable micro- enterprises. The program targets families inrural areas that live below the poverty line, providing them with income-generating assetsthrough a mix o fbank credit and governmental subsidy. 71 5. Climate Variability and Change: A Case Study in Flood-Prone Orissa 5.1 IntroductionandBackground 1. Floods are a natural feature o f India's river basins. Seasonal floods are necessary to support agriculture, deliver topsoil and nutrients to farmland in otherwise infertile regions, sustain valuable ecosystems, and contribute to groundwater replenishment. But excessive flooding comes at a considerable cost and causes great havoc and damage. Floods are a major contributor to the poverty and vulnerability o f communities. 2. It is estimated that about 12% o f India's geographic area (40 million hectares) is affected by floods.45This is almost double the estimated 19 million hectares affected by floods in India about five decades ago. This has occurred despite rising government spending on flood protection, which has increased dramatically from Rs 0.13 million inthe First Five Year Plan(1951-1956) to Rs 106billioninthe Tenth Five Year Plan(2002-2007), while 39% (15.8 million hectares) o f flood-prone landi s protected with embankments, drainage channels, bunds, and similar structures (Sekhar 2007; Planning Commission 2002). Yet frequent breaches o f embankments and other protective structures are a common occurrence. 3. Orissa i s among the most flood-affected states in the country and i s the focus o f this chapter.46The chapter examines the impact o f floods on selected districts in Orissa and explores the consequences o f climate change and variability for agriculture. The assessment i s in two districts: Jagatsinghpur and Puri. The analysis draws upon field surveys and fkture climate projections derived usingthe I M S described in chapter 2 and appendix A. 5.2 Characteristicsof StudyArea 5.2.1 ClimateandGeography 4. Orissa is located in eastern India, with a coastline that extends over 480 kilometers. Its geography has rendered it highly vulnerable to natural calamities and extreme weather events such as cyclones, droughts, floods, and storm surges. Between 1965 and 2006, the state experienced about 17 droughts, over 20 floods, 8 cyclones, and a super cyclone in 1999, which had devastating consequence^.^^ To compound these problems, there are years inwhich the state simultaneously endures droughts and extensive floods. 5. The state's population o f 37 million resides mainly in rural areas (85%), with a large population o f marginal farmers. The average size o f landholdings is less than 1 acre (0.4 hectare). The main crop i s rice, often grown in a two-crop combination with oilseeds, 45According to the National Flood Commission (Rashtriya Barh Ayog) this estimate reflects the maximum flood- affected area. 46The focus on floods and not on other natural disasters such as cyclonic storms, whch are also frequent in this state, emanated from the consultative process, as noted inchapter 1. After all, cyclones tend to lead to flooding and are, therefore, regarded as beingpart o fthe flood dimensionrather than separatefrom it. 47Data compiled from the International Disaster Database(http://www.em-dat.net/) and TEN 2006. 72 pulses, groundnut andjute. The state is located within the Mahanadi basin, which drains about 42% o f its total area. Large tracts o f cultivated land in the state have no access to irrigation, with 75% o f the cultivated area being rainfall dependent. O f the state's territory, 21% (3.34 million hectares) is considered flood prone (Government o f Orissa 2 0 0 6 ~ ) The~populationmost affected by floods is inthe diara lands o fthe deltaic region . ~ andother low-lying areas. 6. Assessing the full impact o f floods on communities i s a challenging exercise. There are both direct and indirect costs associated with flooding. The direct costs are closely connected to a flood event: the magnitude, extent, and duration determine the resulting physical damage. By contrast, indirect costs are incurred over an extended time period in the aftermath o f a flood. They include loss o f opportunity, reduction inproperty values, foregone tax revenue, and disturbance to ecosystems. Measuring these is more difficult.49 The focus here i s on the direct costs, but even for these more visible impacts accurate measurements remain elusive and are difficult to verify against actual expenditures and losses. 7. Table 5.1 shows that the extent o f flood damage inOrissa has varied greatly from year to year. A feature o f the flood damage is that public property losses are far greater than either agricultural or private property losses. This suggests scope for better protection of public assets through improved planning and climate risk assessments, an issue that i s addressedlater inthis chapter. 5.2.2 Anatomy ofthe SampleVillages 8. The choice o f districts selected for this case study is based on a vulnerability profile developed for the state, which found the districts o f Puri and Jagatsinghpur to be especially susceptible to the impacts o f climate change because o f the regular occurrence o f floods (see appendix B for details). The districts are located on the state's eastern coast. Puri has a long coastline o f 155 kilometers, which i s almost twice that o f Jagatsinghpur. Both districts are very similar interms o f socioeconomic indicators (table 5.2). Puri has an average annual rainfall o f 1,449 millimeters per year compared to Jagatsinghpur's slightly higher average o f 1,501 millimetersper year. For the purposes o f this case study, three villages were selected from each district (see figures 5.1 and 5.2 for ~ocation).~' 48 Of these flood-prone areas, 75% spread across eight coastal districts; Cuttack, Kendrapara, Jagatsinghpur, Puri, Balasore, Bhadrak, Jeypore, and Ganjam 49l k s i s because it requires modeling hypothetical scenarios that would have eventuated inthe absence of floods. 50 The villages are Naugaon, Tarasah, and Sunadiakandha in Jagatsinghpur; and Raibidhar, Gadasampat, and Deipur in Puri. In these villages surveys using stratified sampling were conducted to inventory the existing coping strategies and to assess adapting capacities and vulnerability of communities across villages and landholding categories. 73 Table 5.1 Floods and ResultingDamageinOrissa2001-2007 Year 200112 200314 200415 200516 200617" Administrative areaslpopulation affected Number o f districts 24 27 5 14 - Number ofblocks 219 230 20 72 - Number o f villages 18,790 13,404 564 4,318 22,381 Population affected (million) 9.68 7.62 0.3 1 1.91 8.06 Physical losses and damages Numberofhumanlives lost 102 93 10 10 90 Number oflivestock lost 18,149 2,956 - - 1,656 Crop area affected ('000 ha) 799 1,490 37 94 309 Numberofhousesdamaged 187,575 185,483 2,097 18,099 120,446 Financial loss (Rs million) Crop loss 667 2,538 79 - - Private property loss 564 633 5 - - Public property loss 8,834 11,937 662 2,434 - Total 10,065 15,108 746 2,434 - -Notavailable. a. At the time of writing no figures were available for several o f the categories inthe 2006/7 period. Sources: Memoranda on floods for various years submitted by Government o f Orissa to Government o f India, and annual reports on natural calamities o f the special reliefcommissioner for the years cited. Table 5.2 Socioeconomic Characteristics of Puri and Jagatsinghpur Districts Comparedto OrissaState Indicator Orissa Jagatsinghpur Puri Total area (million ha) 15.6 0.19 0.14 Total population (million) 36.8 1.1 1.5 Ruralpopulation (as % o ftotal) 85 86 87 Population growth rate (%, 1991-2001) 15.94 13.15 14.8 Literacy rate (% o ftotal) 63 79 78 Normal rainfall (mm/yr) 1,482 1,501 1,449 Gross irrigated area (%,2004/5) 34.5 52 40 Maincrops grown Rice, pulses, Rice, pulses, Rice, groundnut, oilseeds groundnut millets Sources: Government o f Orissa 2001a, 2002a, 2005a; Government o fIndia 2004. 74 Figure 5.1 Location of Study Villages in Figure 5.2 Location of Study Villages in Puri Jagatsinghpur District District 5.3 Impacts of Floods 9. Evidence of preadaptation to floods. Households in these villages are accustomed to floods. Contrary to what is observed inmost rural districts, agriculture is not the primary source o f income in the study area. Livelihoods and occupations have responded and adjusted to the predictability o f floods. In Puri 54% of income is derived from nonagricultural sources, and in Jagatsinghpur the figure is higher at 70%. Inaddition to casual nonskilled labor, aquaculture, fishing, dairy, and petty business are the main nonagricultural activities in which households engage. Many o f these initiatives have flourished because o f the proactive interventions of community groups (including self- help groups) and the state government, but the scale o f these activities is still minor. Figure 5.3 summarizes the sources o f income. Large landholders remain the most dependent on agricultural sources of income, mostly from cultivation. The landless and marginal farmers earn the bulk o f their income from casual labor. Figure 5.3 Average Percentage of Sources of Income inTotal Monthly Income in Sampled Households (by Landholding Size) Source. World Bank calculationsfromTERI survey data Sample size' 552 75 10. Volatility of incomes. Figure 5.4 depicts the impact o f the 2003/4 flood on households, illustrating vulnerabilities across sources o f income. Large farmers and the landless suffer the greatest reduction in their incomes. For the large farmers, this reflects their heavy reliance on agriculture and rainfall-dependent sources o f income. There are also differences in the impact o f floods on nonagricultural incomes across landholding size. The nonagricultural incomes o f the marginal farmers and the landless register the largest decline after a flood, reflecting their fragile economic status, typically as unskilledcasual laborers with little job security. Inpolicy terms, it suggests the need for greater attention to these vulnerable groups. Figure 5.4 Changes in Agricultural and NonagriculturalIncomes as a Result of a FloodEvent Source; World Bank calculations from survey data. Sample size: 552 1 Impacts on agriculture. Looking more closely at damage to agriculture, rice in the flood-prone kharif season i s hit hard by inundation; production declined by 67% in the flood year (table 5.3). A direct consequence o f the drop in agricultural output i s that agricultural incomes inthe surveyed households are reduced, on average, by about 20%. However, floods boost productivity in the rabi season as a result o f increased soil moisture and fertility; consequently, there i s a growing emphasis towards more intensive cultivation duringthe rabi season. Table5.3 AverageSeasonalCrop Productioninthe SurveyedHouseholds Normalyear (2005/6) Floodyear (2003/4) % Changein Season Crop Average production Average production average (kg) (kg) production Kharif (June-October) Rice 2,739 893 -67 Rabi Rice 4,863 4,580 -6 (November-April) Groundnut 1,211 1,381 14 Blackgram 273 215 -2 1 Parbal 1,380 1,560 13 Vegetables 930 975 4 Greengram 317 350 10 Note; Total production (kg) =number ofbags x size o fbags (kg). Sample size: 552 Source: TERI2006. 76 12. Broader impacts of flood. The effects of floods go beyond economic and financial hardship. Data from both districts show that the health of the population i s considerably affected by floods: a large number o f households (59%) report a higher incidence of sickness due to water-borne diseases as a result of flooding. The incidence i s reported to be higher in Jagatsinghpur (67%), which i s more prone to waterlogging because o f its low-lying geography, than in Puri (59%). As a result, health expenditures increase by 15%, with the medium farmers and the landless being worst affected. With lower incomes, it i s unsurprisingthat expenditure on food declines, on average, by 6%. 13. Educationis also disrupted by floods intwo ways: children are either forced out ofschool to engage in work to supplement the household income or they cannot attend because school facilities suffer damages. Of the population surveyed, 27% reported that their children drop out of school following a flood event. Inaddition, flooding forces people to abandontheir homes (48%). Forcedmigrationto towns insearch o f alternative sources of employment i s another consequence of natural disasters (9% of the sampled population). Box 5.1 describes strategies followed inthe study areas to cope with floods. Box 5.1 Flood Coping Strategies in Study Area Income diversification away from agnculture appears to provide a robust way o f ameliorating flood impacts. But in addition to this, households adopt a host o f strategies to ameliorate the effects o f floods: Proactive approaches, which anticipate future costs and are designed to avoid or prevent damages. Strateges reportedby the surveyed households include: 0 Safe storage o ffood grains inhouses or public shelters (37% o fhouseholds); 0 Construction o f special facilities for grain storage, for example high shelves (22% o f households); 0 Crop insurance, largely concentrated among large landholders, reflecting their greater purchasing power and needto protect assets o fhigher value (20% o f households); 0 Floodproofingdwellings (16% o fhouseholds). Reactivemeasuresattempt to ease the immediate impact and cost o f flood damage: 0 The most common measure reported was borrowing o f money through credit or loans (reported by 54% o f sampled households). Of these, 48% come from formal credit sources such as banks and cooperatives. This i s likely a consequence o f the existence o f farmers' cooperatives and government programs that have been establishedto address rural poverty and disaster management in h s s a . The success o f these initiatives i s further indicatedby the fact that access to credit i s relatively evenly distributed across all landholdings (see table below). The fact that 53% o f large farmers do not borrow, compared to about 34% o f medium, marginal, and landless, speaks to the fact that large farmers have more assets and a wider set o f options to rely on intimes o fcrisis than do their less fortunate counterparts. 0 Distress sales o f cattle and jewelry (15% o f households) constituted another reactive measure and was largely concentrated among the landless and medium farmers. 77 Distributionof Credit across LandholdingSize Source of Large Medium Small Marginal Total credit (% of households) ~ ~~~ ~ Informal(%) 6.4 6.0 3.9 3.9 7.1 Formal(%) 40.0 57.3 62.7 62.7 49.1 No loan (%) 53.6 36.7 33.3 33.3 43.8 Sample size: 552 5.4 FutureProspectsunder Climate Change 14. Turningto the future, the I M S projects the impacts that a changing climate may have on the agricultural sector. The projections are for the lower Mahanadi basin in the A2 and B 2 scenarios. The basin-levelprojections are illustrated infigure 5.4. 5.4.1 Projectionsof ClimateChange 15. The model projections indicate: 0 In the coastal districts located on the deltaic area the model projects a 23% increase, on average, in the annual mean rainfall in A2 and 19% in B 2 (see appendix H). 0 A projected shift inthe monthly rainfall patterns, with more rainfall inthe already wetter months o f May and July (kharif planting season), and less rainfall from October to December (rabi). The implication i s clear - heightened flood risks in the kharifseasoninthe six districts studied. 0 There are also important changes in the spatial distribution o f rainfall, as the central section o f the basin becomes somewhat drier but the already flood-prone areas on the coast will receivemore rain. 0 At the district level, average annual maximum temperatures are expected to increase in both scenarios by between 2.4"C and 3.7"C, but in A2 the increase i s more pronounced than inB2. Minimumtemperatures are projected to increase by aneven greater amount -as muchas 4.2"C inA2. 16. The assessment suggests that the probability o f flood frequency and its intensity could increase dramatically. The hydrological model projects daily outflow discharges at a gauge station (located in Naraj northwest o f Jagatsinghpur and Pun). The results show that inthe A2 andB 2 scenarios, the probability o f flooding will increase substantially. As an example o f the impliedmagnitudes, the probability that the discharge might exceed, say, 25,000 cumecs (cubic meters per second) i s low (about two to three times in 100 events) in the baseline. But under climate change, this i s projected to rise to about ten times in every 100 events (figure 5.6). Changes o f this scale will have significant implications for the type and location o f hydrological infrastructure that will be needed to protect communities and their assets and highlight the needfor better forecasting tools to identifypriorities for structural interventions. 78 Figure 5.5 Spatial Distribution of Average Annual Rainfall in Lower Mahanadi Basin (Baseline and Climate Change Scenarios) RAINFALL[mm). I STUDY MSTRICTS STATE ~ ~ U N ~ A ~ ~ Figure 5.6 ExceedanceProbability Curves for Annual Peak Flows at Naraj Gauge Station Flow EP curve at NamjGauge Station 0.5 0.45 0.4 0% 0.3 w 025 E 0.2 0.15 0.1 a m 0 IDOOO 1- mm 25oM 3oM[I 001wXl4Emm ?zmm Flow (cumecs) Source: R M S Icalculations, 2006. 79 5.4.2 Crop Responsesto ClimateChange 17. To determine how climate change may affect agriculture it is instructive to distinguish the influence o f climate (the flood event) from other possible changes inthe economy and in agricultural practices. Increases in flood damage due to changes in climate might require different remedies from damage due to changes in economic activity. Accordingly the results presented in this section investigate scenarios with changes in climate, holding economic and technological factors at their baseline levels. 18. To compare future impacts on agriculture, four crops were selected: rice, groundnut, maize, and sunflower. These crops collectively account for almost 70% o fthe agricultural output in six coastal districts, with rice accounting for the largest proportion. Projections were run to predict yields for these crops in both districts. For brevity, the focus is on Puri, as the patternis very similar to that inother areas inthe study. 19. Inthebaselinethetypical farmgrowsricebecauseofitssuitabilitytothe climate. Under climate change the results show that yields o f all crops suffer a decline inboththe A2 and B2 scenarios, but the extent o f the decline i s less in B2 than in A2, given the milder nature o f the changes intemperature and rainfall (figure 5.7a). Figure 5.7b illustrates the distribution o f rice yields under climate change. It shows that climate scenarios that are unfavorable to rice yields emerge more frequently, whereas climate events that generate highyields become less common. Figure 5.7a Yield Changes in A2 and B2 by Figure 5.7b Distribution of Yields under Crop, PuriDistrict ClimateChange, PuriDistrict I -c Baselinel I b e b w O 6 0 5 1 0 1G14 14-18 18-22 22-26 Qqs I Ybld (Ilk.) Source:R M S I 2006b. Source: World Bank calculations. 