MODELING VULNERABILITY AND RESILIENCE TO CLIMATE
CHANGE: A CASE STUDY OF INDIA AND INDIAN STATES
ANTOINETTE L. BRENKERT and ELIZABETH L. MALONE
Joint Global Change Research Institute, College Park, 8400 Baltimore Avenue, Suite 201,
College Park, MD 20740, U.S.A.
E-mail:
[email protected]
Abstract. The vulnerability of India and Indian states to climate change was assessed using the
Vulnerability-Resilience Indicator Prototype (VRIP). The model was adapted from the global/country
version to account for Indian dietary practices and data availability with regard to freshwater resources.
Results (scaled to world values) show nine Indian states to be moderately resilient to climate change,
principally because of low sulfur emissions and a relatively large percentage of unmanaged land. Six
states are more vulnerable than India as a whole, attributable largely to sensitivity to sea storm surges.
Analyses of results at the state level (Orissa, and comparisons between Maharashtra and Kerala, and
Andhra Pradesh and Himachal Pradesh) demonstrate the value of VRIP analyses used in conjunction
with other socio-economic information to address initial questions about the sources of vulnerability
in particular places. The modeling framework allows analysts and stakeholders to systematically
evaluate individual and sets of indicators and to indicate where the likely vulnerabilities are in the
area being assessed.
1. Introduction
India is one of the most important countries in the world with regard to climate
change sources and impacts. With a large and growing population, India’s emissions of greenhouse gases are increasing. At the same time, potential climate impacts in India are severe; they include sea level rise, changes in the monsoon,
increased severe storms and flooding, and more drought. Recently, climate variability in the form of floods and cyclones has resulted in destruction of crops,
property and infrastructure, as well as in negative impacts on human health and
well-being. All of these impacts set back general socio-economic development.
And continuing dependence upon agriculture for food and livelihood (25% of
GDP and 60% of the labor force, 2002 and 1999 estimates) (CIA, 2003) makes
the Indian people particularly vulnerable to climate variability and change. Most
studies project decreased yields in non-irrigated wheat and in rice, and a loss
in farm-level net revenue between 9 and 25% for a temperature increase of 2–
3.5 ◦ C (Tata, 2002). The costs of a 1-m sea level rise could include up to 7
million people displaced, and 5764 sq km of land and 4200 miles of road lost
(Tata, 2002).
Since the Indian economy is intrinsically linked with the annual monsoon cycle through its dependence on agriculture, a better understanding of the future
Climatic Change (2005) 72: 57–102
DOI: 10.1007/s10584-005-5930-3
c Springer 2005
58
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
behavior of the monsoon and its variability is warranted for disaster mitigation and
for developing adaptation strategies to cope with climate variability and climate
change (Lal, 2001). Rainfall in India is highly variable both in time and space, and
shows significant interannual variation. Linear trends of monsoon rainfall during
1871–1998 at each of over 300 observing stations spread over India show various
statistically significant trends in some broad contiguous areas.
The two monsoon seasons (the southwest monsoon during June–September
and the northeast monsoon during November–December) bring forth rains—many
times in intensities and amounts sufficient to cause serious floods, creating hazardous (and often disastrous) situations. Cyclonic storms in the pre-monsoon
months (April–May) and the post-monsoon months (October–November) may
cause large-scale inundation and destruction. The eastern coast of India along Bengal, Orissa and Andhra Pradesh is especially prone to such tropical cyclones.
An analysis of seasonal and annual surface air temperatures for India, using data
for 1881–2001 for 25 or more stations, shows a significant annual mean warming of
0.68 ◦ C per hundred years. The warming is mainly contributed by the post-monsoon
and winter seasons. The monsoon temperatures do not show a significant trend in
most parts of the country except for a significant negative trend over Northwest
India. Maximum daytime temperatures show more of a trend than minimum nighttime temperatures (Lal, 2001), in contrast to general expectation. Almost 67% of
the glaciers in the Himalayan mountain ranges have retreated in the past decade
(Lal, 2001).
Intense deforestation activities have taken place along the foothills of Himalayas
and in the Assam region, and land use patterns have undergone definite changes
over parts of Rajasthan and Punjab (Northwest India). Surface cooling with significant increase in rainfall has been observed in the peripheral regions of the
Rajasthan desert; increased area under irrigation may be one of the main causal
factors (Lal, 2001).
Observed trends in the mean sea level along the Indian coast indicate a rising
trend of about 1 cm per decade, which is close to that recorded in other parts of
the globe. Today, coastal regions in India and Bangladesh are subjected to stronger
wind and flood damage than in the past because of storm surges associated with
more intense tropical storms (Lal, 2001).
The diversity and extremes of India’s climate and geography are characteristic
of its society as well. Religious and cultural diversity is a major feature of Indian
life. The strong Hindu traditions have been synthesized with and challenged by
other religions, notably Islam, Christianity, and Sikhism. There are at least 300
known languages in India, 24 of which have at least one million speakers each.
There are differences, sometimes amounting to estrangement, between the North,
with its history of grand-scale invasions, and the relatively stable South. Religious divisions became geographical divides when Muslim Pakistan (1947), then
Bangladesh (1971), were created, but ethnic strife among groups persists. However,
the extended family is a core feature of Indian life. Together with a sense of civil
CASE STUDY OF INDIA AND INDIAN STATES
59
Figure 1. Map of income estimates by state (http://www.mapsofindia.com/maps/india).
society’s claims on individuals and families, the extended family knits the society
together and emphasizes interdependence (Gannon et al., 1994).
