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
/>
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

2


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.

3


Contents
Abstract

2

Acknowledgements

3


Chapter 1: Introduction

7

1.1 Problem Statement

7

1.1.1 Problems Being Addressed

8

1.2 Defining Water Scarcity

8

1.2.1 Government Criteria for Drought and Water Scarcity

9

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
4


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
5


1. Results in R for Regression Model: Social Vulnerability

79

2. Images of 19 Variables in Social Vulnerability Index, Created by Riddhi Shah

80

3. System Identification Matlab Code


90

4. 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 1 and Principal Component 2

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: 8 Principal Components Score

103

5. Principle Component Code in Rstudio

104

6. Reference Table of Risk Score and Components for 16 Observation Well Villages


106

7. Format for Water Security Plan Household Survey, provided by GSDA Aurangabad

107

Citations

108

6


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 5 Year Average


SILLOD
KANNAD
FULAMBRI

I
Ii~
LI
I
I

DABAD
r

jiABAD
WIPR

No
Groundwater
0.1- 1 m Groundwater
1-2 n Groundwater
2-3 n Groundwater
m"

PATHAT

.\i.

d


Source: GroundwaterSurveys and Development Agency, Aurangabad
7


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.
8


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

4 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
9


receiving less than forty liters per capita per day (LPCD), which increased to fifty-five LPCD in
2017.

Figure 1.2.2 District 2015 Water Scarcity Map - January to March
MAP SHOWING SCARCITY AREA - DISTRICT AURANGABAD
Jan. to March 2015

Legend
TALUKA BOUNDARY

Villages

Status
Tanker

Fed Villages

Scarcity Area
No Scarcity

Source: GSDA, 2015
In addition to agriculture and meteorological drought, it is commonly accepted that there are
socio-economic drought, hydrological drought, and ecological drought (UNL, 2016). While these
definitions for drought are defined at the national level, the state of Maharashtra is given the
authority to write policy for managing both drought and scarcity. Currently, Maharashtra has
certain protections for groundwater when it is considered "overexploited". As defined by the
Groundwater Survey and Development Agency of Maharashtra, any watershed that is withdrawn
in a single year over 70% is considered overexploited. This label triggers certain groundwater
withdrawal restrictions within the state, and also directs GSDA's attention to villages who rely
on groundwater from overexploited aquifers.

10



1.2.2 Broader Criteria for Water Scarcity
Seasonal drinking water scarcity in India is strictly defined as persons receiving less than 55
LPCD for drinking and living for a given season. This standard was introduced in the 2012-2017
XII 5- year plan by the National Rural Drinking Water Programme (NRDWP) of India, in order
for a habitation or village to be considered "fully covered" (NRDWP, 2016). While LPCD is
measured annually, it often fluctuates during the year, giving rise to seasonal water scarcity.
Seasonal water scarcity is seen as temporary, and has historically been alleviated when the
monsoon season arrives, but Aurangabad has experienced villages with as many as 9 months of
water scarcity for up to 5 years in a row, a historic high for water scarcity magnitude and
longevity (AAP, 2015-16). For the purposes of this study, it is important to add complexity to
this definition to ensure that we understand the root causes and risk of water scarcity. Water
Scarcity will be defined as the deficiency of drinking water supplies leading to a lack of water
for normal and specific needs, leading to health risks, diminished livelihoods and socioeconomic
vulnerability (UNL, 2016). Rainfall may remain unchanged, but water scarcity may occur in
groundwater, surface water, or elsewhere as it is the supply of water that is insufficient to meet
demand.
The ultimate goal of this study is to consider not only rainfall, but other causes of water scarcity,
and factors which lead a Gram Panchayat vulnerable to water scarcity. By having a broader
understanding of the causes of water scarcity, one can foresee regional vulnerabilities in advance
of a crisis. This is particularly useful for government responses and interventions.

