Adapting a hazards risk model to water scarcity in rural india aurangabad case study

Adapting a Hazards-Risk Model to Water Scarcity in Rural IndiaAurangabad Case Study by Paige K Midstokke B.A Political Economy University of California, Berkeley 2013 Submitted to the Institute for Data, Systems, and Society and the Department of Civil and Environmental Engineering in Partial Fulfillment of the Requirements for the Degree of Master of Science in Technology and Policy and Master of Science in Civil and Environmental Engineering ARCHIVES at the MASSA Massachusetts Institute of Technology OF T ECHNOLOGY February 2018 C2018 Massachusetts Institute of Technology FEB 28 2018 All rights reserved +- jiia Uu Signature of Author: Ir acue LIB RARIES Paige Midstokke Technology and Policy Program redacted ~ignature Certified by: ,partment of Civil and Environmental Engineering December 8, 2017 James L Wescoat Jr / (1 Aga Khan Professor, Department of Architecture Department of Urban Studies and Planning Certified by: $ i gnature redacte Thesis Supervisor Dennis McLaughlin H.M King Bhumibol Professor of Water Resources Management Department of Civil and Environmental Engineering, Thesis Reader Signature redacted Accepted by: Munther Dahleh William A Coolidge Professor, Electrical Engineering and Computer Science Director, Institute for Data Systems and Society //I redacted Accepted by: _Signature Jesse Kroll / x Professor of Civil and Environmental Engineering Chair, Graduate Program Committee 77 Massachusetts Avenue Cambridge, MA 02139 http://Iibraries.mit.edu/ask MITLibraries DISCLAIMER NOTICE Due to the condition of the original material, there are unavoidable flaws in this reproduction We have made every effort possible to provide you with the best copy available Thank you Some pages in the original document contain text that runs off the edge of the page p.93 Adapting a Hazards-Risk Model to Water Scarcity in Rural IndiaAurangabad Case Study by Paige Midstokke Submitted to the Institute for Data, Systems, and Society and the Department of Civil and Environmental Engineering on December 8, 2017 in Partial Fulfillment of the Requirements for the Degree of Master of Science in Technology and Policy and Master of Science in Civil and Environmental Engineering Abstract The objective of this project is to improve the responsiveness of District Planning to rural water scarcity in India Through engagements with the Groundwater Survey Development Agency, and Maharashtra State Government Water Supply and Sanitation Department, we selected Aurangabad District to conduct field visits and develop a model that can spatially represent risk of villages to water scarcity Within Aurangabad District, Vaijapur block was selected as a case study due to its drought effects and high water tanker usage in the past five years This thesis develops a disaster risk metric for water scarcity, using an analysis of potential hazards, socioeconomic vulnerability, and policy responses to assign a "disaster risk score" to each village Risk is seen as a function of hazard, vulnerability, and government capacity, so all three factors of risk are addressed Villages are assigned a risk score in Vaijapur block of Aurangabad District By providing a risk score a season in advance of drought, planners are able to select an alternative capacity measures rather than the quickest tanker option The aim of this research is to assist district governments in Maharashtra state in predicting, between one season to two years in advance, the risk of villages to drinking water scarcity in order to respond before incurring a drinking water crisis Secondly, this model is used to prioritize infrastructure projects over the coming two years in order to best use limited financial resources to alleviate the burden of water scarcity at the village level This research could ultimately be integrated into the existing state website for statewide planning and allocation of resources Thesis Supervisor: James L Wescoat Jr Title: Aga Khan Professor of the Department of Architecture Acknowledgements This thesis is a product of months of fieldwork, and the hard work, financial support, and mentorship of many people My departmental support at Civil and Environmental Engineering, and my home department Technology and Policy Program were incredibly supportive of my academic goals and thesis research Barbara DeLaBarre and Dr Kenneth Oye were particularly helpful in their advice on framing the problem, and incorporating the methods of policy and engineering into a single, cohesive thesis I would like to thank my advisor, Dr James Wescoat, for introducing me to field research and proper methods for conducting academic research with integrity You have provided guidance that has allowed me to understand the depth of analysis required to understand a problem before attempting a solution Thank you for supporting my interests in drought research, in incorporating environmental engineering, and in developing a proper framework This work could also not have been possible without our community partners in India, including Murthy Jonnalagadda, consultant to the World Bank in Mumbai Our partners in Aurangabad, including the Zilla Parishad and Groundwater Surveys Development