Since 1990, India as a country has moved aggressively from a centrally
planned economy to private ownership of businesses and trade liberalization.
It has “developed a diversified industrial base and sophisticated financial sector. Its software subsector—one of the most dynamic in the world—has experienced a sustained and rapid growth” (World Bank, 2002). India has made substantial strides in reducing infant mortality, increasing life expectancy, and improving literacy. Yet poverty (see Figure 1) and malnutrition (World Bank, 2002)
continue to plague India, as well as serious environmental issues. In addition,
the ongoing dispute with Pakistan over Kashmir and ethnic strife claim national
attention.
India’s broad spectrum of highly articulated national policies includes goals
in the areas of economic development, human development, and environmental
protection. India has put in place its tenth Five-Year Plan, and each state prepares an
annual plan that is commented on by the National Planning Commission. National
60
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
goals are, of course, differentially implemented in each of India’s 28 states, which
exhibit widely varying degrees of social and economic development.
At the national level, India’s climate change policies are subsumed in its
economic-industrial and human development policies, which come first. Local policies have had some success in limiting a significant urban air pollution problem.
Substantial improvements in local air quality in Delhi, for example, have been
caused by recent government programs to improve the quality of petrol and diesel
fuels, introduction of emissions standards for vehicles, and conversion of buses
and three wheelers to compressed natural gas fuel.1 Generally speaking, climate
change policy has been reactive rather than proactive and focused largely on the
energy sector. In India, coal accounts for 55% of primary, commercial energy,
and energy demand is growing quickly; burning coal is also a major source of
greenhouse gas emissions and air pollution. For the first time, in the eighth 5-Year
Plan (1992–1997), funds were provided for energy efficiency measures. According to reports from the Organization for Economic Cooperation and Development
(OECD) and the Pew Center (Parikh and Parikh, 2002, p. 25; Chandler et al.,
2002); India, through normal policy developments, is “making significant progress
in limiting greenhouse emissions” through energy efficiency improvements and
environmentally friendly energy development. More aggressive measures, India
feels, should be financed by developed nations as they lead by reducing their own
emissions.
In the area of disaster mitigation, much has been done to document conditions
leading to vulnerability. For example, there are a Flood Atlas of India and a Disaster
and Vulnerability Atlas of India;2 the latter assesses the vulnerability of housing
and infrastructure to earthquakes, cyclones and floods to improve zoning and construction.
Internationally, India has played a key role in climate negotiations at several
points. India broke the impasse at the first Conference of the Parties by leading the
development of a common statement that became the basis for the Berlin Mandate.
More recently, India hosted the eighth Conference of the Parties to the Framework
Convention on Climate Change in 2002. India, bolstered by nongovernmental organizations such as the Tata Energy Research Institute (TERI) and the Centre for
Science and Environment (CSE), focuses on per capita emissions (low in India and
high in most developed countries) and on cumulative emissions (also low in developing countries and high in developed countries), as the indicators that developed
countries should undertake mitigation first. Government officials press developed
nations to establish and conform to emissions reduction goals and to engage in technology transfer to developing countries. Researchers at CSE authored the widely
cited report, Global Warming in an Unequal World (Agarwal and Narain, 1991).
CSE rebutted the second World Resources Report (Agarwal and Narain, 1992),
making the distinction between “subsistence emissions” of the poor (mostly in developing countries) and the “luxury emissions” of the rich (mostly in developed
countries). CSE has also characterized “green” policies dictated by the North (e.g.
CASE STUDY OF INDIA AND INDIAN STATES
61
debt-for-nature swaps) as unwarranted interference in other nations, as exacerbating
inequality among nations, and as likely to foster unsustainable management. India
has criticized the control of Global Environmental Fund by developed countries,
and has not submitted its national communication under the Framework Convention
on Climate Change.
In this context, assessing the vulnerability to climate change in India is an important component of formulating adaptation strategies both nationally and internationally. A vulnerability assessment can point to areas and sectors where vulnerability
is high and thus adaptation strategies should be developed. A structured approach
to vulnerability assessment should provide comparability among regions, so that
common and integrated adaptation policies can be identified.
2. Assessing Vulnerability
Vulnerability to climate change, as a research concept, is both overdetermined and
underspecified. The literature is replete with alternative definitions of vulnerability
(Downing et al., 2003; McCarthy et al., 2001; Kelly and Adger, 2000; Comfort et
al., 1999; Vogel, 1997; Cutter, 1996; Ribot, 1996; Bohle et al., 1994; Cannon, 1994;
Watts and Bohle, 1993; Dow, 1992; Downing, 1991, 1992; Chambers, 1989).
Cutter (1996) identifies three distinct clusters of definitions for vulnerability:
as risk of exposure to hazards, as a capability for social response (what we call
coping or adaptive capacity), and as an attribute of places (e.g. vulnerability of
coastlines to sea level rise). Cutter (1996, p. 532) proposes a “hazards of place”
model that bridges various definitions and states: “Vulnerability is the likelihood
that an individual or group will be exposed to and adversely affected by a hazard. It
is the interaction of the hazards of place (risk and mitigation) with the social profile
of communities.” She ultimately argues (1996, p. 536) that “it is place that forms
the fundamental unit of analysis” for vulnerability.
The multiple meanings of vulnerability are attributable, at least in part, to its
relative newness as an area within climate change research. Research on vulnerability to climate change extends impacts research. Impacts are typically assessed on natural systems and managed resource systems, such as agriculture.