1.2.3 Intersection of Drought and Water Scarcity
We have discussed four of the five types of drought: (1) meteorological being the most referred
to, then (2) hydrologic, (3) agricultural drought, (4) socio-economic, and (5) ecological drought.
These types of drought are seen as forms of a diminished water supply, and all are forms of
water scarcity. Water Scarcity can also occur without the supply being diminished by
environmental factors or drought. Ecological drought, while outside the scope of this research,
addresses the impacts of drought on multiple ecosystems such forests, vegetation, and livestock
(USGS, 2016). Ecological drought is a crucial element in considering the impacts of drought on

the environment and on farmers or irrigated land.
This study looks at Water Scarcity under agricultural, socio-economic and hydrologic drought
conditions, meaning there is a diminished water supply. Consumption patterns at the individual
household are relatively low, with the highest rural households consuming around 100 LPCD
and peri-urban consuming around 135 LPCD (AAP, 2014-15). The government standard for
urban is 135, the standard for peri-urban is 70 LPCD, and the rural standard is now 55 LPCD
(GOI, 2017). For most rural households, consumption rates are much closer to 40-55 LPCD,
11


meaning policy targeting a decrease in household consumption would greatly affect livelihoods.
The experience of Water Scarcity in times of hydrological and meteorological drought, meaning
subsurface and surface water supplies are insufficient for normal household activities is life
threatening. This form of water scarcity will be identified, and Gram Panchayats' vulnerability to
this form of water scarcity will be visually displayed for district governments to make policy
decisions.
Figure 1.2.2 Groundwater Depletion Heat Map of Aurangabad District comparing 2015 levels to
average of the previous 5 year levels

-~

Khuidabad

Index
No Depstion

a3

j 0-1 m Depletion
n 1-2rn Depletion

2-3 Depletion
>3 M Depletion

Source: GSDA Aurangabad, 2015
Methods for understanding the impacts, effects, and causes of drought have evolved over time in
both academic and political fields. It is important to assess this evolving notion of drought, and
the related fields of natural hazards and risk, in order to assess the best way to incorporate data
into anticipating risk of drought for villages in India.

1.3 Literature review of Methods for Managing Water Scarcity in India
Issues pertaining to the management of water supply and demand in India have been documented
for centuries. In order to understand the historical context for water scarcity in Rural India, we
first examine the historical distinction between drought and scarcity to understand their
12


differences. This literature review also outlines the array of modeling methods for basaltic
fractured watersheds, as it defines the geomorphologic challenge and will provide insight into
methods used to understand groundwater fluctuations. Finally, the literature review will
reference the current field of water scarcity planning as the starting point for this research.

1.3.1 Literature Review Abstract
My research entails the development of a water scarcity decision support tool for the state of
Maharashtra to identify Gram Panchayats (multi-village administrative units) most vulnerable to
water scarcity. In order to develop such a tool, it is crucial to first explore the concepts of water
scarcity and drought to best understand how the notion of each term shapes the reasons for its
perpetuation. For India, a long history of farming has made water for irrigation a central focus in
the livelihood of the nation, but now even drinking water and industrial water face scarcity. This
review addresses literature on the history of drought and water scarcity in India, as well as the
history of how to measure, model, predict and remediate water scarcity. A mixture of academic

articles, government literature, books, and doctoral theses are referenced in order to develop a
robust catalog of water scarcity resources.