5.4.3 ProjectedConsequences of ClimateChange 20. Overall, groundnut, a crop that is usually grown in arid regions, is expected to be hit hardest, followed by maize (which i s dependent on soil moisture, but also needs time to dry), rice, and sunflower (which needs moist soil, but also full sun to grow optimally). The economic model estimates that the decline in yields leads to a reduction in farmer profits, as illustrated in figure 5.8. Since the area i s dominated by rice, which i s among the most flood resistant o f crops, there is little that farmers can do to shift planting patterns, suggesting that the prospects for agriculture inthe region may deteriorate under climate change. Climate change consequently reinforces the benefits that would accrue from further income diversification. 80 Figure 5.8 Impact of Climate Change on Farmer Profits Derived from Selected Crops,PuriDistrict 5.5 PolicyImplications 21. Against a background o f more intense and frequent flood risks, it is necessary to ask whether current policies and institutional structures will provide effective protection to vulnerable communities. Orissa, like other states in India, has achieved remarkable success in countering the most extreme effects o f floods (Dreze and Sen 1989). When floods strike, an elaborate relief machinery comes into operation with rapidly arranged protective polices, including employment schemes, cash and food disbursements, health care, and shelter (box 5.2). 5.5.1 Fiscal burden 22. Disaster relief is a high policy priority for the state government, but it comes at a substantial price on the public purse. With the frequency o f natural disasters, Orissa spends more on relief and damages than on planned sectoral and departmental schemes in the rural sector (Go0 2003a-c, 2004, 2005a-b 2006a). Tables 5.4a and 5.4b show that the cost o f relief injust four years exceeded that o f development expenditures for selected major sectoral programs delivered through the Tenth Five Year Plan. The fiscal strain is substantial since central government reimbursements against flood relief are usually lower than the requests for assistance. 5.5.2 Floodmanagementsystem 23. Reflecting the importance given to climate disasters, administrative oversight resides in the highest political office in the state - the chief minister and the ministerial cabinet. Implementationis left to various departments under an elaborate organizational structure. The main thrust o f the flood management system i s on two issues: relief and protection. The relief mechanisms are comprehensive and are judged effective. Similarly, there is a wide-ranging action plan prepared by the Water Resources Department envisioning a host o f structural measures, including dam construction, raising and strengthening o f embankments, and interbasintransfer o f water. Though most state government resources are spent on structural measures, there are also nonstructural initiatives for flood management either planned or under way in Orissa. These include legal measures, flood plain zoning, institutional coordination, drainage plans, and the use o f satellite maps (box 5.2). However, as tables 5.4 a and b show, the state is not fully utilizing its outlays planned for either sectoral programs or relief. These funds could therefore be earmarked for adaptation programs. 81 Box 5.2 Flood Management in Orissa Orissa has a complex flood management system. The state Flood Management Organization in Orissa i s headed by the revenue minister, who controls statewide flood management and relief operations. The Onssa State Disaster Management Agency oversees relief operations and the special relief commissioner, Revenue Department, i s in charge o f coordinating flood relief activities in the state. The Revenue Department coordinates with a host o f other state departments. The district-level heads, the collectors, directly monitor and operate the flood management system and relief works. Likewise there are complex flood reporting protocols involving numerous agencies. Relief The relief system i s comprehensive and comprises: (a) evacuation services; (b) the assessment o f crop loss and other damages; (c) housing and the provision o f other financial assistance in specially deserving cases; (c) health services; (d) drinking water supply; (e) employment; (f) the provision o f free food, polythene, and kerosene; (h) emergency facilities to preserve animal welfare, such as cattle camps, provision o f fodder, veterinary treatment, and vaccination; and (g) furnishment o f daily situation reports to the regional divisional commissioner and the special relief commissioner untilthe danger i s over. NonstructuralMeasures for Flood Managementin Orissa Legal measures. The oldest form o f nonstructural flood management measures i s the legislation that has been enacted over the years, from the Bengal Embankment Act o f 1882, which provided for the construction, maintenance, and management o f embankments and watercourses and gave powers to the collector to move any obstruction to general drainage or flood drainage inany tract o f land, to the Onssa Gram Panchayat Act o f 1965, which requires grampanchayats to construct, maintain, and clean drains and drainage works. Flood plain zoning. Recognizing zoning as an important strategy for flood management, the Government o f Onssa has constituted a subcommittee to finalize a draft flood plain zoning bill. Institutional coordination. The Onssa State Disaster Management Agency, as the nodal agency for coordination and preparedness, coordinates across government agencies during flood preparation, forms community disaster management committees at village, block, and district levels, promotes community-based disaster preparedness through participatory planning, roles and responsibility distribution, and coordination with different departments at different levels, maintains communication networks, keeps ready the OIlssa Disaster Rapid Action Force, and develops and maintains rescue, shelter, and storage infrastructure at village level. Watershed programs. Watershed development programs are currently not focused on flood management but rather may achieve some part o f this as a by-product. The Orissa Watershed Development Mission, an independent agency under the Department o f Agriculture, i s responsible for planning, implementing, and monitoring all programs inthe state. Master plan for drainage clearance in 17 doabs o f Orissa's coastalbelt has beenprepared. Satellite imagery. The Orissa State Remote Sensing Application Centre has several satellite- based images and maps, but basic flood inundationmaps are still not available. 82 Table 5.4a Orissa: Outlays and Expendituresfor Table 5.4b Orissa: Allocation and Expenditures for Sectoral Programs 2002-2007 (Rs billion) Drought and Flood Relief 2002-2006 (Rs billion) Tenth Five Year Plan Drought and flood (2002-2007) Outlay Expenditure relief (by year) Allocation Expenditurea Agriculture and allied activities, forest 7.98 2.69 200213 20.34 18.02 Rural development 8.98 7.96 200314 27.41 21.49 Special areaprograms 16.92 18.80 200415 26.41 23.60 Irrigation, flood control 39.92 24.42 200516 31.80 3.65 Total 73.79 53.87 105.96 66.77 a. Figures for 200213 and 200314 are actual expenditures; for 200415, the figure is provisional; for 200516 it is an average estimate. Sources: 5.4a: Economic Survey o f Orissa, 200415 and 200516; 5.4b: GOO,2006a. 24. The current emphasis on relief and structural solutions i s necessary, but not sufficient to buildflood resilience. Relief is an essential part o f flood policy and is neededto alleviate suffering and distress. But it is limited in its effectiveness, as it does not address one o f the root causes o f vulnerability - the exposure to climate risks. With climate change projected to bring far greater flood risk, the already high costs o f relief could rise dramatically. More importantly, an overemphasis on relief could blunt incentives to shift to more climate-resilient activities and prolong dependence on flood-sensitive livelihoods. 25. Structural solutions are also necessary, but are seldom sufficient to assure full protection. History shows that the public (rather than decision makers) tends to become complacent about the level o f protection that any engineering solution can provide. Structural measures can never offer complete safety, for all possible events, so residual risks on the community will remain. With climate change and the expectation o f more extreme events, these risks will increase over time. This suggests that there i s a need to complement structural and relief solutions with policies that build community resilience to floods. 26. Experience elsewhere suggests that a comprehensive and effective flood management strategy must include a suitable combination o f reactive measures (relief), proactive interventions (both structural and nonstructural measures), and economic policies designed to build flood resilience and lower exposure to flood risks. Countries such as Argentina, China, Poland, and Turkey have struggled for decades with recurrent floods andhave addressed the problem by undertakinga systematic series o f investments inall these areas (box 5.3). 83 Box 5.3 Poland Flood Emergency Project (1997-2005): Good Practice in Nonstructural Measures for Flood Management The Flood Emergency Project was implemented following the 1997 flood in the Odra k v e r basin, which entailed costs o f US$2.3 billion and the loss o f 57 lives. The project's main objective was to increase flood preparedness and management in Poland through the modernization and expansion o f the country's flood forecasting and warning system which, after the completion o fthe project, became one o f the most modem inthe world. A subcomponent on basin flood management planning calls for updating and developing new basin management strategies and plans, with an emphasis on economic assessment o f flood hazard and formulation o f alternatives for different levels o f protection, including structural and nonstructural options. Three regional coordination and information centers were established at Krakow, Wroclaw, and Gliwiceby, and an advanced database, usingremote sensing data, enabled hydraulic modeling o f flood plains and the generation o f flood hazard maps, contributing to the development o f early warning systems and the preparation of flood protectionmeasures. Local flood loss reduction and flood prevention plans, including postdisaster components, were also prepared and implemented in 12 territorial self-government units by local governments and residents, with the support o f technical support units.Cofinancing for reconstruction o f hydraulic structures was provided to eight municipalities. The project included elements that were (a) proactive, including restrictions on flood plain development, preparation o f a handbook on flood mitigation measures, and refinement o f early warning systems: and (b) reactive, including the puttinginplace ofrescue and evacuation plans. A basin approachto water management, consistent with the European Union's Water Framework Directive, was strengthened under the project. The Odra k v e r Basin Flood Protection Project (2007-2014), which aims to protect over 2.5 million people against loss o f life or damage to property caused by severe flooding, will build on the achievements and experience o f its predecessor. Source: World Bank 2006a. 5.5.3 Elements of a Strengthened Strategy 27. The current system in Orissa provide a strong foundation upon which to build a strengthened flood management strategy that is capable of meeting the challenge o f more frequent and intensifyingfloods. There would be three components, all designed to build flood resilience and mitigate damage through an integrated flood management plan. The components are: 0 Advanced systems for the detection and forecasting o f floods; 0 Anticipatory and proactive actions designed to minimize flood risks and build capacity to withstand flood events; 0 Reactive actions that deal with the aftermath o f floods and include compensation andrelief. Strengthening Systemsfor Detection and Forecasting 28. Climate change projections suggest that there will be changes in the spatial distribution, intensity, and frequency o f floods. Advanced forecasting and risk diagnostic tools will be neededto guideinvestments inhigh-value flood protection assets. The current forecasting system for the seven flood-prone basins o f the state are in need o f upgrading and 84 improvement to match the scale o f risks that climate change could bring. The forecasting authority (Central Water Commission) employs a single hydrograph for a very large basin and uses an insufficient number o f forecast points. The system's effectiveness could be enhanced by combining data collection, telemetry, flood forecasting, and flood warning elements into one integrated flood management and information system for the basin.51Flood inundation mapping is another important planningtool and provides local authorities with important information for emergency flood responses. Generating such data for the Mahanadi delta area shouldbe a priority for the Government o f Orissa. Strengthening Anticipatory Measures 29. Although technology can help detect and even forecast floods in a timely way, the informationneeds to be integrated into planningandpolicy for longer-term measures that reduce (a) the magnitude o f the flood; and (b) vulnerability to a flood o f any given magnitude. 30. The assault on floods: importanceof structuralprotection.Any improvements made to existing facilities, or construction o f new ones, will need to take into account the prospect o f more intense flooding and spatial shiRs in flood incidence. Improvements (where needed) and expansion (where possible) o f flood control infrastructure are vital in reducing flood damage. The Government o f Orissa has been prudent in recognizing that absolute protection to all flood areas, for all magnitudes o f floods and for different probabilities o f occurrence, is neither possible nor economic. Economic considerations argue for an emphasis towards the protection o f the higher-value assets (for example urban areas, infrastructure), with greater importance given to building adaptability and flood resilience elsewhere. So carehl monitoring and planning o f new settlements in these flood-prone areasneeds to remain a priority for government authorities. 31. The accommodationof floods: Importance of nonstructuralresiliencebuilding.52A number o fmeasuresmay be considered within this context: 0 Agricultural adaptation. Flood-resilient agriculture provides a way to insulate incomes against flood damage. Numerous pilots have been attempted with more rainfall-tolerant or short-duration varieties o f certain crops to minimize flood- related losses. Though economically viable solutions remain elusive, these initiatives have potential and warrant continued support, as the benefits from research and extension are always uncertain and take time to produce results. A greater emphasis on rubi cultivation could be hrther facilitated by improving irrigation access inthe drier months. 0 Economic instruments. For those with few or no assets, who depend on agricultural wages, the situation i s more challenging. The economic instruments that are relevant for encouraging income diversification for drought management -suchascreditandinsuranceschemeslinkedtojobdiversification-areequally 51As an example the Hirakud Dam, which is the main control structure on the Mahanadi, was originally designed for a flood o f 42,500 cumecs, whereas more recent calculations indicate that the maximumprobable flood i s 69,500 cumecs. Therefore, floods need to be partially regulated by advance reservoir depletion, which in turn calls for a basinwide flood forecasting system. 52Many o fthe arguments presented here echo and extendthose inWorld Bank 2007a. 85 pertinent in the context o f floods. The spread o f self-help groups in Orissa provides a potential community base for launching such schemes. 0 Income diversification. There are already numerous and successful pilots in Orissa that aim to promote flood-based livelihoods. This i s the quintessential form o f flood adaptation. Aquaculture i s one option with considerable promise for unleashing rural growth. With the escalating demand for fish in India, and across the world, there i s scope to increase aquaculture production in flood-prone areas. But for this to eventuate, two obstacles need to be addressed. First, research and extension on biological sustainability, environmental impacts (externalities), fish productivity, and the choice o f species i s in its infancy and requires considerable further investment. A second and more important constraint arises from the outdated marketing system for perishables. Investments in addressing both the research and supply chain obstacles are obvious policy priorities for the flood- prone regions. 0 The primacy of planning and zoning. With the pressures o f rapid population growth and land scarcity coupled with intensifying flood risks there i s a need for better and more careful planning and flood zoning. Land use planning and water management need to be combined into a synthesized plan and call for coordination between various departments and levels o f government. A greater challenge i s the implementation o f a plan that would affect many interests and would need processes that involve public participation and stakeholder engagement. 0 Improved watershed programs. Watershed development programs should take into account flood management as an explicit design practice. International experience shows that planning watersheds upstream in the river basin can significantly contribute to improved flood management. 5.5.4 Reactive Strategies 32. Interms ofreactive policies, India has established disaster relief systems that are among the most comprehensive in the world (box 5.2). There effectiveness could be improved through further fine tuning. But ultimately adapting to floods remains the most sustainable and effective way to protect communities and harness rural growth. Improvements inthe reliefsystem might include the following: 0 Improved training of panchayats and communities in flood preparedness. Community contingency plans for relief have beenprepared through government initiatives but inmany cases the plans have not been updated and have lost their relevance as circumstances change. 0 There i s a gap between what i s being planned for relief and what is being implemented on the ground (for example, what medical facilities are provided), largely due to poor accountability andmonitoring o freliefwork. 0 The protection o f critical infrastructure and shelters, surprisingly, does not figure in the list o f disaster preparedness actions at the state, district, or subdistrict levels. Safe shelters have been constructed in the coastal belt primarily to protect the coastal communities against cyclones, not floods. The maintenance and effectiveness o fthese shelters are therefore a concern. 86 6. A Way Forward 1. India has achieved impressive economic growth inthe past decade but its farm sector has languished, with stagnating yields, low productivity, and pockets o f poverty and indebtedness. Aware o f the growing disparity between agrarian and industrial India, the government has assigned the highest priority to revitalize the farm sector to its fullest potential inthe Eleventh Five Year Planperiod (2007-2012). Targeted growth o f 4% per year in the agricultural sector is deemed necessary to achieve the aims o f poverty reduction, inclusiveness, and increased income opportunities, with a special focus on rain-fed areas. But this i s a challenging task and is made more so by the impacts and consequences o f ongoing andfuture climate variability andchange. 