Vulnerability-resilience assessment focuses on societal systems and individual humans. This difference in emphasis parallels the difference between research into
climate change, which has typically focused on physico-chemical systems, and sustainability research, which has emphasized social systems (see Cohen et al., 1998;
Robinson and Herbert, 2000).
Studies have often conflated impacts and vulnerability, assuming that changes
in the environment are self-evident problems for human systems and for humans
themselves. For example, changes in the timing or amount of precipitation may be
assumed to cause damage to agriculture and, thus, to food security and the livelihood of farmers. In part, such a conflation reflects assumptions that a static world is
62
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
better than a changing world and that any change will carry costs. Another common
assumption is that negative impacts of climate change will be so great that projections of these impacts would drive mitigation policies. Although many instances of
large-scale, high-cost environmental change can be cited, the general assumptions
could be countered by examples of favorable changes. Both the physical and the
social worlds change—and have changed—for the better.
In vulnerability research, analysts are committed to some level of relatedness
between information about the environment and information about society. There
are several possibilities:
• Vulnerability assessment may be virtually indistinguishable from impacts assessment (e.g. Smith et al., 2001).
• Vulnerability assessment was initially an extension of (climate) impact assessment; second-generation vulnerability assessment includes evaluation of
impacts on goods and services (F¨ussel and Klein, 2003).
• Vulnerability assessment may be virtually indistinguishable from sustainability assessment.
• Case studies may be the method for integrating mostly data descriptions of
specific places and their geographical characteristics, considered together with
human systems such as agriculture (e.g. Ribot et al., 1996).
• Indicators and indexes may be used to integrate quantitative information about
both the environment and the society (e.g. Downing, 1991; Moss et al., 2001).
This study is an example of moving from descriptive impacts of climate and
climate change to quantitative indicator sets that allow comparisons of regional,
country, state or provinces, or smaller localities in terms of their vulnerability
and resilience to current and changing climate. It employs a set of indicators of
both environmental and human factors, aggregated into sectors, which are further
classified into sensitivity or coping/adaptive capacity. The method is explained
more fully in the following section.
3. Methodology for Vulnerability Assessment
Vulnerability-resilience analysis (which is associated with adaptation to climate
change) does not merely extend impacts research but also changes the focus from
relatively more physical and quantifiable impacts to meaningful consequences
for human societies. Research on vulnerability is part of the shift from physical/economic representations of climate changes, emissions-producing activities,
and effects of crop yields and water availability to (usually more qualitative)
representations of human attributes and institutions such as land tenure systems,
modes of production, and governance. The tools of impacts research were predominantly models; the tools for vulnerability research have been predominantly case
studies.
CASE STUDY OF INDIA AND INDIAN STATES
63
Both qualitative and quantitative studies contribute to the analysis of
vulnerability. A major focus of the Intergovernmental Panel on Climate Change
(IPCC) is on analyzing candidate methods to summarize and compare socioeconomic and environmental conditions that contribute to vulnerability. However, no
generally agreed-upon methodology or framework has appeared. Criteria for a
framework include the following:
• capturing the important aspects of vulnerability (the subject of many debates
within the research community);
• allowing for comparison among countries, regions, and smaller geopolitical
areas;
• using quantified indicators where possible to facilitate comparisons and integrated analysis;
• utilizing a transparent methodology so that users of the framework understand
the results and are able to analyze them with qualitative information.
One area of research concerns the development of indices of vulnerability, based
on sets of indicators of vulnerability’s various aspects. Summary numbers must be
used with caution, of course, since it is difficult to boil down complex culturalsocial-economic-resource-political structures into one number or even a set of
numbers. The IPCC Third Assessment Report reviews the efforts made to perform vulnerability assessments using indicators (Ahmad et al., 2001). This paper
is an addition to that research.
Previously, we developed a Vulnerability-Resilience Indicator Prototype (VRIP)
model to compare national vulnerability-resilience indices against a global index
(Moss et al., 2001). A country’s or region’s vulnerability to climate change is
assumed to be a function of three factors (see Figure 2):
• Exposure–the nature and extent of changes that a place’s climate is subjected
to with regard to variables such as temperature, precipitation, extreme weather
events, sea level; exposure is location-dependent.
Figure 2. A climate change and variability impacts framework: The Vulnerability-Resilience Indicator
Prototype (VRIP).
64
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
• Sensitivity–how systems could be negatively affected by the change in climate,
e.g., how much land could be inundated by sea level rise, how much might
crop yields change, or how much might human health be affected.
• Adaptive capacity–how much capability a society has to adapt to the changes
so as to maintain, minimize loss of, or maximize gain in welfare.
To assess vulnerability we look at the response to climate exposure, expressed
as sensitivities to climate, and societal coping and adaptive capabilities.
Evaluation of a society’s coping and adaptation capacity is based on society’s
human resources, economic capacity and natural capital. Sensitivity of a society to
climate variability and change is based on an evaluation of its food and water security, its settlement security, aspects of the health of people, and natural resources.
We used six criteria to design and evaluate a methodology for vulnerability
assessment:
• Relevant aspects of potential vulnerability of both physical and social systems
should be included, but emphasis should ultimately be placed on social systems
and their expected changes over time with and without considering climate
change.
• Proxies ought to represent in quantitative form the qualitative concepts of the
relevant potential sensitivities to change and the capacities to respond.
• A hierarchical aggregation of proxies may be formulated as well as the explicit
weighing of proxies, based on well-described assumptions.