1.3.2 Historical Water Scarcity
The notion of water scarcity and drought has evolved over history and geographic boundaries. In
India, drought becomes well documented in the early 1 9 th century as the cause of famine, and
drought management was defined in terms of famine relief (Arnold, 1993). Famine relief came
most commonly in the forms of irrigation works, where the baseline goal was for every farmer to
receive enough water for their crops so that communities had enough food to subsist (Arnold,
1993). This form of drought management is now called "deficit irrigation"- irrigation which
provides enough water for crops to survive, but no more. This led to lower crop yields in India,
as the goal of drought management was to provide farmer subsistence (Burgher). These survival
goals for drought management were prevalent during British colonial rule, when farmland was
vast and there was opportunity for higher revenues with higher crop yields. Drought management
became drought mitigation, as the British diverted more surface water flows to irrigation to
ensure a cash flow from Indian exports such as shampoo and cotton (Peckham). The concept of
drinking water scarcity brought with it health implications in 1 9 th century India, as the British
sought to curb contagion of disease by encouraging social hygiene, which involved regular
bathing and handwashing (Peckham).
Since Indian independence, the Indian Meteorological Department (IMD) has defined drought as
when annual rainfall for a region falls below 75% of expected rainfall. IMD has historically
categorized drought into three categories, hydrological drought, meteorological drought, and
agricultural drought. Their monsoon forecasts predict rainfall deficits and declare "drought
years" in the three drought categories. In January 2016, India Meteorological Department
13


decided to replace the nomenclature of drought with "more precise" language (Vasudeva, 2016).
Instead, the word drought is now being replaced by "deficit".
The Groundwater Survey Development Agency (GSDA), operating in Pune, Maharashtra,

operates as a state-level agency and measures water scarcity based on the percentage of
groundwater withdrawn from a watershed or aquifer. GSDA deems a watershed "overexploited"
when a community extracts 70% or more of its watershed in a given year, leading to longer-term
depletion issues and dropping of the water table (GSDA, 2015). The term "overexploited" is
used to indicate extreme groundwater scarcity. While the GSDA does not govern surface waters
or canals, it does work closely with drinking water municipalities as groundwater is the dominant
source for drinking water in Rural Maharashtra.
The Government of India defines water scarcity in India for households as a function of water
received. Any household receiving less than 40 liters per capita per day (LPCD) is experiencing
water scarcity. The State Government of India has a new target of 55 LPCD by 2017 for rural
India, meaning a home is water scarce in 2017 if each person has access to less than 55 LPCD
(NRDWP, 2015).
Water Scarcity in Maharashtra is currently attributed to a lack of rainfall catchment and
groundwater availability in basins across the state, as described by engineer and Maharashtrian
water storage expert, M.M. Dighe (2003). Dighe attributes scarcity to the increasing population
and increasing demand for water in rural Maharashtra, accounting for the numerous bore wells
competing for, and depleting, the water table. Dighe's water scarcity entails a lack of sufficient
drinking water for households to live comfortably on a daily basis, and it is threatened by a lack
of dams, groundwater recharge, and overall catchment of the sporadic rainfall Maharashtra
receives. Water Scarcity has evolved to become something that is understood based off its
categorization, causing it to be fragmented into more and more types of categories from
hydrological to socioeconomic and meteorological. Similarly, the field of hydrology is
continuing to expand drought and water scarcity to a problem of not only precipitation changes,
but also human demands and climate change as culprits and social vulnerability as a side effect.
These changes in academic understanding of water scarcity shape the conceptual framework
through which one addresses scarcity.

1.3.3 Water Scarcity Frameworks
Now that we have addressed the evolving historical notions of water scarcity, we are able to
define conceptual frameworks for thinking about and categorizing water scarcity. Timing is

important in the creation of a conceptual framework for water scarcity. In India, drinking water
scarcity is deemed as receiving less than 40 LPCD (NRDWP, 2015) regardless of for how long,
although government intervention usually requires an expectation of three months of future water
scarcity.