2. There is a litany o f factors that account for agriculture's slow growth. These include fragmented landholdings, prevalence o f subsistence and marginal farming, inadequate levels o f technology, poor market access, and a shrinking natural resource base, endangering agricultural productivity and farm livelihoods across India. Added to this, projections indicate that climate variability and change will be characterized by more frequent and severe droughts and floods, posing significant economic, environmental, and social risks. Since 60% o f India's workforce depends on agriculture and natural resources for livelihood, employment, and income, they are highly vulnerable to future climate risks and have limited adaptive capacity to effectively deal with the increasing unpredictability o f climate change. 3. Delivering on the promise o f faster and more inclusive growth will therefore require investment strategies that more effectively capture the potential effects o f future climate trends and are accompanied by a portfolio o f adaptation options that can diminish the associated risks. Living up to this commitment will require proactive and integrated tools, and interventions and policy measures, that are specifically targeted and tailored to the high-riskand vulnerable areas. Improvementsininstitutional capacities must also be met at an equal pace to facilitate adaptation. 4. This report suggests that adaptation to climate variability and change can be tackled through a comprehensive climate risk management approach, starting with the most severe vulnerabilities arising in regions that are subjected to frequent, intense and damaging climate events. Using this approach, the report makes a strong case for moving towards integrated risk management solutions for enhancing the resilience and adaptive capabilities o f rural residents in selected areas o f semiarid India. The report emphasizes that there is no single adaptation solution. Effective and sustainable adaptation to climate variability and change requires a combination o f measures that must be implemented at multiple levels (national, regional, local, and community). 5. The study findings and recommendations coupled with a rigorous review o f the literature and consultations with government officials and NGOs suggest four interrelated strategies that would help lower the exposure to climate risks and build adaptive resilience: e Strengthening climate information systems and mechanisms and related management tools to match current andfuture needs; e Fosteringclimate-resilient reforms inagriculture andwater resource management; 87 0 Supporting the management o f climate risks with economic mechanisms and instruments; 0 Improving institutionalcapacities and linkages among sectoral programs. 6.1 StrengtheningClimateInformationSystems andMechanisms 6. The study has shown that climate change projections are characterized by a high degree o f spatial variability in rainfall and temperature trends, which translates into spatial heterogeneity in drought and flood incidence. Generating high-resolution climate information i s an important first step to factor climate risk into development decisions in the states. This can be done with more localized meteorological data and climate models, which can help to better identify areas at risk. 7. Indealing with floods, there is an overwhelming case for strengthening the current flood forecasting system to guide investments in high-value flood protection assets and to support more effective development and targeting o f nonstructural approaches. The effectiveness o f such a system can be strengthened by combining data collection, telemetry, flood forecasting, and flood warning elements into an integrated flood management and information system for the basin. Developing more robust flood inundation maps for planning and use by local authorities i s another important priority that would facilitate risk assessmentinzoning and planning decisions. 8. However, climate and meteorological data processing is complex and lies at the frontiers o f climate change research. The capacity to generate climate information rests with research centers, while the need and demand for such information and products lies with local communities inrisky areas. This argues in favor o f a climate information system at the national or state level to disseminate information for planningand management to end users, such as irrigation or farm sector managers. This i s also recognized inthe Planning Commission's working documents for the EleventhFiveYear Plan. 6.2 Fostering Climate-Resilient Reforms in Agriculture and Water ResourceManagement 9. There i s a strong case for more aggressively pursuingwater conservation across drought- prone states. The projections indicate that even when farmers have largely adapted to arid cropping patterns, increased demand and consequent water stress could severely jeopardize livelihoods and render agriculture less viable in these regions. Reliance on supply-side approaches does little to curb the escalating and unsustainable demand for groundwater. Greater attention must therefore be given to hybrid approaches that emphasize the efficiency o f groundwater use and increase the effectiveness o f watershed activities to conserve soil moisture and harvest rainwater. Such adaptive measures are not a substitute for much needed economic incentives to enhance the productivity of water and water policy reform aimed at, the control o f groundwater demand at the wider geographical scale necessary for effective management. However, they provide interim and feasible measures for reducing vulnerabilities. 10. The study makes a strong case for a shift in agriculture systems in order to overcome h t u r e climate change pressures. It recognizes that there are manyneeds and opportunities related to farming systems. The use o f interim smart subsidies may help shift incentives and cropping patterns to modes that are better suited to state-level agroclimatic 88 conditions. Additional measures, including strengthened support for agricultural research and extension and opportunities for reduction in costs o f production, are essential to promote more sustainable modes o f dryland farming, for example, water-efficient processes for paddy cultivation, promotion o f millets, oilseeds, and pulses and use o f low-external inputs inagriculture. 11. There are other ways inwhich rain-fed farming systems can be made more sustainable in semiarid areas. It is not within the scope o f this report to examine these in detail, but some innovative methods include non-pesticide management, intensification o f biomass used by small ruminants, water-saving composting methods, seed production, livelihood opportunities associated with agro-forestry and livestock-based production systems, and pooling o f farmbore wells, packaged with appropriate incentives. 12. In the case of floods, solutions in the water sector would involve a combination of infrastructure and nonstructural approaches such as flood plain zoning, institutional and community coordination, and drainage planning measures. This is not to diminish the importance o f structural measures, but as indicated earlier, flood forecasting systems can guide such investments to maintain long-termperformance and efficacy. 6.3 Economic Mechanisms and Instruments to Promote Income Diversification 13. This study has emphasized the need for innovative and cost-effective ways o f reaching poorer farmers to help reduce their risk exposure. Indebtedness has beenidentifiedas one o f the major impediments to occupational mobility. Debt instruments offer the potential to protect many vulnerable sections o f society and overcome the limitations o f crop insurance schemes. The study has also suggested caution in the design o f insurance schemes arising from the potential mismatch between payouts and actual losses. Coupling debt relief with new business start-up capital provides a way o f encouraging income diversification by lowering risks and transaction costs. These initiatives could be channeled through public andprivate microfinance institutions. 6.4 ImprovingInstitutionalCapacitiesandProgramLinkages 14. Several government programs provide a rich and varied platfonn upon which to build comprehensive adaptation strategies. But the programs are largely uncoordinated and operate in isolation. Integration is needed to harmonize the essential ingredients o f drought adaptation by buildingupon and facilitating synergy with ongoing programs, for example those m by the state departments o f water conservation, rural development, agriculture, andwater resources. 15. Apex bodies, such as the National Rainfed Area Authority and the National Disaster Management Authority, could play a strong role in coordination, planning, and identifying gaps and synergies in programs. The panchayati raj institutions and community-based organizations have an important role to play in harnessing opportunities and building appropriate capacities for employment security and asset building, while ensuring the effective use o f such assets inreducing the adverse impacts o fdroughts and floods. 89 16. Numerous opportunities exist to build climate resilience within current programs. For example, states could build in an adaptatiodclimate change dimension into the district agricultural planss3 which would go a long way in mainstreaming the climate risk management agenda as well as in enabling conditions for translating the recommendations o f the report into actions on the ground. Furthermore, there i s scope to introduce innovative activities for rain-fed farming systems in the National Rural Employment Guarantee Scheme. States could also integrate improved agricultural technology into watershed development programs. A number o f state and centrally sponsored schemes being implemented under different guidelines and by different agencies can be brought under a common framework with the dual purpose o f strengthening development outcomes andbuildingclimate resilience. 17. As regards floods, the institutional set-up for relief operations is comprehensive and operational. But adaptation to floods requires greater integration across different sectors, including water resources and irrigation, housing and land development, and forestry and agriculture. Since adaptation i s a new and emerging concept, there is a role for creating state-level climate risk management committees as a first step infacilitating synergy and planningfor adaptation activities. 6.5 FutureWork 18. Inconclusion, there is a clear recognitionofthe need for, andcommitment to, moving forward with a development strategy that can take into account future climate risks, as indicated by the many central and state programs that include climate risk components, andthe emergingpolicies embodied inthe EleventhFive Year Plano fthe Government o f India. This study has developed, through analysis and consultation, a series o f adaptation options (with associated time frames) that support the development priorities to build a more climate-resilient and sustainable India. These are summarized intable 6.1. 19. Strategic actions including barrier removal activities are required by the Government o f India to implement the recommendations o f this report. These would include greater investments in climate forecasting and dissemination activities; economic incentives to promote income diversification inthe form o f both insurance and credit services inorder to give farmers incentives to shift to long-term viable non-farm activities and price stabilization interventions by the government for rain-fed and drought-resistant crop varieties, and finally market-based and demand-drivenmechanisms to help farmers with assuredsales o ftheir crops and livestock. 20. Climate change will have wider impacts that go beyond the flood- and drought-affected areas that are the focus o f this report. Most notably future work i s needed inareas o f high development significance. First, glacial melt remains the most dramatic risk that could threaten the water supplies, food production and life-sustaining ecosystems upon which millions in the Indo-Gangetic plains depend. There is an urgent need to assess the magnitude o fthese risks, the economic implications and identifycost-effective adaptation responses. Independent o f this, changing rainfall and temperatures will affect the 53 The National Development Council Resolution of May 2007 under the chairmanship of the Prime Minister of India gives highpriority to incentivizing states to develop comprehensive district agricultural plans that will include livestock, fishing, minor irrigation, rural development works and other schemes for water harvesting and conservation. 90 productivity o f major food supplies in the rice and grain production regions of India suggesting the need to investigate the impacts on food security goals, livelihoods and agricultural growth targets. Similarly, the threats from sea level rise on coastal communities and cities is another important issue that warrants greater investigations. Finally, though seldom recognized, the interplay between climate risks and agricultural trade needs to be explored in greater detail. Distortions from the protectionist policies in developed countries are likely to increase climate risks indeveloping countries. These are important issues that need to be addressed in future work. 91 Table 6.1 Summary Recommendations for Adaptation54 Strengthening publicly accessible climate information systems/mechanisms and relatedmanagementtoolsto matchneeds Establish a climate information management system at central level for developing climate diagnostic and risk assessment tools with feedback mechanisms to end users. This would include: enhanced data collection systems at the local level, systems for hydrogeological data collection and information for groundwater management, and systems for improved detection and forecasting o f floods. Buildclimate risk assessment as a requirement for all relevant highvalue and long-lived infrastructure projects. Supportingthe management of climate risks through economic mechanisms and instrumentsthat promoteefficiency Explore new innovative financial instruments to promote income diversification, such as >P debt relief instruments coupled with credit for job diversification > debt relief coupled with insurance for new business risks community-based risk financing schemes Introduce interim smart farm subsidies to encourage switch to more suitable and climate resilient cropping practices Fostering climate-resilient reforms in agriculture and water resource management Promote agricultural research and extension services towards systems and cultivars better suited to local climate and its variability Implement demand-side approaches for management o f groundwater resources at watershed and aquifer levels indrought-prone areas 9 Promote basin-level irrigation systems in drought- and flood-prone areas designed to take into account climate risks Improvinginstitutionalcapacitiesandlinkagesamongsectoralprograms Establish capacities and strengthen role o f a central bodies such as the National Rainfed Authority and National Disaster Management Authority in order to strengthen coordination and operational linkages between departments at all levels o f government. This could include establishing convergence committees for management o f drought and flood inthe states. Integrate measures targeted towards management o f future climate risks in the planningprocess indistrict agriculture plans55 54Anumber o f recent projects supported by the World Bank have incorporated some elements o f these recommendations in operations. Examples include: the Hydrology I(1995 - 2003), which focused on nine states including Andhra Pradesh, Maharashtra and Orissa), Hydrology I1(2004 - 2010), which currently covers five new states. These include enhanced information systems as project components. The National Agriculture InnovationProject (2006 - 2012) and the National Agriculture Competitiveness Project (under preparation at the time o f writing) recognize and buildinelements o f climate risk. 55GoI's National Development Council Resolution o f M a y 2007 gives highpriority to building comprehensive district agricultural plans. AppendixA: The IntegratedModelingSystem Framework - andApproach Model Structure The Integrated Modeling System (IMS) developed for purposes o f this study consists o f three sub-components - HadRM3 climate data, a hydrological model (SWAT), and an agro-met simulation (EPIC) model - and their functional links. These sub-components are, in turn, linked to the economic model, described in Appendix G. The modeling toolkit and database binds these sub-components into single modeling tool, in a simple and interactive way. The basic components o f the integrated modeling system include a control unit, database management, model- and knowledge-base management, and user interface. The underlying flow o f data and information in the Integrated Modeling System i s illustrated in figure A.1. The starting point for the I M S is the generation o f climate data based on the IPCC emissions scenarios. The resulting climate data is then used in the hydrological model, SWAT, to generate surface water data, which are required as inputs to runthe agro-meteorological model, EPIC. The latter integrates water and climate data into an agricultural output estimation framework. Detailed information about SWAT and EPIC i s provided later in this annex. Both SWAT and EPIC, are process-based deterministic, each i s governed by a set o f modeling equations. The spatial resolution used in this study i s MandaVBlock resolution. However, calibrations were done with the help o f local scientists - to know more about EPIC and its calibration. The I M S requires a computer running on, at least, Windows 2000. The minimum platform configuration i s a Pentium or equivalent processor running at 100 megahertz with 64 megabytes o f memory, at least 500 Mb o f free disk space, and a display resolution o f at least 1024 x 768. For optimal performance, a Pentiumprocessor running at 400 megahertz, or faster, with at least 128 Mb o f memory, 1Gb o f free disk, and a display resolution o f 1280 x 1024 i s recommended. 