• Quantitative vulnerability indicators should provide assessments of vulnerability to a first approximation. Results of such an assessment ought to be
able to be used to identify important players in the form of leading proxies that may help direct further research and analysis toward sectors where
resilience-building and adaptive strategies are relative priorities.
• The quantitative model should be transparent in its sources, processes, and
results in order to be useful to policymakers.
• The methodology should be usable at multiple spatial and time scales.
The VRIP model calculates indicators of sensitivity to climate change and of
coping and adaptive capacity. It aggregates these into an overall indicator of vulnerability in a three-level, transparent process. Indicators are grouped into sectors;
for example, the food security sector comprises two indicators: cereal production/agricultural land and animal protein demand per capita. The sectors are classified as either sensitivity (food security, water resources, settlement/infrastructure,
human health, and ecosystems) or as capacity for coping and adaptation (environmental capacity, economic capacity, and human and civic resources). The (negative)
sensitivities and the (positive) coping/adaptive capacities are aggregated to a single
vulnerability index.
The primary building blocks for the model are the indicators. Indicators, which
are observable and measurable, are proxies for aspects of vulnerability and adaptive
CASE STUDY OF INDIA AND INDIAN STATES
65
capacity, which are not directly measurable or observable. Each indicator should
simplify or summarize a number of important properties and be quantifiable, preferably based on existing data.
The selection of individual indicators, what they represent and the behavior as
a set are important. Moreover, the interrelatedness of all the indicators must be
considered to identify potential gaps or overlaps. One possibility, for example, is
that many indicators based on per capita data would overemphasize population as
a factor in vulnerability.
Because of interrelatedness, the structure of the VRIP may implicitly weight
indicators, as described. A separate issue is the possibility of ascribing an initial
weight to indicators. In this study, all indicators are given equal initial weights,
since there are no theoretical reasons to ascribe different weights. Other studies
may give different initial weights to individual indicators, based on the goals of
the study, scenario analysis, or particular circumstances of the area or sector under
study.
Specific indicators used in the VRIP are given in Table I, grouped into sectors
and classified as either coping/adaptive capacity or sensitivity.
We describe, below, our approach to proxy evaluation based on our previous
research (Moss et al., 2001) and present work on India’s states. Where relevant, we
describe where and how our comparison of India’s states led to different approaches.
The national data and the data of India’s states were scaled (indexed) against their
world counterparts. National data used were in general obtained from the World
Bank (1998) unless otherwise noted. The data sources for India can be found in the
appendix.
4. Coping and Adaptive Capacity Indicators
The socio-economic conditions that bear on coping and adaptive capacity include
demographic characteristics, economics, politics/governance, management of natural resources and civil society.
4.1.
ECONOMIC CAPACITY
Wealth generally provides access to markets, technology, and other resources that
can be used to adapt to climate variability and change. Hence we include GDP (market) per capita as one of the proxies for national economic capacity. For India’s states
we found 1990–1991 state income and converted these to US dollars. However, in
societies where the distribution of wealth or income is very unequal, coping capacity
will also be unequally distributed. Thus we include unequal distribution of wealth
income within a society, indicated by the Gini coefficient as a component of our indicator of coping/adaptive capacity for our national comparisons3 (from Deininger
Sensitivity
Coping &
adaptive
capacity
Settlement/
infrastructure
sensitivity
Environmental
capacity
Human and civic
resources
Economic
capacity
Sector
Population no access
clean water/sanitation
Population at flood risk
from sea level rise
SO2 /state area
% Land unmanaged
Population density
Literacy
An income equity measure
Dependency ratio
GDP(market)/capita
Indicators/data
Access of population to basic services to
buffer against climate variability and change
Distribution of access to markets, technology, and other resources useful for adaptation
Realization of the potential contribution of
all people
Social and economic resources available for
adaptation after meeting other present needs
Human capital and adaptability of labor
force
Population pressure and stresses on ecosystems
Air quality and other stresses on ecosystems
Landscape fragmentation and ease of
ecosystem migration
Potential extent of disruptions from sea level
rise
Proxy for
(Continued on next page.)
Sensitivity ↑ as population with no access ↑
Coping/adaptive capacity ↓ as population
density ↑
Coping/adaptive capacity ↓ as SO2 ↑
Coping/adaptive capacity ↑ as % unmanaged land ↑?
Sensitivity ↑ as population at risk ↑
Coping/adaptive capacity ↑ as poverty or inequity ↓
Coping/adaptive capacity ↓ as dependency
↑
Coping/adaptive capacity ↑ as literacy ↑
Coping/adaptive capacity ↑ as GDP per
capita ↑
Functional relationship
TABLE I
Data-based indicators, grouped into sectors and into coping/adaptive capacity and sensitivity
66
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
Water resource
sensitivity
Human health
sensitivity
Ecosystem sensitivity
Water use
Life expectancy
Renewable supply and
inflow
Completed fertility
Fertilizer use/ cropland
area
% Land managed
consump-
Cereals
production/crop land area
Food security
Protein
tion/capita
Indicators/data
Sector
Withdrawals to meet current or projected
needs
Supply of water from internal renewable resources and inflow from rivers
Composite of conditions that affect human
health including nutrition, exposure to disease risks, and access to health services
Degree of modernization in the agriculture
sector; access of farmers to inputs to buffer
against climate variability and change
Access of a population to agricultural markets and other mechanisms (e.g., consumption shift) for compensating for shortfalls in
production
Degree of human intrusion into the natural
landscape and land fragmentation
Nitrogen/phosphorus loading of ecosystems
and stresses from pollution
Proxy for
TABLE I
(Continued).