14


To provide contrast, water scarcity in the United States is not deemed severe unless it persists

more than two years, whereas drought that spans 6 months or more in India is seen as severe.
From 1855-63, the West Coast of the United States experienced extreme prolonged drought due
to La Nina, the counterpart to El Niio, where sea temperatures drop and trade winds are
incredibly harsh in the Pacific (Cole). In this time, and even now, U.S. drought lasting more than
2 years is regarded as prolonged and intense (Cole). The El Niflo and La Nifia were seen as
cyclical, causing water scarcity and drought in the Pacific to be viewed as cyclical changes
(Cole). In India, the annual monsoon season is seen as the cyclical 'reversal' of water scarcity.
Sinha in a survey of 900 years of monsoon precipitation shows the increased variability in total
accumulated precipitation each year and variability in the intensity of rainfall in order to show
that sustained drought is likely to become a more common occurrence (Sinha 2007).
With a fixed amount of rainfall, conceptual frameworks have shifted to understanding not only
the supply, but ways to model and curb demand for industrial, household, and agricultural water
use. Malin Falkenmark is a leading expert on not only the effects of human demand on scarcity,
but the concentrated negative effects on low-income, rural and minority populations. Falkenmark
claims that water scarcity is the key strain on water security, and thus on socioeconomic
development (Falkenmark, 1997). Falkenmark, along with colleagues, addresses the concept of
demand-driven water scarcity, how it can be measured by use-to-availability, and postulates the
proper reserve amounts as a percentage of total water supply (Falkenmark and Lindh, 1976).
Demand-driven water scarcity was coined as "water stress" in 2011 to identify the human and
non-human consumption of water as a stress on the overall water system (Kummu and Varis,

2011). The United Nations formally set the mark for high water stress as 40% withdrawal in a
"Comprehensive Assessment of the Freshwater Resources of the World" (UN, 1997), but this
has been expanded by Falkenmark in developing nations to 70% withdrawal as the point of
overexploitation, where a basin should be closed until recharge has occurred (Falkenmark,
2003).
Within drinking water scarcity, it is broadly accepted that there are two forms: demand-driven
water scarcity and population/supply driven water scarcity (Lankford, 2013). Supply driven
water scarcity is the result of a lack of sufficient water, including rainfall, surface water,
groundwater, and treated oceanic water.
The University of Nebraska, Lincoln provides detailed conceptual frameworks for water scarcity
and drought through their Institute of Agriculture and Natural Resources. This institution, which
specializes in drought studies, identifies that drought is a "deficiency of rainfall... over an
extended period of time, generally at least one season", where the meaning of deficiency varies
widely by geography. Donald A. Wilhite provides a method for differentiating drought from
other water crises such as scarcity through a conceptual framework (Wilhite, 2005), while A.F.
Loon uses observation-modeling to distinguish drought from water scarcity (Loon, 2013).

15


The Government of India's Ministry of Water Resources establishes three categories of drought,
each measured differently. The Government of India concretely measures and defines
meteorological, hydrological, and agricultural drought (Ministry of Water Resources, 2013).
These definitions are used to classify a village or administrative unit in India as 'water scarce',
which in turn signals government remediation processes.
1.3.4 Water Scarcity Indices
There are water scarcity and drought indicators or indices, used to identify the relative risk or
vulnerability of a region, watershed, or community to an imbalance in water access. The Water
Poverty Index is a new, holistic look at the aggregate of many indices and is designed for
identifying the vulnerability of a community to risk by aggregating watershed, country, and

regional indices (Sullivan, 2002).
The Palmer Drought Severity Index and Crop Moisture Index were the leading indices for
drought measures in the 1950's-2000's. Both gave a relative measure of moistness based on
temperature and precipitation to estimate the amount of evaporation or evapotranspiration, and
were best suited for those reliant on irrigation or groundwater (Palmer, 1965). The Standard
Precipitation Index, a less complex index to calculate, was developed in 1993 as an alternative to
the Palmer Index to measure the standard deviations away from mean precipitation in a region
and provides early warning of drought (NASA.gov).