93 E c, W v) m h M E 3E .3 3 The Graphical User Interface (GUI) for IMS-EPIC is a Windows-based tool that communicates with the internal database for data entry, editing, and data validation. The GUI was designed to overcome most o f the significant hurdles faced by EPIC users, particularly the lack o f a user- friendly interface. The development environment o f the GUI i s based on Arc View Avenue and Visual Basic; both applications operate under Microsoft Windows. Figure A.2 provides an example o fthe appearanceo fthe interfacewhich will pop up duringinstallation. Figure A.2 Interface for Installation of the Application Once the application is installed on the computer, the Arc View application i s launched. Using the toolbar one can open a view to show the regions o f interest (figure A.3). Figure A.3 I M S Toolbar inArc View I I I ~~~1 , Yields inlayout In order to runprojections and create charts depicting the projected parameters, one makes a connection to the preferred climate scenario database by choosing the scenario: Baseline, A2, or B2. The dataset i s uploaded by the system and the data corresponding to the block or region desired will be displayed in a series o f interfaces showing the block, the crop(s) being analyzed, 95 the type of soil, the weather, and the crop management information. It is also possible for the user to edit such parameters. Finally, crop yields are displayed using bar charts and layouts as shown infigures A.4 and AS. FigureA.4 CropYields ShownUsingBar Charts FigureA.5 SpatialRepresentationofAverage CropYieldinthe Chittoor District 96 HydrologicalModeling (SWAT Model) SWAT, which stands for Soil and Watershed Assessment Tool, was developed to predict the impact of land management practices on large, complex river basins or watersheds. SWAT2000 is capable o f performing continuous, long term simulations for watersheds composed o f various sub-basins with different soils, land uses, crops, topography, weather, etc. Being a physically- based model, SWAT2000 requires specific inputs to model any river system rather than use o f regression equations to describe relationships between the inputs and outputs. The driving force i s water balance. Another important characteristic o f SWAT2000 i s that it provides relative accuracy as well as absolute accuracy. This model i s best used to predict long-term outcomes o f management practices. SWAT2000 can handle hundreds o f sub-basins. The soil profile for each o f these sub-basins can be divided into ten layers. The movement of runoff, sediment, nutrients, and pesticide loadings to the main channel in each sub-basin is simulated considering the effect o f several physical processes that influence the basin's hydrology. SWAT2000 requires data relating to daily precipitation, m a x i m u d minimum air temperature, solar radiation, average daily wind speed, and relative humidity. This information can come from observed data or it may be generated from a weather generator database. The precipitation may be homogenous for the entire watershed; however, spatial variability may leadto unique climate conditions for the various sub-basins inthe model. Box A.I Stochastic Weather Generator A stochastic weather generator allows the generation of synthetic daily weather data for any number o f years using the statistical properties of a weather variable. In this study, a stochastic weather generator was used to generate scenarios of future climate at block level from the low-resolution RCM-derived scenarios. The climate scenarios developed from the outputs o f the R C M include not only changes inthe mean o f the climate but also changes inits variability. The historical annual cycle o f means, standard deviations, probability o f weddry days, and number of rainy days at block level were computed using the daily rainfall data at block level and other daily weather data obtained from the India Meteorological Department (IMD). The stochastic weather generator was used to simulate daily weather for 50 years for historical (Baseline), A2, and B 2 scenarios ineachblock o f the selected districts o f each studyregion. The weather generator i s designed to preserve the dependence in time, the internal correlation, and the seasonal characteristics that exist inthe actual weather data. Precipitation and wind are generated independently o f the other variables. Maximum temperature, minimum temperature, and solar radiation are generated subject to whether the day is wet or dry. A first-order Markov chain i s used to generate the occurrence o f wet or dry days. When a wet day is generated, the precipitation amount is based on a skewed normal distribution. With the first-order Markov chain model the probability o f rain on a givenday is conditioned on the wet or dry status o fthe previous day. The procedure to generate the daily values o f maximum and minimum temperature and solar radiation is based on the weekly stationery generating process given by Matalas (1967). The wind component o f the model provides for generating daily values o f wind speed and direction as described by Richardson(1982a). Source: RMSI, 2006b. 97 Assumptions andLimitations A watershed may be broken down into several sub-basins, each consisting o f different Hydrologic Response Units (HRUs). The HRU i s the primary modeling unit for the SWAT model. Within a sub-basin, an HRUoften consists o f areas with the same hydrology, soil type, and management practices. The model assumes that the hydrological routing paths are the same for all the areas that belong to the same HRU, even if the areas are distributed in different parts o fthe sub-basin. Since the model i s designed to simulate and predict the impacts o f long-term land management practices in watersheds on the water quality in receiving water bodies, it cannot properly be applied to simulate detailed, single-event flood routing. The stream flow network in SWAT i s designed as one directional flow, which routes runoff from upper streadreach to down stedreach. It cannot simulate backflows. The basic model components simulated by SWAT2000 include weather, surface runoff, return flow, percolation, evapotranspiration, transmission losses, pond and reservoir storage, crop growth and imgation, groundwater flow, reach routing, nutrient and pesticide loading, and water transfer. Agrometereological Model (EPIC) The EPIC (Erosion Productivity Impact Calculator) model was developed by scientists from the US Department o f Agriculture's Agricultural Research Service (ARS), Soil Conservation Service (SCS), and Economic Research Service (ERS). The model was selected for use in the I M S because it provided a more coherent modeling environment and because there was experience available inthe application o f EPIC inrelevant parts o f India. EPIC was originally designed to assess the effect o f soil erosion on productivity. It simulates the effects o f management decisions on soil, water, nutrient, and pesticide movements, as well as their combined impact on soil loss, water quality, and crop yields for areas with homogeneous soils and management. Some o f the important components o f EPIC are: weather generator (WXGEN); hydrology, erosion and sedimentation: nutrient cycling; crop growth; tillage; economics; andplant environment control. Model Resolution The IMS' model resolution is aimed at "blocks"; results are aggregated at "district" level, since this i s the level o f resolution o f the agro-met model. All efforts are made to collect data at basiddistrict level and the results are generated at this level. For regional levels, the results can be aggregated from the districtshlocks. Apart from this, one major driver o f this study i s HadRM3 (a Regional Climate Model), which has a resolution o f 0.44 x 0.44 degrees (approximately 50 kilometers cell-size) on ground covering the average size o f a typical Indian districthub-basin. 98 Model Validation The model was validated through rigorous testing involving the comparison o f reported crop yields to simulated yields based on similar conditions to those prevailingat the time that the real- world reported data were published. Efforts were made to provide accurate and realistic data- input files to close the gaps between simulated andreportedcrop yields. The validation exercise showed that the observed crop yields and those simulated by the I M S were very close. Box A. 2 Regional Climate Models vs. Global Climate Models The simulation o f seasonal rainfall as well as its spatial and temporal variability over the Indian subcontinent has remained rather poor in most Global Climate Models (GCMs). This is mainly due to the fact that GCMs have course horizontal resolution, which restricts the representation o f the topography's and coastlines' complexity. It also limits the parameterization o f sub-grid scale processes. Therefore, GCM scenarios fail to capture the local details needed for conducting impact assessments at the regional level. In addition, GCMs cannot capture extreme weather events and their intensities such as cyclones or heavy precipitation events. The alternative method to obtain detailed predictions o f the future climate i s to use a high-resolution version o f GCMs known as Regional Climate Models (RCMs). AnRCMhas a highresolution (typically 50 kilometers, comparedto 300 kilometers ina GCM) andcovers a limited area o f the globe (typically 5,000 kilometers x 5,000 kilometers; roughly the size of a box around Australia). It i s a comprehensive physical model, usually made up o f the atmosphere and land surface, containing representations o f the important processes inthe climate system (e.g. clouds, radiation, rainfall, soil hydrology). At its boundaries, RCM i s driven by atmospheric winds, temperature, and humidity output from a GCM. R C M predictions o f ideally 30 years (e.g. the period 2071-2100) are needed to provide robust climate statistics, e.g. distributions o f daily rainfall or intra-seasonal variability. The third-generation Hadley Centre RCM (HadRM3) is based on the latest GCM, HadCM3. It has a horizontal resolution o f 50 kilometers with 19 levels in the atmosphere (from the surface to 30 kilometers inthe stratosphere) andfour levels inthe soil. Inadditionto acomprehensive representationofthephysical processes inthe atmosphere and land surface, it also includes the sulphur cycle. This enables it to estimate the concentration of sulphate aerosol particles produced from SOzemissions, which have a cooling effect as they scatter back sunlight and also produce brighter clouds by allowing smaller water droplets to form. Thus, regional models cantake account o fthe effects o fmuchsmaller-scale terrain than GCMs. Inspite the RCMs' advantages, their level o fresolution is not highenough to assess the impact of climate change on natural resources. Inorder for this assessment to be as accurate as possible, the resolution would have to be evenbetter, that is, approximately 10kilometers by 10kilometers. Inan ideal, best case scenario it would be as high as 1kilometer by 1kilometer. Thus, there is a mismatchbetween what climate models can supply andwhat natural resource impact models require. Since the technology to create even lower-resolution climate models is not yet available, other alternatives are available to produce a similar effect. These alternatives involve the manipulation o f the climate data fed to the models using an approach called "Statistical Downscaling". T h ~ means developing hgh-resolution s climate data using the IPCC hture climate scenarios and the observed data o f rainfall, temperature, solar radiation, relative humidity, and wind speed. Source: RMSI, 2006b. 99 Appendix B: Methodology Used for the Design and Analysis of HouseholdSurveys and Data Vulnerability: A Complex Term There i s no one single definition for the term "vulnerability" andno one single way o fmeasuring it.Different disciplines define it and measure it differently, but the one common trend among all o f them is the idea that the concept i s related to levels and types o f risks to which people/communities are exposed. Table B.l summarizes some o f the most commonly used definitions. Table B.l: DefinitionsofV~lnerability~~ Discipline/ Definitionof Vulnerability What is MeasuredHow Criticism Literature it is Measured Itis an outcome o faprocess o f The fall o f incomebeyond There is an underlying Economics household responses to risks, the poverty line or presumption that all 57 given a set o f underlying changes inconsumption losses canbe measured in conditions. Often times, the are measured. monetary terms. outcome i s poverty. It is the probability that The loss o f livelihood, ~~ Ittends to use terms and ~ Sustainable "livelihood stress" will occur - continued vulnerability to concepts that are unclear Livelihoods with more stress or a higher subsequent shocks and or not widely accepted. It probability implyingincreased vulnerability changes over i s not clear how changes vulnerability. Also, "the balance time are the subjects of invulnerability wouldbe between the sensitivity and interest. The assessments evaluated over time when resilience o f a livelihood are specific to population some indicators show a system." or society. It uses a case positive change while study approach. others a negative one. It is the risk of irreversible Vulnerability mapping Itusually lacks a Food physical or mentalimpairment and indexes. A number o f b e n c k r k to which Security due to insufficient intake o f analytical techniques are indicators can be macro or micronutrients. used to examine the compared. It recognizes degree o f correspondence that vulnerability is made between the concept o f upo fdifferent food security and the components, but it ignores indicators chosen to the specific process by measure it. which the components interact to determine overall vulnerability. Itis the characteristics o fa Vulnerability = There is a lack of Disaster person or group interms o f their Hazard- Coping precision inthe language Management capacity to anticipate, cope with, Household characteristics used, which leads to and recover from the impact o f a are key determinants in confusion. At times, it natural disaster. It i s an that they affect either side fails to be specific about underlying condition separate o f the equation. what constitutes loss or 56Alwang, Jorgensen and Siegel. 2001. Vulnerability: A View from Different Disciplines. Washington, DC: World Bank. 57Variants o f this approach are the "poverty dynamics" and the "asset-based" approaches to vulnerability (see source for details). 100 from that of the risky events that The use o f vulnerability damage, or whether it may trigger the outcome. It mapping is also matters who endures refers to risks as "hazards". widespread. these. The differences between the approaches can be reduced to the tendency o f each discipline "to focus on different components o f risk, household responses to risk and welfare outcomes". All approaches have their strengths and weaknesses: some are considered strong intheir conceptual framework but weak in their empirical approach (i.e., how it i s measured) and vice versa. The definition used inthis study i s eclectic: it borrows from all o fthese disciplines. Selectionof Study Areas-Vulnerability Mapping Vulnerability indices are commonly used in the field as a way to measure vulnerability by different researchers and institutions. Two such indices are the "Food Insecurity and Vulnerability Information and Mapping System" (FIVIMS), developed by the Food and Agriculture Organization o f the United Nations (FAO); and the "Vulnerability Analysis and Mapping (VAM)", producedbythe World FoodProgram incooperation with FIVIMS.58 Inthis study, avulnerability index was also developed to guide district selection. Case studysites were identified based on a vulnerability analysis using a "Principles, Criteria and Indicators" (PC&I) framework, together with a Geographic Information System (GIs). Rather than assigning weights and scores on an ad hoc basis, Principal Component Analysis (PCA)59was employed to provide a statistical basis for determining the effect o f each variable on the target variable, i.e. agricultural vulnerability.60 The drought- and flood-prone areas were demarcated and then overlaid with other maps containing information on other biophysical, social, and economic parameters. The basin was usedas the geographical unit inthe development o f the maps. By superimposing maps with the different parameters and showing their fluctuation from one year to another over a reasonable period o f time, a map depicting different degrees o f variation is produced which serves as the basis for selecting specific sub-areas for analysis. Inthis study, the secondary data on biophysical, social, and economic indicators such as gross cropped area, cropping patterns, groundwater availability, and an Infrastructure Development Index (IDI), among others, was compiled over different years spanning a 10-year period, for comparison purposes. The data was collected from various sources including the Survey o f India (SOI), the Census o f India (COI), the Central Ground Water Control Board (CGWB), the Central 58 Aandahl, Guro and Karen O'Brien, Vulnerability to Climate Change and Economic Changes in Indian Agriculture. 59 The statistical basis provided by this methodology provides better results than the conventional practices that "create" weight on an ad hoc basis. 6oThe area o f different vulnerable zones in the states selected for the study was estimated using the spatial analyst function of Arc GIS and drawn on GIs-produced maps. 101 Water Commission (CWC), the National Bureau o f Soil Survey and Land-Use Planning (NBSS & LUP), the National Atlas & Thematic Mapping Organization (NATMO), the Center for Monitoring Indian Economy (CMIE), the Indian Agricultural Statistics, Volumes I& 11, the Agricultural Census, and the Maharashtra and National Information Center (NIC). Using PCA, a vulnerability index was created which allocates degrees o f vulnerability to districts: low, moderate, high, very high, and extremely high. Districts were classified according to the index and maps were then developed for the states o f Andhra Pradesh, Orissa, and Maharashtra. An overlay o f different profiles for these states thus forms the basis for the selection o f the districts in each state, except for Orissa where official data was not available to allow for a comparison o f vulnerability over time. Consequently, the district selection in Orissa was guided by a combination o f (a) analysis o f secondary data and (b) the extent o f the geographical area which is considered to be liable to floods. Based on this analysis, some districts were deemedto face greater threats than others due to a combination o fhighbiophysical and social vulnerability andlimited infrastructure development. The final selection o f districts inthe selected river basins was made to purposely capture a range o f vulnerability patterns given their different socio-economic, technological, and biophysical conditions. FieldSurveys Despite the fact that more than two districts in each o f the three states were selected for climate projection (five in Maharashtra and four Andhra Pradesh), further prioritization o f districts was necessary in conducting field surveys due to limits in time andbudget. Thus, field surveys were carried out in two districts in each state. In the end, the districts o f Anantapur and Chittoor in Andhra Pradesh, Jagatsinghpur and Puri in Orissa, and Ahmednagar and Nashik in Maharashtra were chosen for the study. The objectives o fthe surveys conducted were the following: to assess the coping capacities and vulnerabilities o f communities indealing effectively with droughts and floods; and 0 to determine the factors that influence the effective implementation o f coping measures at field level. Institutional surveys were carried out to collect information on the central and state government plans and programs beingimplemented in the state and to ascertain their efficacy in enhancing the capacities o f communities in dealing effectively with climate variability and conditions o f extreme weather, including drought and floods. The field surveys sought to collect information on the communities' perceptions on (a) the intensity o f droughts/floods, (b) the crops grown in the region, (c) the change in cropping patterns, irrigation, livelihood options and migration, (d) infrastructure, (e) the availability o f financial services and schemes, and (f) the importance o f insurance. Through these surveys, an attempt i s made to undertake a critical review o fpolicy and community-oriented interventions that enhance the capacities o f communities to cope during extreme climate situations. In all, 1,640 households were surveyed: 570 households in Andhra Pradesh, 650 households inOrissa, and 420 households inMaharashtra. 