Sensitivity ↓ as life expectancy ↑
Sensitivity calculated using ratio of available water used: Sensitivity ↑ as % water
used ↑
Sensitivity is ↓ if use < 60 kg/ha or > 100
kg/ha; neutral when use ≥ 60 and < 100
kg/ha
Sensitivity ↓ as fertility ↓
Sensitivity ↑ as % land managed ↑
Sensitivity ↓ as consumption↑
Sensitivity ↓ as production ↑
Functional relationship
CASE STUDY OF INDIA AND INDIAN STATES
67
68
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
and Squire, 1996, 1998). For India, we used, instead of the Gini, a measure of state
poverty and inequity. For missing values we substituted India’s national measure.
4.2.
HUMAN RESOURCES
Human and civic resources are another critical component of coping and adaptive
capacity. This category includes literacy, level of education, access to retraining programs, and other factors that determine how flexible individuals may be in adapting
to new employment opportunities or shifts in living patterns brought about by climate variability or change. As indicators, we selected the dependency ratio and the
literacy rate. The dependency ratio measures the proportion of economically active
and inactive individuals in a population; a higher rate of dependency would indicate that economically active individuals had many others to support, and resources
for adapting to changes in climate would be more limited. We calculated India’s
state-specific dependency ratios from state-specific percentages of male and female
urban and rural populations in the workforce and averaged these. The literacy rate
was included as a measure of the skills that individuals would have to have in order
to adapt.
4.3.
ENVIRONMENTAL COPING AND ADAPTIVE CAPACITY
As discussed previously, natural systems are sensitive to climate stimuli and thus
will need to adapt to climate variability and change. Adaptation may involve a
variety of eco-physiological changes, changes in species mix, migration, or even
the loss of some species or ecosystems. The survival of current ecosystems will
depend not only on the degree of climate variability or the rate and magnitude
of climate change but also on the baseline condition of the systems themselves.
For proxies of the resilience or coping and adaptive capacity of ecosystems, we
take three available measures of the amount of natural capital that is available:
population density, SO2 emissions/area, and the percentage of unmanaged land in
a country.
From the global emissions inventory activity’s (GEIA) gridded data of SO2
emissions and from the Carbon Dioxide Information Analysis Center’s (CDIAC)
gridded national and state codes, we generated national and state sulfur emission
estimates per hectare land. Population density was calculated for each state from
population data and land area data and checked against population density data.
National land use data were obtained from the Food and Agriculture Organization (FAO) (FAOSTAT98 and World Bank, 1998). Indian land use was explored
extensively because of the importance of its present use, potential use, and relationship to general environmental conditions, and climate and proxies. The data used
were calculated based on data referenced in the appendix.
CASE STUDY OF INDIA AND INDIAN STATES
69
5. Sensitivity Indicators
Sensitivity sectors include settlements/infrastructure, food security, ecosystems,
health, and water resources. For each sector, 1–3 indicators were selected to represent aspects of the sector that could be quantitatively modeled.
5.1.
SETTLEMENTS/INFRASTRUCTURE SENSITIVITY
Settlement sensitivity includes effects on economic activities in the industrial, energy, and transportation sectors, as well as effects on human settlements. Climate
variability and change have direct impacts through flooding, droughts, changes
in average temperatures (e.g. leading to thawing of permafrost), temperature extremes, and extreme weather events (e.g. hurricanes, cyclones). In addition, climate
variability and change can affect markets for goods and services in these sectors,
as well as natural resource inputs important to production (Acosta Moreno et al.,
1995). Settlements in coastal margins and on small islands are affected through sea
level rise and through storm surges, while these areas and inland settlements can be
affected by weather-related events that act directly on infrastructure (e.g. leading
to river basin flooding, landslides, and the like) and indirectly through effects on
other sectors (e.g. water supply, agricultural activity, human migration patterns).
Patterns of effect are different for urban and rural settlements, but both have been
shown to be sensitive to climate variability and change (Scott et al., 1995).
Three proxies are used for approximating sensitivity of settlements and infrastructure in industry, energy, and transportation: population at flood risk due to sea
level rise, and populations without access to clean water and sanitation. The projected number of people affected in a country by potential rise in sea level is based
on the estimated number of people in each country affected by sea surges (from
Delft Hydraulics, 1993).
Information on Indian state populations potentially affected by 1-m sea level
rise was scaled against national data for people affected by present-day storm
surges and people affected by 1-m sea level rise. The national data we used include
preventive measures against sea level rise, while the Indian state 1-m sea level rise
information does not include these, resulting in a possible overestimation of people
who are currently affected.
India’s available data on access to safe water were obtained from state information on water availability classified on the basis of the source of drinking water
as ‘safe’ or ‘otherwise.’ If the household had access to drinking water supplied
from a tap, handpump or tubewell situated within or outside the premises, it was
considered as having access to safe drinking water. Indian state data on access to
sanitation were not available and national data were used for all states.
The World Bank reports through the India water resources management reports,
which are co-published and distributed by Allied Publishers Limited of India, that
70
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
more than 75% of the rural population, some 520 million people in India, do not
have access to public water supply facilities and that achievements in sanitation
coverage have been even less extensive: only 3.6% of the rural population is covered at present. Actions to improve coverage of sanitation have been stepped up
recently through provision of subsidies and technical assistance for household construction of sanitation facilities. However, national guidelines and investments in
the rural water supply and sanitation sector neglected to ensure that the quality of
services to rural areas remained adequate. The demands on the urban water supply
and sanitation sector, which serves both urban domestic and industrial needs, are
tremendous. To date, the sector has under-performed, resulting in harmful impacts
on consumers, especially the poor, and on the environment.