1.3.5 Impacts on the Rural
Drought impacts on rural communities, particularly in India, differ widely from drought impacts
on the urban. The University of Yamanashi, Japan explored the impacts drought have on
Maharashtra, a state with a large rural and farming population and a state that produces 15% of
India's gross domestic product (Ichikawa et al, 2014). This study found that the depletion of
water resources in rural Maharashtra had high impacts on agriculture and food security for the
state as a whole. They point out that the 2012 drought in India caused the nation's gross domestic
product to decrease by 0.5%. The study also shows the varying degree of water quality among
private water tankers, and the cumbersome process involved with retrieving water out of a
depleted well.
There are also large questions of livelihood and gender inequality during rural drought in India,
addressed by Krishna in an exploration of community resource management (Krishna, 2004).
Krishna addresses the dangers of women retrieving water late at night, the inability to attend
school, and how rural industries rely more heavily on water, and thus does their livelihood
(Krishna, 2004).
Falkenmark, Lunddqvist, and Wildstrand provide insight into the micro-scale approaches to the
vulnerability of drought in semi-arid regions of India in order to develop large-scale strategies
for mitigation (1989). The vulnerability to drought is assessed against the counterfactual, where
16



drought iessons over time, in a study by Dr. Gopalakrishnan in order to illustrate that poverty and
environmental degradation are likely effects of drought (1993).
International studies on rural drought in arid regions on impacts of the family, the farm, gender
equity, education, and livelihoods are found on Tanzania (Krishna, 2004), Nepal (Merz et al,
2003), Australia (Bettini et al, 2013), Brazil (Garcia-Torres et al, 2003), and Canada (Sanyal,
2015).

1.3.6 Modeling Scarcity in India
There are a variety of hydrological modeling methods relevant to water scarcity, as well as some
more general climatic, natural disaster related modeling and general vulnerability modeling
methods. Although it deals with geohazards, a relevant survey of modeling techniques includes
Pradham's "Terrigenous Mass Movements", which explores methods for modeling and mapping
vulnerability to natural disaster. These methods include risk mapping, "data modeling,
topography, geology, geomorphology, remote sensing, artificial neural networks, binomial
regression, fuzzy logic, spatial statistics and analysis, and scientific visualization" (Pradham et
al, 2012).
Remote sensing of bore wells has become a successful way to monitor water levels and water
management, though it has limitations in hardrock terrain (Rao, 2003). Remote sensing has also
been found effective for measuring evapotranspiration of crops, and is well suited for rural, arid
India, as was found in a 2001 study (Srinivas 2001). Remote sensing data have also been
successfully incorporated into GIS for mapping of water resources by the International
Astronautical Congress (Jeyaram et al, 2006).
Demand modeling of water using non-spatial modeling, such as system dynamics modeling, has
been demonstrated by the Massachusetts Institute of Technology in a 2011 study on Singapore
(Welling, 2011). Methods for multi-variable econometric regression for water demand and
prediction based on population and environmental factors are analyzed by the Institute for Water
Management in Dresden, Germany (Koegst et al, 2008).
Maharashtra faces unique challenges in modeling its groundwater due to the nature of basaltic
fractured hard rock. A notable study on the formation of basaltic aquifers in India was conducted
to address the complexities of the structures to be modeled. Measuring potentialities for

groundwater in basaltic hardrock in India is a method that allows for error and interquartile
ranges. Kriging, a geostatistical method used in the Oil and Gas industry is also used in some
groundwater models when one can assume uniformity of the soil or rock underneath the surface
(Khan et al, 2016). An alternative method for randomizing water levels across an unmapped
aquifer is Monte Carlo simulation, which requires large amounts of data for the strata types and
depths below all observation wells (Khan et al, 2016). The most commonly used method today
for modeling groundwater in hard rock is still the pumping test, which is an empirical method
17


that requires pumping all of the water out of an irrigation well and measuring recovery rates
(Shah, 2012).
Service delivery models for drinking water in India are analyzed in a study by the Naandi
Foundation (Kumar et al, 2014). Modeling monsoon rainfall variability using national Indian
data has been completed in a 2014 study (Ranade et al, 2014). Drought characterization,
modeling future predictions of drought, and modeling change and down-scaling are all addressed
in a 2015 study of international drought modeling and mitigation (Senaut, 2015).