102 Development of Tools for Institutional and Field Surveys Questionnaires designed for implementation in drought and flood circumstances as well as other Participatory Rural Appraisal (PRA) tools were used. Secondary data including sketch maps, transect walk, collation o f time-line information and trend-lines, seasonal cropping calendar mapping, institutional mapping, problem tree analysis, and problem and opportunity rankingwas collected. In addition, group discussions, interviews, focus group discussions, and institutional surveys were carried out. The questionnaires were pre-tested in pilot surveys in Rajasthan (a drought-prone area). It provided insights about the available quantitative information and its usefulness for the purpose o f the survey, and it was improved and modified accordingly. The lack o f proper recorded information at the village level posed a major constraint to quantitative/statistical analysis. Selection ofVillagesinIdentifiedDistrictsBasedon Analyses of Secondary Data The selection o f villages in each district was based on the screening o f village-level secondary data collected from the census office. This data was collected for parameters including village land area, land use, cultivated and irrigated land, and availability o f infrastructure including education, bank/credit, society, communication, power facility, and services like health care. The data was used for the preliminary selection o f villages within each district. These were later confirmed by discussions with officials ingovernment departments at the district level as well as other localizednon-governmental organizations and communities at the village level. Village Classification Based on Irrigation All villages lyingwithin a district were classified into one o f three levelsbased on their irrigated area as a percentage o f their total agricultural area: low (0-33%), moderate (33-66%), or high (66-100%). Village Classification Based on Infrastructure Development An infrastructure index was developed by considering the existence or level o f certain facilities and services at the village level including the availability of drinkingwater, education facilities, medical facilities, electricity, banks, agricultural society, andcommunication linkages. The villages were assigned to one o f four categories according to their irrigation- and infrastructure-based classification. The purpose o f this categorization was to select villages that were representative o f different contexts which may further the understanding o f the factors underlining the different levels of vulnerabilities. These broad criteria on irrigation and infrastructure are used to classify the villages ina matrix as the one shown here below. 103 Figure B.l Example of the Village Classification in the Infrastructure-Irrigation Matrix H i g h Forinstance: Forinstance: Infrastructure Korhate in Manesaudram in Maharashtra Andhra Pradesh Low Infrastructure For instance: Forinstance: N e r a m a t l a i n h d h r a Brahmanapalle in Pradesh Andhra Pradesh H i g h Irrigation L o w Irrigation Sampling The sample size, n, for eachtarget population i s computed usingthe formula by Murthy, (1977): n = N*n" N+Z-1 Here, n" =c2/e2 where c is the population coefficient o f variation and e i s the allowed percentage o f error. N i s the target population size (480). To obtain a representative sample, proportionate sampling based on landholdings were conducted. Records indicating household land category were collected from the Tehsildar61. The survey was conducted based on the various landholding categories: >4 acres (large farmers); 1-4 acres (medium farmers); < 1 acre (small farmers)62; and landless. After conducting the survey, all the data was coded and entered, andwere used for quantitative analysis. A reference manual was also developed to facilitate viewing and referencing. Measureof IncomeVolatility The Coefficient o f Variation (CV) i s used to understandthe extent to which household incomes are volatile to the impact o f drought/floods events63. The CV is simply a measure of the deviation of 'impact year' incomefrom 'normal year' income. C V i s defined as the ratio o f the Standard Deviation to the Mean (p) and is defined by the following formula: c, = - Q- P Person incharge o f maintaining landrecords o f villages ina panchayat 62 1ha =2.5 acres 63 Iti s imperative to mention here that income is considered a proxy for well-being, but is not tantamount to well- being. The latter is a holistic concept, encompassing livelihood security, food security, ownership characteristics, and respect insociety. Of these, the focus o f this study i s on livelihood and income security, and hence the nature o f employment emerges as an indicator inthe status o f well-being. 104 Inthis formula, 0 is standarddeviation and p is average income64.It is often represented as a percentage bymultiplyingthe above by 100. An advantage o f the C V is that it is free from the units o f the variables, and it thus permits comparisons with respect to their variability. The C V i s commonly used since it is a quantity without physical units. Although the C V indicates the magnitude o f variations, it fails to capture the directional shifts in income. As a substantial majority o f the surveyed households experienced drops inincome inan impact year, few 'outlier' households that showed an increase inincome duringanimpact year were segregatedout. 64p = (Normal income+ Impact income)/2, and02=[(Normal income -Meanincome)/'2 +(Impact income- mean income)"2]/2. 105 Appendix C: InternationalConference on Adaptation to Climate Variability and Change-Towards a Strategic Approach The International Conference on Adaptation to Climate Variability and Climate Change: Towards a Strategic Approach was held in New Delhi, India on December 7 - 8, 2006. It was organized by the World Bank,in collaboration with the Government o f India's Ministry o f Environment and Forests (MoEF), andthe European Union (EU). The Department for InternationalDevelopment (DFID), the Swiss Agency for Development and Cooperation (SDC), and the German Technical Cooperation Agency (GTZ) co-sponsored and participated in the event. The logistics o fthe eventwere organized by The Energyand Resources Institute (TEN). The main objective o f the conference was to take stock o f the latest progress in adaptation knowledge and practices, including policy and financing aspects in India and globally. This was made possible through a consultative process involving all stakeholders, from the donor community to the policy makers. The conference provided a forum for the exchange o f information on recent developments, programs, and challenges in India. The aim was to strengthen the commitment to future strategies and programs and to stimulate discussions on a possible action framework for integrating efforts towards strengthening adaptation in India. The outcomes include: 0 anincreased awarenesso fgovernment officials inIndia o fthe adaptation challenge and its implications for the development and implementation o f climate-related policies and programs; the enhanced ability o fparticipating state governments to develop climate-related adaptation policies andprograms; 0 the improvement o fthe stakeholders' knowledgebase inclimate-related areas; and 0 an enhanced coordination o f adaptation-related work programs among development partners. The event was well attended and received coverage in the international and localmedia. The conference's agenda and the presentations made by participants at the conference can be found at: http://www.teriin.ordevents inside.php?id=l7797. Below i s a summary o fthe keypoints raised at the conference on a session-by-session basis. Session 1. Inaugural Session The sessionhighlightedthat climate change evokes the need to consolidate and share information on contemporary initiatives aimed at reducingvulnerability and strengthening adaptation. The vulnerabilities o fdevelopingcountries to climate change underscore the urgency for taking proactive measures on the mitigation andthe adaptation fronts. Session 2. Adaptation and SustainableDevelopment The challenges posed by climate variability and change require prompt action. U s e h l information can be drawn from local communities' historical and current coping measures in dealing with climate variability and extreme events. The key challenge, however, i s to 106 incorporate principles o f equity and conflict resolution, beneficiary selection andbenefit sharing, transparency in project implementation, information sharing, and adequate representation o f minority communities inthe decision makingprocesses. Keypoints: Targeted research and development is required to enhance adaptive capacities o f communities - some initiatives in this context include revitalizing cooperatives and credit institutions, enriching the technological base, and strengthening the insurance system through the promotiono finnovativemeasures such as index-based, weather-risk insurance. National development policies and government programs need to integrate adaptation initiatives - increase resilience o f livelihood and infrastructure, improve governance, empower communities, and mainstream climate risk management. Watershed development and management should be pushed, especially in rain fed areas - priority intervention areas include provision o f potable water for consumption and o f protective irrigation for crops, while integrating livestock management and considering equity concerns. Session 3A. ClimateRiskAssessment:EmergingApproaches and Tools Screening tools provide a broad overview to project managers and development planners about the key climatic risks that could affect the implementation o f development projects and related investments bythe government, donor agencies, andother institutes. Keypoints: 0 Screening tools are required to mainstream climate risks into development planning and to provide guidance to project managers anddevelopment planners. 0 A framework is requiredinorder to integrate disaster risk management efforts with climate risk and development concerns. By strengthening adaptive capacities at the macro level, it is possible to create win-win situations that result in improvement in overall well-being o f the community/ecosystem. Session3B. Sector Impacts: Policyand EconomicImplications The implications o f climate change on the agricultural sector demands urgent attention. Stakeholders, such as scientists andpolicymakers, will be required to work together to integrate research into policy successfully. Key points: 0 There isneed for effective policy support to bolster the adaptive capacity o f farmers. 0 Autonomous adaptation in the agricultural sector faces constraints such as the time lag in responses, the unpredictability o f extreme events, and the lack o f extension services and technical guidance. 107 0 The cost o f adaptation should include support for buildinginfrastructureto promote research, development o f models for integrated assessments, and provision o f information on policy research. There i s a need for structural and non-structural investments in win-win strategies for better preparedness to climatic stress. This includes strengthening the methods for impact prediction o f climatic trends and strengtheningpublic services such as health care, sanitation, disease surveillance, access to vector and disease control services, integration o f health concerns inpolicies, efficient urbanplanning, housingregulations, water treatment etc. Session4. Adaptation Strategies: EmergingApproaches Planned and autonomous approaches to adaptation are not unrelated; while some risks, and necessary adaptation, can be identified and planned for, responses to other risks depend on the inherent flexibility o f systems. Keypoints: 0 Resilient systems are better equipped to meet basic environmental, economic, and social needs when faced by sudden climatic extremes. 0 Adaptation to droughts and floods at the local level depends on certain key factors such as: o the extent to which people are able to diversify into livelihood based on less climate- sensitive sectors; o the extent to which people have the ability to access information relaying early warnings and transport channels for goods and services; o the presence o f assets and security such as insurance and the resilience o f key infrastructure; o the condition ofkey environmentalresources, such as groundwater ina drought-prone area, and the presence o f institutional arrangements, self-help groups (SHGs), and credit groups; o the presence of institutional or self financial mechanisms (credit systems or remittances). Synergies need to be developed between plannedand autonomous adaptation by focusing on disaster management strategies, scientific information, and the use o f financing andinsurance mechanisms. 0 Targeted sector-wide interventions can be used to reduce the risks associated with climatic variability and extreme conditions. These include: o water-based intervention- the adoption o fbetter irrigationpractices, the rechargingo f groundwater through strengthening water harvesting structures, the revival and restoration o f community-based water conservation measures, and the revitalization o fwater user groups; o land-use-based interventions - control of soil erosion losses, crop advisory, development o f ago-forestry, and kitchen gardens which can promote nutritional security; o energy-based interventions - provision o f improved cook-stoves, promotion o f bio- energy crops, and promotion o fbriquettemaking; and 108 o livestock-based interventions - livestock management, pasture landdevelopment, and development o f fodder banks. The perceived barriers to adaptation are the lack o f timely weather-forecasting information, o f credit and savings, ando f appropriate technology. Experiences from various fields need to be pooled - including institutional, financial, and technical systems and participatory planning and implementation - to contribute towards livelihood diversification, provision o f a low-cost and input-internalized production system, natural resource management, resolution o f property-rights issues related to land and water, andenhancement o fcommunity-based local institutions' roles. 0 Communities employ various measures to deal effectively with the consequences o f climate change. These include looking at temporary, i.e. reactive, and permanent, i.e. proactive, models o f response. There is a need to identify and incorporate these measures into the enhancement o f communities' adaptive capacities. 0 Linkages should be drawn between the local level o f delivery and implementation o f adaptation measures and the efforts to include adaptation ininternational agreements. Session5. MainstreamingAdaptation: PolicyIssuesandOptions Key vulnerabilities to climate change need to be identified. There is also a need to strike a balance between adaptation andmitigation. Keypoints: 0 Partnershipswith other countries (both developing and developed) are requiredto address the challenges o f sustainable development and climate change in accordance with the goals o f the UNFCCC. There is also need to look at partnerships or linkages between governmental andnon-governmental sectors. 0 At the community level, sustainable livelihoodsneed to be promotedinorder to avoid future risks. In the case o f climate-sensitive occupations, income diversification i s the key to resilience. Agro-based industries are critical inproviding the support structure. Social capital forms an extremely important factor to enhance community resilience. Also, a balance needs to be struck between finance mechanisms o f credit, savings, increased engagement o f SHGs, development o f insurance structures, asset-building at household and community level, incentive-driven watershed management, anddisaster preparedness and mitigation. Session6. StrengtheningGlobalCooperationon Adaptation Key points: 0 To strengthen global cooperation on adaptation, there is a need to promote knowledge sharing and communication platforms among global actors, establish institutional frameworks for adaptation that are in line with development priorities, and assist in integratingadaptation inthe planningprocess. Some examples where global collaboration could be forged include the initiation o f small pilot projects planned to be integrated into ongoing larger scale development programs such as the watershed development programs and early warning systems. This would include risk 109 assessments, risk management including the development o f a national adaptation strategy, integration into investmentprogramming, and application o f adaptation tools and methods. 0 There needs to be adequate research on the economics o f adaptation, on methods to mainstream adaptation indevelopment, and on the involvement o f other stakeholders. 0 Current sustainable development practices related to rural livelihoods, disaster risk management and urbandevelopment need to be strengthened. Session7A. EU-India Cooperationon Climate Change and Adaptation Policies: Researchand PolicyActivities on Adaptation Key points: 0 Several EU-India collaborative efforts are already in place and a number o f new funding instruments have also been introduced in addition to individual member state programs. An integrated framework for the assessment o f climatic hazards, vulnerability, and risks, along with spatial planning, mapping, and modeling would be required for development o f robust adaptation and mitigation strategies. 0 For development and implementation o f strong adaptation projects stakeholder engagement i s critical. 0 Uncertainty related to the resolution o f results can often lead to faulty predictions and recommendations; these should be adequately researched before packaging the climate agenda. 0 Time and effort needs to be invested to promote researcher exchanges between India and the EUto buildhumanresource capacity anddevelop technical expertise. 0 Itis essential to recognize the obstacles to adaptation includinginstitutionalbarriers. Session 7B. State-LevelDialogueonAdaptationIssuesand Options: The Way Forward The state-level discussions were aimed at attaining a regionallstate-level understanding o f the diverse aspects o f adaptation, including the key components o f a state-level adaptation plan and institutional structures and delivery mechanisms. Keypoints: 0 Ecosystem management should be promoted as part o f a multi-pronged adaptation strategy. 