5.2.
FOOD SENSITIVITY
Sensitivity to food security is defined as the potential for changes in the availability
of food in a particular geographic area. It encompasses both production of principal foodstuffs (e.g. crops, livestock, fish) as well as socio-economic issues such as
type of production system, access to production inputs that can offset changes in
climatic conditions, and access to markets for purchase of food. Climate variability
and change can affect food sensitivity through a variety of mechanisms, particularly related to food production. Variability in temperature and precipitation affects
crop production directly, as well as through impacts on soils (e.g. erosion), pest
and disease outbreaks, and other mechanisms. In addition, floods, droughts and periods of extreme temperatures can affect livestock and fisheries production (Stern
and Easterling, 1999). Climate change is projected to have impacts on agricultural
production through these mechanisms and through changes in atmospheric concentration of CO2 , which affects productivity and water use efficiency, particularly in
C3 plants. Impacts on agricultural production may also be felt through changes in
availability of water resources for irrigation (Reilly et al., 1995). Climate variability and change also cause changes in livestock and fisheries production through a
variety of mechanisms (Allen Diaz et al., 1995).
Two proxies represent food sensitivity in the VRIP. Cereals production per unit
area is intended to capture the degree of modernization in the agriculture sector
and the access of farmers to production inputs that can be used to buffer against the
effects of climate variability and change. Systems with high production per unit area
are presumed to be less sensitive than those with low production. Animal protein
consumption per capita is an imperfect proxy for the degree of modernization
in processing and distribution of agricultural goods for consumers. Populations
with high levels of animal protein consumption are presumed to have lower food
sensitivity than those with low levels of consumption. FAO data were used for
national information.
Because 60% of the population in India is agrarian, we paid particular attention
to agricultural data for the different states. Rice is the major agricultural product;
CASE STUDY OF INDIA AND INDIAN STATES
71
80% of the rice produced in India is grown in rain-fed lowlands. Increasing use
of technical means like tractors, harvesters, irrigation and fertilizer is apparent.
Imports of agricultural products have been steadily increasing, while exports have
been much more variable over the last 10 years.
For India, the animal protein indicator is probably an inappropriate indicator of
food security. Thus, in our within India state-by-state comparison we substituted
total protein consumption for animal protein.
5.3.
HUMAN POPULATION HEALTH SENSITIVITY
The health of human populations is affected by climate variability and change
through both direct mechanisms (e.g. heat waves in conjunction with episodes of
poor air quality, especially in urban areas) and indirect pathways (e.g. changes in
prevalence of vector-borne and non-vector-borne infectious diseases). Populations
with different levels of technical, social, and economic resources would differ in
their sensitivity to climate-induced health impacts. Sensitivity to climate variability
and change would be expected to be higher for those populations with poor basic
living conditions such as overcrowding, malnutrition, and inadequate access to
health services. Thus, sensitivity of human population health to climate conditions
can be expected to be highest in developing countries and among the poor in
transitional and developed countries.
Two proxies represent sensitivity of health to climate variability and change:
completed fertility and life expectancy. These variables represent a variety of conditions that affect human health, including nutrition, exposure to disease risks, and
access to health services.
5.4.
ECOSYSTEMS SENSITIVITY
Ecosystems and the functions they provide to individuals and society (e.g. providing food, fiber, medicines and energy; processing carbon and other nutrients;
purifying and regulating water resources; providing recreation and intrinsic value)
are sensitive to variation and change in climate. The composition and distribution of ecosystems has changed in the past in response to shifts in climate, and
models project future shifts in response to both the rate and magnitude of climate
change.
Mechanisms through which climate impacts are felt are similar to those for
agriculture, i.e. variation or change in precipitation and temperature, changes in
atmospheric composition that affect the competitive balance among different types
of plants, changes in soils, and changes in the incidence of diseases and pests.
Ecosystems are also influenced by other environmental stresses, including pollution
(both runoff in water courses and deposition from the atmosphere), increasing
extraction of resources, and incursion/fragmentation.
72
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
Two proxies represent the sensitivity of ecosystems: percentage of land area
that is managed, and fertilizer use per unit land area. The percentage of land under management is a proxy for the degree of intrusion of human activity into the
natural landscape and the potential fragmentation of land, which would increase
the sensitivity of ecosystems to climate variability and change. The percentage of
unmanaged land in a country consists of unmanaged and old forestlands. Fertilizer
use per unit area captures nitrogen and phosphorus loading of ecosystems and is a
proxy for ecosystem stresses resulting from pollution.
National land use and fertilizer use data were obtained from the FAO. India’s
land use was estimated from the referenced data in the appendix. The relationship
between fertilizer use and sensitivity to climate variability and change is nonlinear.
Values of 60–100 kg/ha are considered to result in the lowest ecosystem sensitivity.
If fertilizer use is less than 60 kg/ha, the deficiency in fertilizer use is projected
to increase sensitivity because nutrient deficits and low productivity in agricultural
systems may potentially result in cultivation of adjacent lands. As use increases
above 100 kg/ha up to a cap of 600 kg/ha, sensitivity increases due to increasing
loads of pollutant runoff. Between 60 and 100 kg/ha fertilizer use, sensitivity was
considered to be neutral.
5.5.
WATER AVAILABILITY
Climate variability already has had a large impact on the general hydrology of
a landscape and on the availability of water at the local and national scale, and
climate change can be expected to have as large or larger an impact. Presently, 19
countries around the world are classified as water-stressed (Watson et al., 1998).