Agency, were also incredibly helpful in providing data, coordinating meetings and village visits, and providing expert guidance on the water scarcity dynamics in the region The MIT Tata Center, supported by the Tata Trust, provided financial and academic support without which this project could not be possible I would also like to thank Michael Bono and Chintan Vaishnav for their advice on designing metrics for risk and the different forms of sensors available to measure water levels I would like to thank Riddhi Shah for her exceptional GIS mapping skills and her work on this project, including making a trip out to Aurangabad for surveying the Zilla Parishad in Marathi I would also like to thank Dennis McLaughlin for his guidance as my engineering thesis advisor, and his mentorship for developing the system identification and PCA models Finally, I would like to thank my family and Jeremy Elster for their support in my research, my travel, and my graduate education Their compassion and support allowed me to dive deeply into my research, and to commit to developing myself as a hydrologist and policy analyst Contents Abstract Acknowledgements Chapter 1: Introduction 1.1 Problem Statement 1.1.1 Problems Being Addressed 1.2 Defining Water Scarcity 1.2.1 Government Criteria for Drought and Water Scarcity 1.2.2 Broader Criteria for Water Scarcity 11 1.2.3 Intersection of Drought and Water Scarcity 11 1.3 Literature review of Methods for Managing Water Scarcity in India 12 1.3.1 Literature Review Abstract 13 1.3.2 Historical Water Scarcity 13 1.3.3 Water Scarcity Frameworks 14 1.3.4 Water Scarcity Indices 16 1.3.5 Impacts on the Rural 16 1.3.6 Modeling Scarcity in India 17 1.3.7 The Modem Field of Planning: Drought and Scarcity 18 1.3.8 Literature Review Conclusion 19 1.4 Research Questions and Objectives 20 1.4.1 Gaps in Current Water Scarcity Planning and Management 20 1.4.2 Connection of Data Sources to Planning Process 21 1.4.3 Expanding the Range of Choice 21 1.4.4 Summary of Research 21 Chapter 2: Aurangabad District Case Study 22 2.1 Historical Water Context and Landscape 26 2.2 Existing Planning Practices 28 2.3 Climatological Conditions 31 2.4 Hydrologic and Geologic Conditions 31 2.5 Socio-Economic Conditions 34 2.6 Policy and Regulations for Water Scarcity 34 2.7 Case Study Synthesis 35 Chapter 3: Risk Model Methodology 36 3.1 Conceptual Framework 36 3.2 Hazard Score Development 37 3.2.1 Variables and Sources 37 3.2.2 Rainfall Statistics 38 3.2.3 Groundwater Statistics 41 3.2.4 Irrigation Demand and Temperature 49 3.2.5 Methodology 50 3.2.5.1 Systems Identification 50 3.2.5.2 Results and Interpretation 51 3.3 Vulnerability Score Development 54 3.3.1 Variables and Sources 55 3.3.2 Methodology 58 3.3.2.1 Data Cleaning 58 3.3.2.2 Variable Exploration 58 3.3.3 Vulnerability Score Development from Percentage Variables 60 3.3.4 Vulnerability Results 61 3.4 Capacity Score Development 62 3.5 Overall Risk Scores 65 Chapter 4: Planning Implications and Conclusions 69 4.1 Current Planning Process 69 4.2 Key Findings and Implications 73 4.2.1 Recharge Rate 73 4.2.2 Timing of Planned Government Interventions 74 4.2.3 Spatial Patterns and Planning 75 4.3 Model Recommendations 76 4.4 Conclusion 77 79 Appendices Results in R for Regression Model: Social Vulnerability 79 Images of 19 Variables in Social Vulnerability Index, Created by Riddhi Shah 80 System Identification Matlab Code 90 Principal Component Description 93 Figure 4.1 Principle Component Results: Cumulative Variance Explained 93 Figure 4.2 The Scaling applied to each variable in PCA 94 Figure 4.3 Mapping of Cumulative Variance Explained by first 10 PCs 95 Figure 4.4 Mapping of Principal Component and Principal Component 95 Figure 4.5 Eigenvalues 96 Figure 4.6 Distribution of Ten Principal Components and Summation 98 Figure 4.7 Summed Principal Components and Score 99 Figure 4.8 Map of Vulnerability Score: Principal Component Based 100 Figure 4.9 Percentage Variables in Second PCA Analysis 101 Figure 4.10 Variance Reduction in Second Principal Component Analysis 102 Figure 4.11 Map of PCA for Percentage Variables: 10 Principal Components Score 103 Figure 4.12 Map of PCA for Percentage Variables: Principal Components Score 103 Principle Component Code in Rstudio 104 Reference Table of Risk Score and Components for 16 Observation Well Villages 106 Format for Water Security Plan Household Survey, provided by GSDA Aurangabad 107 Citations 108 Chapter 1: Introduction 1.1 Problem Statement Severe and sustained water scarcity, predominantly in the form of depleted rainfall, has limited the availability of groundwater resources, and thus drinking water, in Central Maharashtra Aurangabad district, located in central Maharashtra, has a complex array of challenges in managing water scarcity Aurangabad has a growing population, water-intensive industries including soda and beer manufacturing, small farmers who rely on rainfall, and the district has lower rates of rainfall absorption to groundwater due to elevation changes and runoff It is expected that regions with below average rainfall will have declining groundwater levels A newer challenge for districts is managing areas that are receiving the expected amount of rainfall but at a higher intensity