0 Before incorporation o f climate change concerns into ongoing programs and plans, an assessmento f ground situations andrequirements should be made. 0 The various innovative adaptation strategies being employed by communities to cope with current climatic stress need to be documented and firther supported. 0 Prior planning i s essential for efficient adaptation and this can be reflected in terms o f changes ininfrastructure and resource use pattern. 0 Models o fdevelopment need to be revisitedto reduce unsustainable patterns o f development. 0 Identification o f `vulnerability' hotspots is needed for better preparedness. 110 0 More research i s required in the agricultural sector to promote growth o f tolerant crop varieties and address food security concerns. 0 Weather data management needs to be strengthened at all levels, as well as the requirement for instituting better warning systems. 0 There is a needto buildinter-andintra-departmental cooperation to address the challenges of climate change inan integrated manner. 111 Appendix D: Programs that Address Droughts and Floods in the Case Study Areas Water Resources Programs State-sponsored schemes Major, mediumand minor irrigation projects: Investments in irrigation were done throughfive irrigation corporations created between 1996 initiatives implementedby and 1998.Around 40% of irrigationpotential is yet to be tapped, but the WaterResources the cost of such investments is very high. Department, GoMexcept where noted Minor irrigation projects: Schemes with a command area of less than 2,000 hectares such as irrigation and percolation tanks, diversion bandharas, kolhapur and konkan type weirs, lift irrigation schemes, tube wells, renovation of malgujari tanks, and land drainageschemes. Cooperative lift irrigation schemes (implemented by the Apculture Department): Schemes which make irrigation water available to cultivators who cannot access rivers, canals, dams, etc.; they include isolated patches of land which cannot be served by major or medium projects. Water auditing: Pioneered for irrigation projects in 2003-04, the objective is to see whether water-use eflciency is as per design and to assess the eflciency of irrigationprojects' operations. Benchmarking o f irrigation projects: Since 2001 all major and medium irrigation projects have been evaluated against a set of 12 indicators covering system performance, agricultural productivity, financial aspects, assessment recovery ratios, and environmental and social aspects. Centrally-sponsored Command Area Development (CAD) Projects: Projects thatprepare schemes land to receive irrigation water, taking an integrated area development approach with an emphasis on balanced and comprehensive development of irrigated areas, water user associations, andfarmer participation. Accelerated Irrigation Benefit Programme (AIBP): I t largelyfunds major and medium irrigationprojects to create additional irrigation potential. Rural Infrastructure Development Fund(RIDF):This scheme by the National Bank OfAgriculture and Rural Development (NABARD) is meant to create rural infrastructure including major and medium irrigation schemes. National Finance Commission Grant: A Rs. 10crore grantfrom this source has been allocated in the TenthFive Year Planfor Orissa to augment traditional water resources. WatershedPrograms State-sponsored schemes Employment Guarantee Scheme (EGS): I t guarantees unskilled labor employment in rural areas, and uses this labor to create community assets, implementedby the such as soil and water conservation structures. Since the early 1990s, the Department of Water EGS has been extended to cover individual assets such as wells and Conservation,the horticultural plantations; it does not have a coordinated or integrated Employment Guarantee approach to watershed development. Scheme (EGS), and the 112 Department of Integrated Watershed Development Program: I t is intended to arrest soil Agriculture, GoM erosion, recharge groundwater and control runoff by treating the watershed from ridge to valley, strengthen and protect drinking water sources, develop waste land, degraded land and highly eroded lands, bring sustainability to agricultural production and generate employment in rural areas. Adarsh Gaon Yojana (Ideal Village Development Program): Theprogram aims to develop self-reliant and self-suficient villages in Maharashtra, with an emphasis on the development and regeneration of land and water resources; core activities include afforestation of watersheds, soil and water conservation (on both arable and non-arable land), drainage line treatment, and construction of water storage structures. Marathwada and Vidarbha Watershed Development Missions: Created by government resolutions to develop fallow and undeveloped land and groundwater tables in large rain-fed areas withpoor vegetation and water scarcity. Centrally-sponsored River Valley Project: Its main objectives are: (a) to prevent land schemes degradation by adopting a multi-disciplinary integrated approach to soil conservation and watershed management in catchment areas: (b) to implementedby the improve land capability and moisture regimes in watersheds; (c) to Department of Water promote land use to match land capability; and (d) to prevent soil loss Conservation,the from catchments in order to reduce siltation of multi-purpose reservoirs Employment Guarantee and (e) to enhance in-site moisture conservation and s u ~ a c erainwater Scheme (EGS), and the storage in the catchments to reducefloodpeaks and volume of runoff Department of Desert-Prone Areas Program (DPAP): The earliest area development Agriculture, GoM program (1973-74) tackles the special problems faced by areas constantly affected by severe drought conditions: these projects take up land and water resource development, aforestation, and pasture development using a watershed approach in watersheds measuring 500 hectares. Funding of Rs. 6,000 per hectare is shared by central and state governments in the ratio of 75:25. Integrated Wastelands Development Program (IWDP): I t implementsJive- year projects with community participation for sustainable rural developmentfollowing a ridge-to-valley `watershed approach' for in-situ soil and water conservation, aforestation, and water resources development in areas not covered by the DPAP or the Desert Development Program (DDP). National Watershed Development Program for Rain-fed Areas (NWDPRA): It takes a demand-driven participatory approach to develop local watershed plans and converge ongoing (agricultural) production programs; it is implemented with community contributions with the assistance of Project Implementing Agencies (PIAs). Watershed Development Fund:A Rs 200 crores-fund set up at the National Bank for Agriculture and Rural Development (NABARD) to promote integrated watershed development in 100 priority districts in 18 states (including Maharashtra) through aparticipatory approach. Additional Central Assistance (ACA): Under the Revised Long Term Action Plan (RLTAP) for KBK districts in Orissa, this Planning Commission program aims at drought prooJing and improving the moisture regime in these micro-watersheds to improve agricultural 113 productivity. NGOschemes Watersheds Organizations Trust (WOTR): Established in 1993, WOTR is the capacity building organization of the Indo-German Watershed Development Program (IGFDP); it is financially supported by the Government of Germany and NABARD. WOTR currently supports several NGOs in implementing a multi-sectoral, multidisciplinary approach to watershed development which involves continual interaction and exchange between various sectors and disciplines. Bharat Agro-IndustriesFoundation(BAIF): Set up in 1967 as a non-profit public charitable trust, BAIF has developed an array of watershed interventions topromote sustainable rural development,food security, and clean environment. These include water and land-resource development, livestock development, tribal rehabilitation, empowerment of women, community health, renewable energy and environment, and training in sustainable development. Agragame: This NGO works on natural resource management, watershed management in particular, in the tribal dominated uplands of Orissa. Agragamee spearheaded Sanjojana (meaning co-ordination), a network of 45 NGOs and individuals implementing projects in watershed development, N M , poverty alleviation, and community empowerment. Currently the Sanjojana network has seven projects in operation, six of which are located in southern tribal belt and one in the northern plateau. International donors The Western Orissa Rural Livelihood Project: Implemented by the Orissa Watershed Development Mission (GOO)with support from the DFID the project covers 274 watersheds in Western Orissa. The project's Watershed-Plus approach targets additional resources for the poor and marginalized; it has adopted innovative institutional arrangements to address the issues confronting the poorest and aims at organizing and enabling them to plan and implement participatory livelihood-focused development effectively. Orissa Tribal Empowerment & LivelihoodsProgramme (OTELP): Funded by DFID, the World Food Programme (WFP), and the International Fund for Agricultural Development (IFAD), theprogram facilitates the transfer of land rights to tribes before taking up any land- and water-based development measures. I t is being implemented in seven predominantly tribal districts of Orissa by the Scheduled Tribe and Scheduled Caste Development and Minorities and the Backward Class Welfare Department, Agricultural Programs State-sponsored schemes DevelopingDrought Resistant Varieties: Although several varieties have reportedly been developed in agricultural universities and state research implemented by the institutions, the extension staff of the Department of Agriculture hasfound Department of it diflcult to transfer thesefindings to thefield. Agriculture, GoM Dry LandHorticultureDevelopment Program: Its objective is to accelerate the coverage of new areas withfruit crops. Drip and Sprinklers: There is a state-funded sprinkler and drip irrigation scheme for sugarcane and a centrally-sponsored program for sprinkler and drip irrigation for horticulture and other crops. Sprinkler sets are also distributed under the National Oilseeds Production Program and the HorticultureProgram. 114 Rural Livelihoods Programs/Schemes State-sponsored schemes Swamajayanti Gram Swarozgar Yojana (SGSY): The SGSYfocuses on the implementedby the organization of thepoor at the grassroots level through a process of social Department of Rural mobilization. Its aim is to establish a large number of micro-enterprises in Development, GoM order to bringpoor families above thepoverty line by providing them with income-generating assets through a mix of bank credit and governmental subsidy. It is a holisticprogram of self-employment which covers activities such as organizing the poor into self-help groups, building capacity through training, selection of key activities, planning activity clusters, building up infrastructure, and providing technology and marketing support. Sampooma Grameen Rozgar Yojana (SGRY): The main objectives of the program are (a) toprovide additional wage employment andfood security and (b) to create durable community, social, and economic infrastructure in rural areas suflering from endemic labor migration or distress and in backward or calamity-prone rural areas. National Food for Work Programme (NFFWP): The objective is toprovide additional resources to the most backward districts of the country to further intensifj, the generation of supplementary wage employment and provision of food-security through creation of need-based economic, social, and community assets. National Rural Employment Guarantee Scheme (NREGS): The GoI enacted the National Rural Employment Guarantee Act to guarantee 100 dayslyear of rural employment to every BPL household from which an adult member volunteers to do unskilled manual work. The GoI meets the entire cost of wage payments, 75% of material cost, and a certain percentage of the administrative cost. The state Government meets the cost of unemployment allowance, 25% of material cost, and the administrative cost of the state council. State-sponsored livelihood Animal Husbandry: This Department's goal is to create self-employment support schemes opportunities for the educated unemployed youths and rural poor by implementedby the supplying milch animals and goat units to thepoorer sections of society. Horticulture, Animal Dairy Development: The dairy development sector aims to ensure Husbandry and Dairy guaranteed remunerative price to milk producers for their milk. Ongoing Development work under the state plan include establishment of milk chilling plants in Departments, GoM several districts in theMarathwada and Vidarbha regions. Horticultural Development: In Orissa, the National Horticulture Mission promotes plantation development and post-harvest activities such as production of planting materials, establishment of new fruit gardens, cultivation of powers, rejuvenation of old plantations, protected plantations and organicfarming. 115 Appendix E: A ConceptualFramework for a Maharashtra Drought Adaptation PilotinRain-FedAreas This appendix presents a framework for a drought adaptation pilot, using the state o f Maharashtra as an example. Background Inthe drought-prone regions ofMaharashtra, current copingstrategies appearto beincreasingly inadequate inreducing people's vulnerability to the effects and impacts o f climate variability and climate change. This suggests an urgent need for more effective integration of focused interventions to enhance the resilience o f communities to current and potentially more disruptive fkture climate conditions. An ongoing World Bank-supported study entitled Addressing Vulnerability to Climate Variability and Climate Change through an Assessment of Adaptation Issues and Options (hereafter "the Study") has analyzed issues and options for a more comprehensive integration o f climate-related matters into the Government o f Maharashtra's operations. This Study also supports the development o f a policy framework that integrates climate risk management into the development process. Preliminary findings o f emergingissues and options from this study were discussed inhigh-level meetings with Government o f Maharashtra (GoM) officials in May and August 2006. These meetings concluded with the recognition that the effects o f climate variability (to be further compounded by climate change) were costing the government significant time and resources, and that current crisis management was both expensive and sub-optimal. GoM officials supported the idea o f looking at opportunities to better address climate-related considerations in Maharashtra's development programs. To this end, GoM officials requested that the Bank team develop a framework for a state-level pilot project on drought adaptation, as part o f the ongoing Bank Study. It was agreed that the Department o f Water Conservation would be the designated nodal agency for this exercise, and it would work in collaboration with other departments, including the rural development, agriculture, andhorticulture departments. A draft conceptual framework was preparedinresponse to this request and is used as a basis to facilitate fkrther discussions with GoM officials. The note was presented to a group o f senior GoM officials at a meeting chaired by the Secretary o f Water Conservation and EGS in October 2006. Following a detailed discussion, it was decided that two designated officials, the Director o f Soil Conservation and the Director o f Social Forestry, would prepare responses to the draft document, which would then be integrated into a concept note. Responses were received by February2007 andhave beenincorporated into this document. DevelopmentContextand Rationale About a quarter o f India's drought-prone districts are in Maharashtra, with 73% o f its geographical area classified as hot and semi-arid regions. Maharashtra's 13 drought-prone districts account for 60% o f its net sown area. Even districts in the moderately assured rainfall zone are increasingly affected by vagaries in monsoon rainfall. As a result, a large part o f the state's predominantly rain-fed cultivable land suffers from crop failures and associated 116 hardships. The state faced consecutive years o f drought from 2001 to 2004, and the agrarian crisis has become so acute that farmers in the rain-fed regions o f the state have resorted to extreme measures, indicating a complete break-down o f copingmechanisms. At a macro-economic level, state-wide growth inthe agricultural sector has been slower than in either the industrial or service sectors, and agricultural productivity i s much lower than the national average. Notwithstanding unfavorable agro-climatic conditions, agriculture appears to be increasingly unproductive or only marginallyproductive due to a number o freasons including (a) highand unsustainable input costs (with heavy mono-cropping in some areas), (b) inadequate extension and knowledge services (resulting in suboptimal cropping practices), (c) little value added to support agri-businesses, (d) inadequate availability o f groundwater and poor soil- moisture conservation, and (e) market policies that do not support the majority o f farmers with small and marginal holdings (less than 5 hectares o f land). Furthermore, irrigation, which covers only 16% o f the total agricultural area, i s mainly accessible to the large farmers due to the inequitable distribution o f water resources governed largelyby power subsidies6'. Nonetheless, agriculture continues to be the main source o f livelihood for about 58% o f the state's population. All this suggests that much greater attention to rain-fed ecosystems and agriculture is needed in terms o f improved productivity, income, employment, and marketing. More importantly, providing support to nonfarm livelihood approaches and income-generating activities is also neededinrural Maharashtra. The Government o f Maharashtra supports a number o f development programs aimed at helping poor rural communities during drought episodes. Inresponse to the severe drought o f 1970-73 which affected 15 to 30 million people, Maharashtra led the way by introducing (for the first time inthe country) the Employment Guarantee Scheme (EGS) to provide gainful employment through relief works. Over the past thirty-odd years, the EGS has provided a substantial amount o f demand-led manual employment through labor-intensive public works (roads, percolation tanks, contour and `nala' bunding, horticulture-linked works), especially during off-season periods o f low employment opportunities. While the EGS is today considered a successful drought-relief program, it has not made a significant impact on reducing the drought-proneness o f the state. Nor has it reduced poverty through the creation o f productive assets and their maintenance or through buildinglong-term capacity and awareness on drought resilience. Apart from EGS, there are a number o f state and centrally-sponsored programs that seek to improve community resistance to drought, such as the Drought Area Development Program (DPAP), the Integrated Wasteland Development Programme (IWDP), the National Watershed Development Programme for Rainfed Areas (NWDPRA), the Jalswarajaya Water Supply and Sanitation Programme, the National Food for Work Program (NFFWP), the Swaranajaynati Gram Swarazgar Yojana (SGSY), the Sampoorna Gram Rozgar Yojana (SGRY) ,the Jawahar Gram Samridhi Yojana (JGSY), and the recently introduced National Rural Employment Guarantee Scheme (NREGS). There is also a rich heritage o f donor and NGO supported watershed development and rural livelihood enhancement programs. Despite these programs, the rural poor appear to be increasingly vulnerable to drought conditions inthe state as a whole. This 652% of the farmers inthe state have access to 70% of the irrigation; 80% o f the state's rural population does not benefitfrom any irrigation schemes. 