This number can be expected to change due to population growth, changes in
land use, precipitation, and evapo-transpiration (linked to temperature increase).
Moreover, not only will socio-economic aspects of society be affected through
changes in water availability, but also government policies can be expected to
respond.
At the country level, we defined the sensitivity to water availability through one
proxy, composed of withdrawals to meet current or projected needs and (divided by
the sum of) the supply of water from internal renewable resources and inflow from
rivers. For water availability in India’s states we constructed a composite indicator
for water resources from five statistics:
•
•
•
•
•
The lengths of any rivers flowing through the state (per capita)
The surface area of reservoirs, tanks, and ponds within the state (per capita)
The total replenishable groundwater resource (per capita)
The ratio of domestic use to other uses
The level of groundwater development in each state (per capita).
The choice of these five statistics conforms to our knowledge of Indian practices. In rural areas (and some urban areas) clothes and bodies are likely washed
CASE STUDY OF INDIA AND INDIAN STATES
73
in rivers, while standing surface water and groundwater are used for drinking. To
make our comparisons of the state indicator values with those of India as a whole
and with global values we took the geometric mean of the five proxy values and
scaled these against India’s original water availability value in the prototype model
exercise. The result is that India’s water availability sensitivity remains identical to that in the previous comparisons and that the (scaled) water availability
sensitivities differ among the states but remain comparable to India’s and global
values.
6. Limitations of the Set of Proxies
Several critiques of the general methodology can be made. It may be argued that
the proxies are not representative enough, that the set is incomplete, and/or that the
proxies within the set are overlapping. These are issues for any indicator set; most
discussions of sets address the first concern but not the latter two. See, for example,
the environmental sustainability index documentation (YCELP and CIESIN, 2002),
the Wellbeing Index (Prescott-Allen, 2001) and the indicator set used by Aandahl
and O’Brien (2001) for India.
First, each proxy’s representativeness can be questioned. It may be argued,
for example, that a land use proxy may not explicitly account for soil characteristics or historic civil strife in an area. We would argue that current land
use is the result of several factors such as these and thus is an adequate proxy.
Although no proxy can perfectly represent an abstract category, judgment of
its adequacy should be based on the criteria described above: the proxy’s ability to summarize a number of important properties and its capability to be
quantified. In addition, there is an extremely important practical constraint that
good-quality data must be available, especially when comparisons are to be
made.
Second, the set of proxies used in the VRIP does not include relevant and
important characteristics such as governmental capacity, cultural worldviews and
beliefs, and institutional opportunities and limitations. Those proxies have been
developed and used in other quantitative studies, but they are both especially open
to bias and often do not include enough and/or good-quality data. For example, the
environmental sustainability index (YCELP and CIESIN, 2002) includes variables
such as “IUCN member organizations per million population,” “Civil & political
liberties,” and “Democratic institutions.” The data for civil and political liberties is
taken from Freedom House and incorporates the results of a survey: “The Survey
derives its information from a wide range of sources. Most valued of these are
the many human rights activists, journalists, editors, and political figures around
the world who keep us informed of the human rights situation in their countries”
(“About the survey,” Freedom House, 2002). The Wellbeing Index (Prescott-Allen,
2001) includes a category called “Freedom and Governance”; three of its four
variables are taken from Freedom House data. (The fourth is from Transparency
74
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
International’s Corruption Perceptions Index.) There is an obvious bias toward
western-style freedom and democracy in both of these indexes. In the present
study, we chose to include these factors in the qualitative rather than the quantitative
analysis.
Third, many of the proxies in the VRIP set are not independent. Protein intake
has obvious implications for human health, fertilizer use on cereals production, and
so forth. This is an issue for all proxy sets that seek to provide a complete picture
of a system or systems, and it is probably impossible to have proxies that are completely independent. Dependencies must be acknowledged as part of the analysis
of results and taken appropriately into account. Independent components are not a
requirement of an effective vulnerability index, however. Each indicator proxy is
chosen for a different reason and each represents a different aspect of vulnerability.
Therefore, if, for example, population (as reflected in per capita data) is reflected in
multiple indicators, it is because the effects of population are reflected in multiple
areas. Thus, the indicator set will accurately reflect the effect of population on the
overall vulnerability index.
For example, GDP per capita is somewhat correlated with life expectancy, as
expected. GDP per capita is included in the vulnerability index as a proxy for access
to markets, technology, and other resources useful for adaptation. Life expectancy
is a proxy for human health. Since these are completely different aspects of vulnerability, it is important to get a measure of both of them without imposing explicit
weights unless there is a theoretical basis for doing so (e.g. scenario assumptions
or particular situations). Their correlation and consequent larger effect on the index
is acceptable since they are more important than other variables that only affect a
single aspect of vulnerability.
The quantitative analysis methodology can be summarized as follows:
• collect the necessary 17 indicator values for each country listed in Table I,
• scale the indicators against world values,
• calculate index numbers for each sector as geometric means of each sector’s
indicators (the climate-sensitivity sectoral indicators [negatives] and the coping/adaptive capacity sectoral indicators [positives]),
• calculate the overall sensitivity and the overall coping/adaptive capacity as
geometric means of the participating sectors,
• calculate the net vulnerability-resilience indicator value as the simple algebraic summation of a state’s sensitivity and adaptive capacity.
Finally, it must be recognized that there is a degree of arbitrariness in any set of
indicators, and that the availability of good quality data will always place limits on
developing a fully exhaustive set. However, what remains essential is that a dataset
be as meaningful as possible for the issue to be addressed.