1.3.7 The Modem Field of Planning: Drought and Scarcity
In March 2016, U.S. President Obama released a Federal Action Plan for Long Term Drought
Resilience, which included a memorandum, action plans, progress reports, and a tracking of
actions taken'. The US Bureau of Reclamation is offering WaterSMART grants for funding
small, on the ground projects as well as large scale energy efficient and water conservation
projects and planning improvements for drought 2 . This shift towards increased research in the
drought planning field has been gradual, as can be seen in Figure 1.3.7.1, showing the number of
English-written books referencing Drought Planning, peaking in 1990 and again after 2008.
Figure 1.3.7.1 Drought Planning references in English books
0.000000500%
0.000000450%


0.000000400%
0.000000350%

0.000000300%
0.000000250%
0.000000200%

0.000000150%
0.000000100%
Druh
Drought

0.000000050%.
0.000000000%
1900

1910

1920

1930

1940

1960

(cM n hNi1sifar fOcus.

1960


1970

1990

1990

2000

9&ftPWaodfoaactcw=8%*N*)

Source: Google Ngram Viewer, 2017
This trend helps to narrow the scope of drought planning to its relative inception in the 1970's,
its peak in the 1990's, and the current field as it stood in the 2008 time frame.
We are seeing the most data-intensive forms of drought planning research in academic
dissertations and theses, with modeling, prediction, and decision support systems designed, such

I />2 />18


as a recent Texas A&M dissertation that integrates risk, two new multivariate indices, and a
decision support tool (Deepthi, 2014).
The University of Nebraska, Lincoln has a leading center on drought planning (UNL, 2016).
The current framework for rural drought planning in the United States consists of a balance sheet
where supply and demand are calculated, with water quality compromised supplies are
subtracted from the total supply. A target is set for the amount of water in reserves, and when
supply falls below a certain level actions are triggered.
The current framework for water scarcity in India consists of a bottom-up "signal" for scarcity,
and then a top-down response. Water scarcity is experienced at the village level, paperwork is
submitted to a Block Development Officer (BDO), and then either the BDO or District
determines the type of response, if any, to be provided to the village. This process is timely and

does not provide Districts any insight to plan for future water scarcity. The process is designed
for rural villages who receive less than 40 LPCD, for which it receives applicants during three
quarters of the year. Applications are not received during monsoon season. The government
enlists one of seven approved responses to remediate the drinking water, ranging from short to
long-term solutions and vary in cost. The policy responses are listed in the table below.
Figure 1.3.27.2 Policy Responses to Water Scarcity Applications in Maharashtra
1.
2.
3.
4.
5.
6.
7.

Provide Shallow Trenches in Riverbed
Deepening and Desilting of Wells
Acquisition of Private Wells
Providing Water by Tanker/Bullock Cart
Special Repairs of Piped Water Supply Schemes
Provide a Tubewell
Provide Temporary Piped Water Supply

IIT Bombay, IIT Roorkee and a variety of other institutions in Maharashtra, Gujarat, and
Uttarakhand India have produced sophisticated research on predictive techniques for water
management, aquifer mapping, applications of sensor technology to water supply estimation of
surface and groundwater, and techniques for curbing agricultural and household consumption of
water in villages. There is a lack of integration of socio-economic vulnerability with
hydrogeological vulnerability.