117 i s evidenced by significant rural stress, groundwater mining, distress migration, and large and persistent inequalities inincomes and development outcomes. Overall, lessons learned from ongoing development programs show that there has been a limited degree o f success with the different approaches being implemented by different agencies under various framework and guidelines (e.g. seven different watershed conservation and development programs are currently ongoing inthe state). Achievement o f desired adaptation outcomes at the household and community levels has consequently been sub-optimal, even problematic. Therefore, coordinating between the multiplicity o f sectors and agencies involved i s a central challenge for developing a successful adaptation program inMaharashtra. This requires planning o f development programs on the basis o f local priorities and the lessons learned from implementing a variety o f rural development projects. There i s a need for a more strategic and effective integrated programming approach, with proper institutional arrangements, to help communities (a) save and recharge water, (b) adjust farming practices and cropping patterns in view o f water scarcity and market conditions, and (c) provide options other than agriculture for sustaining broader livelihoods. Linksto OngoingActivitiesandAddedValue The focus on long-term adaptation strategies in the context o f global climate change and integration o f climate-risk management into developmental planningi s a key area o f cooperation between the World Bank and the Government o f India. The Adaptation Pilot Program outlined in this concept note is inline with the programs ofthe Government o fIndia and Maharashtra which give highpriority to measureswhich would reduce the vulnerability o frural communities inrain- fed areas to climate-related risks, especially with a view o f reducing and then eliminating rural poverty and regional disparities. The GoM has been allocating significant resources to programs which provide drought relief in affected areas. The pilot will build upon, and facilitate synergy with, a number o f ongoing and relevant programs chiefly run by the departments o f water conservation, rural development, agriculture, water resources, and forests among others. In choosing areas for a pilot activity, coordination and complementarities would be established with the ongoing activities in groundwater management and projects in the water supply, irrigation, agriculture, and forestry sectors (figure E.l). The focus o f the pilot will be on long-term adaptation approaches and outcomes that go much beyond the usual programs for rural development, agriculture, forests, and water resources. 0bjectives The proposed Drought Adaptation Pilot will take an integrated approach to designing and implementing adaptation strategies to droughts in rain-fed areas. The proposed primary objectives o fthe Drought Adaptation Pilot concept are the following: 0 identify, analyze, and then recommend measures for state-level policy framework that i s supportive o f drought adaptation inrain-fed regions; 0 improve institutional and service delivery coordination between government andprograms to focus their actions andoutcomes on increasing resilience to droughts; 118 improve the capacity and awareness of small and marginal farmers living primarily in rain- fed areas, as well as options available to them for adapting to the impacts o fclimate change; test and evaluate measures that help (a) diversify livelihood options, (b) enhance productivity and management o f dryland ecosystems through sustainable management practices, (c) improve production systems through adoption o f innovative, cost-saving, and risk- minimizing technologies, and (d) safeguard vulnerable flora and fauna in their natural habitats. FigureE.l MaharashtraDroughtAdaptation Pilot (DAP) and Convergencewith Ongoing Programs MaharashtcaDrought Adaptation Pilot (RAP) and Convergence with Ongoing Programs Componentsof the Adaptation Pilot The pilot will involve the following main components/stages: Component1: Supportfor Developmentof a State-LevelPolicyFrameworkon Adaptation IncludingMonitoringandKnowledgeManagement There is a need to develop a policy framework that supports adaptation in the state and that can take the lessons and best practices from this pilot into a larger program. The expected outcomes o f the pilot would be improved awareness o f and capacities for drought adaptation options and approaches, a demonstration o f programmatic convergence, and a more effective packaging of focused interventions. The pilot would carry out participatory and real-time monitoring and evaluation to assess its performance in order to identify problems early on and suggest mid-term course corrections. A well-formulated monitoring strategy will therefore have to be an integral part o f the pilot program. Another key area o f work that could contribute to a basis for fbture policy work is knowledge management and sharing on drought adaptation options and experiences. The pilot will support activities on forming learning alliances for information exchange, education and communication, awareness-raising campaigns, training workshops, and creating a network o f knowledge centers including websites that provide information on 119 adaptation. Effective monitoring and evaluation and knowledge management systems will support drought adaptation activities beyondthe period o fthe regionalpilot programs. Component2: StrengthenedConvergencewith OngoingPrograms The pilot will strengthen convergence with ongoing programs to focus the impact o f these programs on drought adaptation. For instance, the convergence with EGSNational Rural Employment Guarantee Scheme (NREGS) would be established to support a range o f allied activities, including public works, land development, soil and water conservation, crop planning, horticulture, and agro-forestry through people's participation. Further, convergence through EGS could include renovation and restoration o f communally-owned malgujari tanks, which if repaired could give significant support to farmers. Apart from the EGS/NREGS Program, the pilot can learn lessons and build upon a plethora o f relevant activities in Maharashtra. These include (a) the existing watershed development programs - the Maratwada and Vidharbha Watershed Missions and the Bank-supported Jalswarajya community-based water supply initiative, (b) the aquifer mapping and management pilots and Maharashtra Water Sector Improvement Project, (c) the Department o f Agriculture's programs for drought-prone area, and (d) the work beingdone by the Forest Department, which owns a major portion o f the degraded areas o f drought-prone districts inMaharashtra. Component3: InstitutionalSupportandCapacityBuilding The pilot's design and implementation would require specialized professional inputs and the participation o fNGOs, line departments, local governments, and community-based organizations (e.g. rural cooperatives and self-help groups which could be formed by government department programs for agriculture, rural development, and forestry). In addition, the pilot can aim to develop farmer groups and user groups to take collective action. Further, these institutional groups could be linked with tied grants and micro-finance programs. The capacities and awareness levels o f all o f these institutional players will be developed through a well- coordinated capacity-building program at the grassroots level. Players insuch a sensitization and capacity-building exercise could also include the Social Forestry Directorate, which has on-going outreach work indrought-prone areas. Component4: Community-LevelPlanningandImplementationof DroughtAdaptation Plans The findings o f the Bank-supported study on climate change and adaptation indicate a strong need for adaptation solutions in Maharashtra. These solutions should be based on multi-sector interventions and take into account the full range o f local conditions. Therefore, it is proposed that the planning o f adaptation interventions and packages be done by the communities themselves. This will require the development o f micro-plans on drought adaptation, possibly linkedwith tied grants, taking into account ongoing programs and missinglinks and gaps inthe village-level activities. Detailed studies on integrated planningmethodologies for micro-planning in drought-prone areas would help identify ways to institutionalize such micro-planning activities infuture government activities inthe field. An initial assessmento f the rain-fed regions in Maharashtra points to the following categories o f interventions, which have the potential to make a difference inbuildingresilience incommunities. 120 Management of Common Natural Resources(Water and Land) Inamajority ofthe drought affected areasinMaharashtra, the degradation ofthe soil inrain-fed cultivate areas due to denudation o f tree and grass cover and the ineffectiveness o f programs to check soil erosion have been major problems. There is a need, therefore, for more efficient use o f local rainfall to improve soil moisture and recharge groundwater. Further expansion o f surface and groundwater irrigation through major or medium irrigation projects and tubewell projects is not feasible andmany o f the irrigation facilities inrain-fed areas require repair andmaintenance. Given these circumstances, the central focus o f the pilot will be on activities at the micro- watershed level, building on the Marathwada and Vidharbha Watershed Missions as a strategic entry point. Specific activities would take into account both surface and groundwater catchmend6 in the early stages o f planning. The focus will be on (a) reducing demand, (b) developing collective protocols and mechanisms for the proper maintenance o fwatershed works, and (c) regulating entitlements for equitable access and use o f water and biomass. This will be done through technical and participatory management techniques and systems involving Village Self Help Groups (VSHGs), grassroots NGOs, and Panchayati Raj Institutions (PRIs) which endorse the Hariyali Guidelines towards empowerment o f the PRIs. These watershed investment activities will also build upon available best-practice models and the promising work o f several NGOs inthe state, which chiefly include the work o fthe Watershed Organization Trust (WOTR) under the Indo-German Watershed Development Program, BAIF, AFARh4, and MSSM. It will also examine feasible mechanisms for introducing community-led management o fwater demand to ensure that additional water, captured through soil and water conservation, is not used up completely for expansion o fimgated areas or to support water-intensive agriculture. Also, biodiversity conservation in these drought-prone areas needs to be addressed in a more holistic manner. Given the tough competition for limited resources, appropriate water and food management i s necessary to reduce the conflicts between human beings and wildlife, such as attacks on humanbeings and cattle over utilization o f habitat and crop raiding o f agricultural and horticultural fields. Diversification of Production Systems and TechnologyInnovation Incertain parts of the rain-fed regions in Maharashtra, there is a significant practice of mono- cropping and utilization o f farming practices (e.g., purchase o f commercial seeds) that are conducive only to imgated conditions and therefore detrimental interms o f overall productivity. In addition, the use of pesticides is excessive in some rain-fed areas. These practices result in sub-optimal production levels and low price realization. The pilot would liaise with ongoing programs to strengthen the advisory and extension support services to farmers inorder to suggest ways to improve existing support services for dryland agriculture. The focus o f the extension services could range from recommendations on the timing and quantum of application o f water and fertilizers, to recommendations on the diversification o f cropping systems to include short- duration, drought-resistant seed varieties and practices that significantly increase the yield 66There are significant variations in the state's hydrogeology that will have to be considered in order to determine the feasibility of, andpotential for, groundwater recharge activities. 121 potential o f coarse cereals, pulses, oilseeds, and fibers that are the backbone o f a rain-fed agriculture. As part o f this process, the pilot will organize learningplatforms that bringscientists and farmers together in order to extend lessons learned from dryland agriculture research to the farmers' fields and agricultural projects. The focus o f the entire exercise will be on optimizing the available gene pool, minimizing the input costs o f agriculture, and increasing the revenue per unit of land and per unit of water inrain-fed areas and irrigated tracts, hence increasingprofits from sustainable agriculture overall. In addition to the focus on agricultural development, the pilot would look at issues such as apiculture, vermi-composting, and the prevention o f forest fires. The pilot would also develop systems to familiarize rural communities with `watershed plus' activities that attend to end-uses o f harvested rainwater and promote livelihoods, including the development o f fodder banks in order to meet the increased demand for stall feeding andthe promotion o f leasing arrangements o f common lands to the landless for the cultivation o f fodder crops and the promotion o f fisheries. Nonfarm Livelihood DiversiJication In some drought-prone areas where agriculture is becoming increasingly unviable, it would be necessary to promote nonfarm livelihoods. This strategy o f income diversification through alternate livelihoods can play an important role in building farmers' long-term coping abilities duringdrought conditions. For instance, women SHGs andother collective action groups can be mobilized to plan and implement a variety o f income-generating activities with the assistance o f grassroots NGOs, market-based institutions, and micro-finance institutions. A wide variety o f approaches have beentried throughout differentparts o f the country, from which lessons will be distilled for locally-appropriate piloting. The emphasis, however, would be on developing institutional and community-based mechanisms to generate and share economically-viable options for livelihood diversification, so as to improve resilience to drought and to reduce stress migration and its attendant strain on urban areas. Economic Support Toolsand Marketing Poor credit availability and poor marketing systems are crucial impediments in drought-affected areas, as a result o f which the poor farmers suffer on various accounts. The farmers in these regions, among the poorest inthe state, usually take out large loans from the informal system, the bulk o f which comprise unlicensed money lenders. Inpractical terms, credit is required for the input cost o f seeds, fertilizers, and pesticides. To strengthen the lines o f credit from cooperative banksto farmers, the pilot will look at ways to troubleshoot the existing initiatives o fpublic and private micro-finance institutions to ensure effective access to credit. For example, the National Bank of Agriculture and Rural Development (NABARD)has the capacity to create systems for providing credit lines and training in credit assessment, disbursal, and recovery to smaller nodal agencies which, in turn, can distribute these lines o f credit to individual entities. There are, however, several bottlenecks that impede effective access, as demonstrated in the low adoption o f farmers' insurance, and more recently, in the disbursal problems o f the drought-relief credit package announced for farmers in Vidharbha. The pilot will also endeavor to develop and test innovative economic instruments for weather-related insurance and other social safety-net funds 122 established as contributory funds at the community level to tide the farmers over during periods o f immediate distress. InstitutionalAspectdPartners The pilot program will help to further clarify the roles and responsibilities o f different institutions/stakeholders, including state and local governments, NGOs, communities, sector support service agenciedline departments, local banks, and other financial institutions, in implementingsuch programs on a larger scale. Adaptation success stories often correlate with the presence o f good NGOs, suggestingthat grassroots civil society organizations can be made an important part o f an adaptation strategy. To the extent possible, the pilot will try to make use o f existing rural development, agricultural programs, and micro-financing schemes running successfully inthe state. Institutional mechanisms for channeling tied grants to community-based groups and PRIs will needto be explored. GeographicalCoverage A pilot assessmentwill be performed intwo or three districts, representing four to five different agro-climatic zones and geographical and socio-economic conditions, with coordinated inputs from different departments and agencies (e.g., Agriculture, Horticulture, Soil Conservation, Forest, Minor Irrigation, IMD, GSDA), research institutions, and NGOs (e.g. BAIF, AFARM, MSSM, WOTR). The final selection o f districts will be made upon assessing a range o f factors, . including: .. the interest o f the districthlock level authorities; the availability/skill potential o f NGOs and other institutions to participate in the implementation; the availability o f a range o f geo-hydrological and agro-climatic conditions inthe area, so as to be able to test and evaluate a mix o f interventions; . the existence o f relatively successful key programs such as EGS/NREGS and watershed activities inthe area; and the potential to maximize learning value and usehlness o f the pilot exercise, including the formulation o f a largerprogram. ImplementationArrangements A source o f financing will have to be identifiedto support this pilot. Some financial and/or in- kind contribution will be expected from the state government. The overall exercise will take place incollaboration with andunder the guidance o fthe Departmento f Water Conservationand EGS. A nodal agencykontact point at