for a shorter period of time, meaning there is higher runoff and less water is absorbed into the ground Additionally, there are variable rates of withdrawal which lead to variation in regional groundwater depletion Below, Figure 1.1.1 shows pockets of groundwater depletion throughout the district of Aurangabad, in part because below average rainfall, (i.e > 10% less than average rainfall) is experienced in the south Figure 1.1.1 Groundwater Depletion in 2015 Compared to Last Year Average SILLOD KANNAD FULAMBRI I Ii~ LI I I DABAD r jiABAD WIPR No Groundwater 0.1- m Groundwater 1-2 n Groundwater 2-3 n Groundwater m" PATHAT \i d Source: GroundwaterSurveys and Development Agency, Aurangabad The population of Aurangabad district, as of 2011 census, was 3,695,928 Of that population, 62.47% or 2,308,846 people live in rural villages (GOI, 2011) The high proportion of rural communities makes water management extremely decentralized and challenging Aurangabad is positioned in the arid Marathwada region of central India, and its district, along with surrounding districts, face difficult decisions in deciding which villages receive aid in times of drought, what types of aid they receive, and in anticipating rural village water needs In the current scarcity planning process, money is set aside each year to be used in one of seven responses, and villages can apply for assistance once they are receiving less than 40 liters per capita per day (LPCD) In 2016, for example, 80 tankers were sent on 2-3 trips per day for three months to villages, costing the district over $3,000,000 USD (ZP Aurangabad, 2016) This is the most-costly of the seven responses a district can make, but it requires the least amount of advanced planning or anticipation 1.1.1 Problems Being Addressed Of the vast challenges faced by a drought-prone arid rural region, there are three systemic problems which should be addressed First, drought planning is currently reactionary rather than anticipatory; second, drought responses are spatially fragmented and thus inefficiently deployed, and third, drought planning is done in the short-term In order to improve resiliency in the Maharashtra, it is crucial to address these three concerns This thesis delves into the plans for how to make district planning proactive, increase intervention efficiency by visualizing spatial patterns of risk, and design a tool for multi-year drought assessment by means of an adapted hazards risk model By improving our understanding of the risks and vulnerabilities rural villages face, the water scarcity planning process can become more proactive and less reactionary, giving districts the ability to respond with longer term solutions than the provision of tankers An integrated regression model of groundwater prospect data, census data, rainfall data, and observation well data is used to assign a hazard score to villages in specific monsoon scenarios, giving districts insight into which villages require intervention before the peak dry season This model of risk assessment will be incorporated into the planning process as a decision support tool that can provide a ranking of water scarcity risks in the presence of different conditions, such as depleted rainfall 1.2 Defining Water Scarcity Drought and water scarcity are often used interchangeably when discussing a depletion in the supply of water to households, agriculture, or industry As this study is focused on the state of Maharashtra, it is crucial to understand these terms as they exist in policy and practice in India as well as specifically in Maharashtra 1.2.1 Government Criteria for Drought and Water Scarcity The Indian Meteorological Department (IMD), a federal agency, has historically classified drought as a rainfall deficiency which deviates from a long-term average Drought has been classified as normal if it deviates 25% or less from the long-term average, moderate drought if 50% or less, and severe drought if it deviates more than 50% from the long-term average (IMD, 2016) These classifications are typically given when a month, season, or year is atypical from the historical long-term average for rainfall This understanding of drought does not consider hourly intensity of rainfall, groundwater absorption, or other forms of water scarcity such as increased consumption Below is a map of rainfall variation in Aurangabad District, using IMD data Figure 1.2.1 The 2015 Isohyetal Map of Rainfall Variation in Aurangabad ISOHYTAL MAP OF AURANGABAD o 30 60 404W0 m 00.to." $20 10 "a0 - ON b60700 704740 Source: GSDA, 2016 The Department of Irrigation of India (DOI) has defined agricultural drought as four consecutive weeks of rainfall depletion greater than 25% from the long-term average (DOI, 2016) More than half of Aurangabad district relies on agriculture as their primary income, making agricultural drought detrimental to the livelihood of the district This definition of drought again refers to rainfall depletion, even if farmland does not rely directly upon rainfall but instead upon surface water or well water The Government of India has defined drinking water scarcity as an amount of liters per capita per day received in the smallest administrative unit, the village Water Scarcity was a village variance