1.3.8 Literature Review Conclusion

This literature review provides historical context for water scarcity research in India, as well as
the general shift in the field of water management to conceptualize, model, and respond to water
scarcity. While the concept of water scarcity is not new to rural Maharashtra, a region heavily
19


reliant on irrigation, we are entering a new era of sustained drought that current literature does
not adequately address. The policy and practical responses to higher severity drought and longerterm water scarcity require a multi-year planning process and a new suite of planned responses.
In Maharashtra, the water scheme investment plans, called "Annual Action Plans" occur on an
annual basis, and serve as systematic planning processes for water scheme construction and
alteration in rural Maharashtra. These plans are not compared year to year, and long-term
solutions are often unable to be reached due to the need for quick solutions to drinking water
scarcity. While water tanker use is widely discouraged as a long-term fix, government funding
limits the response to scarcity-prone villages to one action at a time. This means, provide a
tanker for instant relief and assume the monsoon season will relieve this need, or construct or
mend sources for water withdrawal.
Water scarcity, in the form of depleted rainfall, or scarcity due to water stress is increasingly
hard to predict as rural Maharashtra is covered in publically dug wells, as well as non-sanctioned
privately dug wells for human consumption and irrigation. The lack of proper community
management at the aquifer and sub-basin levels often leads to a scurry in newly prolonged dry
seasons for groundwater.
This research on water scarcity planning in Aurangabad will augment existing practices for
management of a scarce resource for drinking water, while introducing new concepts regarding
the modeling of vulnerability to water scarcity, and the suite of policy responses available to
each block (Taluka). There is a striking parallel between natural disaster modeling and drought
modeling, making it a natural connection to design a drought model with the existing natural
disaster modeling methods already established.
1.4 Research Questions and Objectives
The primary research question is how can the district planning process for water scarcity be
improved with a numeric model to identify drought risk? This research question is further broken

down into two lines of thinking: how can the data being collected by the district and state
agencies be used to anticipate future drought risks, and how can data models be integrated into
the current planning process.
The objectives of this research are to make the planning process more proactive, to understand
how a variety of factors, including rainfall, affect groundwater levels, and to provide a decision
support tool that can be used for scenario-based modeling of Gram Panchayats groundwater
levels.
1.4.1 Gaps in Current Water Scarcity Planning and Management
As is referenced in the Literature Review, there is a lack of sophisticated data integration in
current water scarcity frameworks as it can be cumbersome and up-front costs are high. Much
emphasis is put on predicting rainfall, but not on how that variability disproportionately affects
20


Gram Panchayats. There is also a lack of conjunctive groundwater and surface water
management, as in India different state agencies control the two sources of water, politicizing the
possibility of conjunctive management.
1.4.2 Connection of Data Sources to Planning Process
The District Governments of Pune and Aurangabad, as well as the Groundwater Survey and
Development Agency of Maharashtra have expressed an interest in using their existing data
sources to the planning process. Our partnership began with a statistical analysis of Annual
Action Plan data from 2014-2015 to notice trends in the types of projects, forms of scarcity, and
characteristics of villages being selected for water scarcity remediation measures. These data,
along with Groundwater Prospect Map data, a century of rainfall data, Observation Well data,
the Integrated Management Information System (IMIS) repository, and Government of India
Census data from 2011 have been cleaned, combined, and analyzed for trends, relationships and
statistical significance in advance of the model creation. This process was done as an initial step
to understand how projects have been selected, and the types, timing and recurrence of water
scarcity in Aurangabad District as well as Pune District. Initial reports were presented in person
in Maharashtra in January 2015 and August 2016 to ensure applicability of research and

feasibility of integration with existing planning processes.

1.4.3 Expanding the Range of Choice
Aurangabad district is aware of the sustained drought and the changing nature of monsoon
seasons. Their fractured basaltic hardrock makes it difficult for geologists and hydrologists to
predict exactly where water sits and where rainfall recharges the land without aquifer maps,
which are being mandated by the federal government but have not been completed in
Aurangabad. Many regions of Aurangabad experience water scarcity at some point in the span of
a calendar year, but the timing, cause and severity have become seemingly unpredictable.
By developing a model to help assess the risk of groundwater scarcity in Gram Panchayats,
districts are able to react to the problem earlier, and possibly differently than if they had less time
to react. This expands the range of choices and policy responses a district can employ. The late
scientist, hydrologist and professor Gilbert F.White studied how in disaster preparedness and
response planning, when a government can consider the full range of choices they are less bias to
pick one over another (White, 1986). Ultimately, this would improve the management process of
water scarcity planning in India by limiting a bias towards the quickest solution.
1.4.4 Summary of Research
Chapters two through four will outline the conditions of our study's location , how the risk score
was calculated, and how the results of this research could be introduced to the existing annual
21


District Water Security Plan. Methods tried but not ultimately used for the risk score are located
in the appendix.