the state government level will be identified and will coordinate/facilitate/oversee the work in the state. Implementation arrangements at the district and block levels will be discussed and identified in consultation with the state nodal agency. Pilot programs will involve a range o f stakeholders including state and local governments, NGOs, SHGs, WAS,consulting firms (for training), local banks, etc. A steering committee chaired by the Secretary o f the Department o f Water Conservation and EGS could be 123 established, that includes representatives from other departments (such as the rural development, agriculture, revenue and planning departments), NABARD, andk e y NGOs. Table E.l Potential Impacts and Responses for Addressing Vulnerability to Climate VariabilityandClimateChange inMaharashtra Sectors Economic Ecosystem Impacts Impacts Impacts on Poor PotentialInterventionOptions Irrigation, Drought, Impact on Irrigation 0Restorationandmore efficient management Tank floods, water biodiversity potential, of tank system Management supply, and vulnerabilityto Adoptionofmore efficient methodsof power drought irrigation 0 Desiltationof canals and feeder channels and restorationof lakes Demand-sideresponsestrategies,such as water pricing, water rights, socialregulation of privateborewells, reforminwater laws, etc. Agriculture and Loss o f Pestsand Foodpoverty Cropdiversificationandlivestockproduction Livestock production diseases and andmanagement strategies malnutrition Developmentof new technologiesand improvedextensionservices Changes inlandusepractices, including changesintiming, croppingsequence, and intensityof production;landrightspolicies Minimizing external andhighcost inputs (pest andsoilproductivitymanagement) Seedmanagement for normal andcontingent years Communitymanagement of livestockand fisheries Changesinmarketingsystems andrisk financingoptions (e.g. creditandinsurance) Natural Loss of assets Depletionin Foodand Managementof commonlandandprivate Resource and groundwater, incomepoverty fallow land as buffer for drought vulnerability Management, livelihoods; soil cover, and for humanand Changesingraze-landmanagementpractices Livelihoods migration moisture content livestock (time,location,duration) systems Naturalregenerationof biomassthrough social fencing Promotionof non-fmincomegeneration andlivelihood opportunities Economic Lossof Low Foodand Changesinrisk financingoptions (crop Instruments production productivity income insurance, weather-based insurance) andlivestock poverty,large Socialsafetynets debts, acute economic stress during calamity conditions 124 Appendix F: Programfor Stakeholder Consultations Date & Place Event Objective Participants Responsibility June 3-4,2004 Roundtable To examineissues andoptions MoEF, DEA, IIM, WorldBank NewDelhi Discussion surroundingthe emergingissue TIFAC, Development of adaptationto climatechange Alternatives, TERI andimplications for Indiaand the rest ofSouthAsia August 16,2004 Roundtable To discuss andseek feedbackon MoEF, DEA, IIM, WorldBank NewDelhi Discussion the WorkingDraft ConceptNote TIFAC, Development of the study Alternatives, TERI February1,2005 Review To discuss andfinalizethe study GovernmentofIndia, WorldBank NewDelhi Meeting conceptnoteandterms of (MoEF, MOA, reference MoRD, MoWR, MOST,DEA, PC), Institutions and Experts (IIM, MU, IIT, IPCC), Bilaterals (DFID) May9,2005 Launch To launchthe study andto GovernmentofIndia GovernmentofIndia NewDelhi Workshop discuss the studymethodology ministries and andWorldBank with a wide rangeof agencies, RMSI, stakeholders TERI, NGOs, and TAG extlerts October26,2005 Meeting To discuss and seek feedbackon Membersof TAG, WorldBank NewDelhi with the studymethodology MoEF, andother Techmcal central government Advisory officials Group (TAG) and policy extlerts April 25,2006 Meeting To discuss the status andnext MoEF, WorldBank, MoEF NewDelhi with steps of the study TERI Secretary andother senior officials o f MoEF April 26,2006 Technical To discuss andseek feedback on Membersof the TAG TERIandWorld NewDelhi Advisory preliminaryresultsofclimate (selectIndian Bank Group modelingandhouseholdsurveys experts) (TAG) Meeting April 29,2006 Roundtable To share the objectivesofthe Seniorofficials from DepartmentofRelief Bhubaneswar Discussions studyandmethodologyinorder various state andRehabilitation to set the stage for a morein- departmentsand andDepartmentof depthdiscussiononce the NGOs Water Resources, preliminaryanalysisofdata was OSDMA, complete Governmentof Orissa 125 M a y 8,2006 Roundtable To share the objectives o fthe Senior officials from Department o f Rural Mumbai Discussions study and methodology inorder various state Development and to set the stage for a more in- departments Water Conservation, depthdiscussiononce the Government o f preliminary analysis o f data was Maharashtra complete; to initiate a rapid assessment o f the existing line- department programs for coping with floods and droughts in Maharashtra August 28-29, Bilateral To discuss preliminary study Chief Secretary and World Bank 2006 Meetings findings from the case study Secretaries from the Bhubaneswar withstate Departments o f departments Planning, Revenue, Water Resources (WRD), Rural Development (RD), Reliefand Rehabilitation, and Agriculture (DOA) August 31- Bilateral To discuss preliminary findings Secretaries from the Department o f Rural September 1,2006 Meetings from the case study and the DOA, Environment, Development and Mumbai with state follow-up, in-depth analysis RD,and Water Water Conservation, departments undertaken incooperation with Conservation; Government o f and state government agencies officials from GSDA, Maharashtra Roundtable Irrigation, and other line departments November 3-4, Roundtable To discuss preliminary findings Department o f Rural Department o f Rural 2006 from the case study and the Development and Development and Mumbai follow-up, in-depth analysis Water Conservation, Water Conservation, undertaken incooperation with Government of Government o f state government agencies Maharashtra Maharashtra December 7-8, International To disseminate emerging State governments, Department of Rural 2006 Conference adaptationprograms and Government o f India Development and New Delhi on strategies; to present emerging ministries and Water Conservation, Adaptation international models o f agencies and experts Government o f to Climate adaptation actions and programs; Maharashtra Variability to discuss emerging findings o f and Change the studv December 18-22, Roundtable To initiate a rapid assessment o f U S Geological Department o f Water 2006 and Bilateral the existing line-department Survey, WRD, Resources, Bhubaneswar Meetings programs for coping with floods Watershed Government o f with state and droughts inOrissa Development Orissa departments Mission, Department o f Revenue, DOA, OSDMA, ORSC, Department o f Planning & Coordination, Government o f Orissa 126 Appendix G:A Descriptionof the Economic Model The structure o f the economic model has been determined by the outputs, sequence and constraints imposed by the agronomic model EPIC. The approach in this report extends and draws upon previous work where EPIC has been linked to economic models to determine the impacts o f agricultural policy inthe EU.67 Following the steps in EPIC it i s assumed that once land area has been allocated to different crops, farmers make farm management decisions inresponse to the observed weather pattern, by altering fertilization, watering etc to optimize payoffs. This suggests a two stage modeling process: inthe first stage, the crop mix is determined based on the anticipated weather outcomes andexpectedprofits from each crop. Once crop choices andlandallocation decisions are made, planting, fertilizing and irrigation occurs andis adjusted inresponse to the actual weather. Accordingly, by backward induction, the second stage o f the model, where farm techniques are determined, is solved first. Specifically, for any given crop i (1,n), the farmer varies farm E management techniques to maximize the per hectare payoffs from the crop: where xi is per hectare profits from crop i,pi i s price of crop i,yi(Z,O is yields from EPIC based on farm management strategy Zi and a vector representing climate event T.The properties of yi(Zi,Oare determined by EPIC and typically appear to exhibit single peaked behavior. Specifically: for some zk E Zi 3 a Z such that for z < Z, 2$i /dzk > 0 and for z > Z , 2$i / dz, <0. c@J are the corresponding costs o f farm strategy Zi. Costs are linear ininputs and are expressed as: ci(ZJ = C c k V k ,where ck i s the cost coefficient o f input p, k E (1,h). r h k=l represents the many dimensions o f climate incorporated in EPIC and includes among other factors the daily level and distribution o f temperature, rainfall, soil moisture and carbon dioxide. Thus for an element rj r, tJyi/dT' E >Iandi3yi2/ 0 X j 2>I0. Since the focus is onthe farm household at the district level it i s natural to assume that all prices are exogenous. For purposes o f the simulations a range o f discrete farm management strategies are used covering variations and different combinations o f (a) seeding, (b) fertilization, (c) irrigation and (d) tillage techniques. Let * denote the optimum value o f inputs from the maximization o f equation (l), then in stage two farmers determine land allocation based on the expected profits from each crop: 67See, e.g. FIPM2005. 127 where ni=E(zl:),E isthe expectations operator with expectations definedover climate events that determine yields for any given farm management strategy, Li i s land devoted to crop i, and D(ni)is the sum of negative deviation of payoffs from crop i and 0 is a risk aversion parameter. Together these terms capture risk taking behavior in a simple way using the familiar mean-variance approach. Other more complex methods are left for future extensions o f the framework. Equation (2) i s maximized subject to a series o f technical constraints. The key among these are a land availability constraint: 3. CLiI L n land availability constraint i=l where z i s the given endowment o f land. Andawater supplyconstraint: n 4. C W i L iIF water supply constraint i=l where wi i s water consumptioncoefficient for crop i, and w is total water supply. For completeness a number o f additional constraints are incorporatedinto the spreadsheet for seeds, fertilizer and labor but are not allowed to bindsince there i s little evidence o f quantity constraints o fthese inputson farms. InAndhra Pradesh, anadditionalconstraint isimposed requiringaminimumof0.5 hectares is devoted to rice. This captures survey evidence showing that rice i s grown on these farms to meet subsistenceneeds for fodder and consumption. - 5. L, 2 L, subsistence rice consumption constraint where =0.5 ha, andsubscript r denotesrice crop. Since not all farmers are identical, the analysis distinguishes between three types o f cultivators, depending on the availability o f land. Subsistence farmers are classified as those with landholdings up to 2 hectares. They are driven by subsistence needs and the imperative to survive takes precedence over commercial considerations. This is modeled as a safety-first constraint where the primary objective i s to earn a threshold amount o f Rs. 12,000, which is the subsistence threshold in the National Sample Survey (NSS). Medium farmers have holdings between 2 and 3.5 hectares and large farmers have holdings in excess o f 4.5 hectares. The farmers attempt to maximize the commercialpayoffs from farming, as defined in equations (2) - (5). To capture subsistence behavior an additional simple constraint i s imposed for small land holders: n 6. x{niLi}2Y smallfarm subsistence constraint i=l 128 where = Rs. 12,000. The simulations allow for rational expectations where expectations are based on the known distribution o f events and adaptive expectation where there i s learning and expectations are based on past series o f events. SensitivityAnalysis of the FarmEconomicModel The results o f quantitative assessments are dependent on key modeling decisions and assumptions. It i s therefore important to assess the robustness o f the results to various plausible changes. This section describes sensitivity assessments o f the results presented in the main report. The focus is on (a) crop and input prices bi) levels o f farmer's attitude towards risk (e) assumptions on farmer's knowledge o f climate events and (d) levels o f water availabilityhhortage. The farm economic model finds that inAndhra Pradesh, irrespective o f farm size, the bulk o f the cropped area i s devoted to groundnut. Under climate change scenarios, groundnut still remains the most profitable crop, and consequently there is no change in the planting mix. Table A.l assesses the extent to which changes in input parameters (including price o f crops, total labor cost, and fertilizer cost) would be requiredto alter the current findings. Table A.1reports critical percentage changes that makerice andjowar competitive to groundnut. Table G.1 Critical Percentage Changes for Diversificationout of Groundnut in Andhra Rice Jowar Price Total Labor Cost FertilizerCost Price Total Labor Cost FertilizerCost Baseline 27 -57 -107 44 -229 -188 A2 23 -23 -60 29 -151 -108 82 25 -50 -103 47 -246 -186 Table G.1suggests, for example, that under the baseline scenario the price o f rice has to increase by 27% (from 5.7 to 7.3 Rskg) to induce farmers to start growing rice, while this is 44% (from 5.3 to 7.6 Rskg) for jowar. Fertilizer cost for rice needs to be reduced by as much as 107% to makerice competitive-anunrealistic situation evenifsubsidized. Turningnext to variations inthe risk aversion parameter, the degree o f risk aversion is likely to be negatively associated with farm size - asset holdings and wealth, so results are presented for medium farmers as an illustration o fthe insensitivity o fthe cropping mix results.68 Binswanger (1980) conducted an experimental study with 330 individuals from arid areas inAndhra Pradesh and Maharashtra and found that more than 80% were "moderately" risk-averse. Only 2% o f individuals were found to be "extremely" risk-averse. 129 Table G.2 RiskAversion and CroppingMix inAndhra Pradesh risk Baseline ~ aversion ~ A2 82 coefficient Rice GN Jowar Rice GN Jowar Rice GN Jowar 0 0.5 3 0 0.5 3 0 0.5 3 0 1 0.5 3 0 0.5 3 0 0.5 3 0 2 0.5 3 0 0.5 3 0 0.5 3 0 Table G.2 gives an example o f medium farmers cropping patterns with varying levels o f attitude towards risk. Higher risk coefficient implies a higher degree o f risk aversion. Under all climate scenarios, the crop mix i s not sensitive to whether farmers are risk-neutral or ri~k-averse~~. Turning next to Maharashtra. The economic assessment finds that, on farms where there i s adequate water, sugarcane i s the dominant crop due to its significantly higher profits. Table G.3 shows the levels o f crop prices, fertilizer costs, and user charge o f water necessary to make bajra and jowar competitive to sugarcane. For this, bajra andjowar prices have to be increased more thanthree fold. The level o fuser charge ofwater for sugarcanefarmer is assumedto be minimal (1.2 Rs/mm) to reflect simply costs o f irrigation without water charges. Ifwater charges were increased to the levels shown in Table G.4, the area o f land allocated to sugarcane would shrink and farmers would start growing less water-intensive crops. The analysis with variations inrisk aversion shows that sugarcane remains dominant regardless o f whether farmers are risk-neutral or risk-averse (Table G.4). Table G.3 Competitive Crop Prices, Fertilization Cost, and User Charge of Water in Maharashtra(FWkilogramand Rdmm) Price of Bajra Fertilizer Price of Jowar Fertilizer User Charge of Bajra cost Jowar cost Water Baseline 24 -82 28.5 -123 59.3 A2 17.8 -50 20.2 -80 43.8 B2 19.5 -55 21.3 -85 41.8 Note: current price ofbajra andjowar is 5.25 Rskg, while their fertilization cost is 12 Rskg risk Baseline A2 B2 coefficient Bajra aversion Jowar Sugarcane Bajra Jowar Sugarcane Bajra Jowar Sugarcane 0 0 0 3.5 0 0 3.5 0 0 3.5 1 0 0 3.5 0 0 3.5 0 0 3.5 2 0 0 3.5 0 0 3.5 0 0 3.5 69 A substantial majority o f individuals in these areas are moderately risk-averse, and this is recognized in the literature (e.g. Anderson and Dillon, 1992) as equivalent to the risk-aversion factor o f 1. The exercise allows risk- aversion coefficient up to the factor o f 2, the level beyond which is very unlikely to be found among households in these areas. 130 The Role of Expectations: Expectations on climate events also play a significant role in determining how farmers respond to climate change. The main analysis presented i s with the case where farmers holdrational expectations - based on the actual distribution o f climate events. Further analysis shows that, even when fanners are not as well-informed, the relative profitability among the crops under consideration remains unchanged - groundnut being the most profitable in Andhra Pradesh and sugarcane in Maharashtra. Thus, the findings are not contingent on whether the farmers have hll knowledge o f climate change or make planting decision based on year-by-year past experience o f weather events (table G.5). This reflects the extreme conditions inthese areas. Table G.5 Farmers'Knowledgeof ClimateEventsand CroppingMix Knowledge Andhra Pradesh Maharashtra Base of Climate Events Rice Groundnut Jowar Bajra Jowar Sugarcane Rational Expectations 0.5 3 0 0 0 3.5 Past 20 years 0.5 3 0 0 0 3.5 Past 10years 0.5 3 0 0 0 3.5 Past 1year 0.5 3 0 0 0 3.5 According to EPIC model, farmers are expected to deal with weather events that are more extreme more frequently compared to the past. Although higher returns are possible, the farmers will encounter undesirable outcomes in A2 that lie outside their experience (figure Gl). It is important to note that the assessment so far is based on the optimistic notion that water and irrigation supplies are unaffected by climate change. With warmer conditions and higher rates o f evapotranspiration, water demand i s likely to increase and water tables, which are already in decline, may deplete even further. So, it i s important to assess how growing water shortages, combined with climate change, affect farm incomes. FigureG.l Distributionof GroundnutYields inAndhra Pradesh I 251 I 20 15 10 --------------------- 5 0 baa even% gow events (lower yields) (higher yields) Significant water shortages lead to shifts in cropping patterns, a decline in cropped area and lower incomes. With shrinking water availability the balance shifts towards jowar, (figure G.2). 131 Figure G.3 shows the consequences o f water shortages on farm incomes. Even medium-sized farms fall below the "survival threshold" o f Rs 12,000 in the A2 scenario, when the constraint binds sufficiently. Though these results are illustrative rather than predictive, the implied magnitudes highlight the importance o f strengthening water conservation initiatives across the state. FigureG.2 Area AllocationandWater Supply,MediumFarm,Andhra Pradesh (ha) 3 5 3 2 5 2 1 5 1 0 5 0 240 230 220 210 200 IRice 0aoundnut m Jowar FigureG.3 ProfitsandLevels ofWater ShortageinA2 ScenarioinAndhra Pradesh 20000 - 15000 - 10000 - 5000 - 0 10 20 30 40 50 60 70 80 Levelof Want-r Shortam-