Chapter 2: Aurangabad District Case Study
The federal government, referred to as the Government of India (GOI), mandates general rules
such as national limitations on well-depths (60 meters) and minimums for drinking water (55
LPCD by 2017), but it recognizes that states develop regulations for consumption and
distribution of groundwater and surface water (GOI, 2011). Indian States govern most aspects of

water, including drinking water and irrigation water. This allows for tailored policies to distinct
geographies, but also creates challenges regarding coordination of interstate rivers, canals and
watersheds. The state of Maharashtra, shown in Figure 2.0.1, spans the coastline, desert, and
mountains adding many layers of complexity to state-level water management.
Figure 2.0.1 Maharashtra State Highlighted in India Map

Source: GSDA Mumbai, 2017
The state of Maharashtra is the focal point of this research project. The following is a brief
explanation of the social and environmental landscape of the state. Maharashtra is a state with
extremely high literacy rates relative to the country, and home to two large cities; Mumbai and
Pune. In terms of area, the state is predominantly agrarian, and is comprised of 70% basaltic
hardrock and associated black cotton soils. The state has 114,200,000 people according to the
22


2011 national census, and its cities and peri-urban regions are expected to increase in population
(Hui, 2017).

Figure 2.0.2 Divisions of Maharashtra State
Bhandara
liandurandl
jagaon

db

Choaur

Now*
Suburban


mumrbal

hange

city

a

Amaravati Division
Aurangabad Division
Konkan Division
Nagpur Division
Nashik Division

ur

SPune
Division

Source: Google Images, 2017
Figure 2.0.2 displays the six divisions of Maharashtra state, which each contain four to eight
districts. Maharashtra on its west side is home to the Konkan Division, known for high rainfall,
lush terrain, and high salinity coastal soil due to its proximity to the ocean. On one field visit to a
village in Raigad District, I sampled a glass of water collected from a rice field with a TDS of
over 3000. This brackish water limits groundwater utilization, forcing rural communities to rely
heavily on the vast rainfall during monsoon seasons to carry them through the year. Their soil is
best for rice and coconut crops. While water scarcity is not an issue, storage capacity of rainfall
and water quality are threats to their drinking water supply.
The central part of Maharashtra state, the Aurangabad Division, is home to the Godavari basin,
named after the Godavari river. This area is a part of the Marathwada region, an arid to semi-arid

region with low rainfall in monsoon season and limited surface water. Eight districts within
Maharashtra state rest within the Marathwada region, one of which is Aurangabad district. The
entire Marathwada region is known for its proclivity to drought. Aurangabad is the regional
headquarters and will be the case study for this research.

23


The varying geography and climate of Maharashtra state, as is seen in the contrast between the
Konkan and Aurangabad divisions, makes water planning at the state level challenging. This has
led the State of Maharashtra to empower its district governments to shape planning practices and
standards that encourage financial and water sustainability under the conditions unique to their
regions. This study intends to enhance the existing planning practices at the district level by
identifying regional risk to drought and regional socio-economic vulnerabilities.

Figure 2.0.3 Aurangabad District
Soegaqn

0

Phulambri

Paithan

Source: GSDA Aurangabad, 2016
Aurangabad district, as seen in Figure 2.0.3, is comprised of nine blocks. The district is the
headquarters for the Aurangabad Division of Maharashtra State, governing the arid and semi-arid
Marathwada region. Aurangabad District has a history of water innovation, with some structures
still in operation from the 1600s.


24