The quantitative analysis in such study results in snapshots of current country
or state level vulnerability.
CASE STUDY OF INDIA AND INDIAN STATES
75
Figure 3. Vulnerability rankings of 38 countries and the world.
7. National Results and Results for India and Indian States
In the previous study (Moss et al., 2001), the prototype model produced quantitative
vulnerability-resilience assessments for 38 countries and the world (see Figure 3).
In that study of country-level vulnerability, India ranks second-to-last.
Our VRIP methodology (Moss et al., 2002) was used next for 103 countries.
In that case India again scored very low, indicating high vulnerability. This high
vulnerability as compared to other countries is confirmed by other indexed rankings
of India (see Table II, YCELP and CIESIN, 2002; World Bank, 1998). Note that
the YCELP and CIESIN index is intended for environmental purposes, however,
while HDI indicators are for human development purposes. The VRIP integrates
both kinds of phenomena. When we use the HDI method and VRIP data from 1990
for all 103 countries, we find a somewhat lower HDI score, indicating that it matters
which dataset is used.
In a cross-nation health-only comparison, India ranks 133 out of 191 countries
with regard to Healthy Life Expectancy (HALE). Healthy life expectancy includes
adjustment for time spent in poor health.4 For India, HALEs amount to 51.4 years
for children born in 2000 and an estimated expected loss of healthy life years at
birth of 8.4 years for males and 10.4 years for females.
This first, highly aggregated comparison is only a starting point for an analysis.
We need to examine why India is so vulnerable. To do that, we look at the whole
set of indicators for India.
76
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
TABLE II
India’s ranking using different methods (103 countries)
Method
Notes
India ranking
VRIP world indexing
YCELP/CIESIN—
Environmental
sustainability index
HDI
HDI (estimated)
Method used by Moss et al. (2001)
Using CIESIN data and publication
101/103
80/103
Using UNDP data and publication
Estimated using VRIP data from
1990 and slightly altered
calculation mechanisma
77/103
82/103
a
While the HDI indicator is normally based on ranges with defined bounds, we used the
maximum and minimum values of the country data to determine the ranges.
Figure 4. Scaled proxy values for India.
7.1.
INDIA’S INDICATOR CONTRIBUTIONS TO VULNERABILITY
Figure 4 shows the values of various proxies (scaled to global vulnerability
values) contributing to India’s vulnerability for the year 1990 when countrylevel data were used. World values are −100 for sensitivity and +100 for coping/adaptive capacity; these two values add up to a global vulnerability-resilience
index of zero. Interpretation of these graphs must take into account that the
CASE STUDY OF INDIA AND INDIAN STATES
77
proxies, as part of a hierarchical structure, contribute more or less directly to
the final VRIP index. Obvious negatives that make large contributions to vulnerability are availability of animal protein and water. However, animal protein
is not necessarily an appropriate indicator for India, because many dietary customs in India relate to religious and cultural beliefs. This issue is addressed in
the next section. Water availability is an issue in the country, which is plagued
by drought in some places and by flooding in other parts, affecting the productive use of water for human use, sanitation, industry, and irrigation. India
is trying to realize more of its potential for irrigated crops, thereby lessening
the effects of extreme climate events on agriculture. This vulnerability assessment points us to water, including irrigation schemes, as a potential trouble spot.
This proxy was also re-evaluated according to the description of India’s water
availability.
On the positive side, India’s democracy results in equity that is slightly higher
than the global average. The dependency ratio—the percentage of the population
dependent on the percentage of the population in the workforce—is relatively high.
This indicates that many people are available for the workforce, supporting relatively few people other than themselves. The sulfur emission proxy (representing
lower emissions than the global average) adds to adaptive capacity; the level of
sulfur emissions over this large country is less detrimental to human health than
the global average.
7.2.
VULNERABILITY OF INDIA: MODIFICATIONS TO THE ANALYSIS
We used the same VRIP methodology to analyze the vulnerability-resilience of
states within India to gain further insights into the robustness of our methodology and its scale-dependency. First we collected the necessary 17 indicator
values from various sources (listed in the appendix), and substituted total protein consumption for animal protein consumption. For our analysis of the vulnerability and/or resilience of states within India we used data from various
years between 1990 and 1998, because all data were not available for a particular base year. This is not necessarily a weakness in the data, since data
from one year may not represent the general status of a situation. Often, it
is advisable for indicator calculations to incorporate a representative mean of
the variable information to be processed. A five-year average of a proxy value
might be a better indicator than a single-year value to answer the research
questions.
India’s national data, which in this next analysis were based on state-level data,
differed slightly from previously collected data with regard to impacts of sea storm
surges and potential sea level rise, national income, and fertilizer use, cereal production, and access to safe water. Figure 5 show the changes in the scaled proxy
values. Table III shows that, in sum, India’s VRIP index decreased from −92 to
78
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
TABLE III
Changes in India’s values for vulnerability-resilience, sensitivity, and coping/adaptive capacity from
the prototype experiment to the modified analysis
Original results for India
India with protein consumption re-evaluated (a)
India proxies re-evaluated based on state data
India with proxies re-evaluated based on state
data and with comprehensive water sensitivity (b)
Vulnerabilityresilience
Sensitivity
Coping and
adaptive capacity
−92
−77
−80
−80
−145
−131
−136
−136
54
54
56
56
Figure 5. Differential contributions by the proxies for India based on revised proxy values.
−80 based on these changes. The most significant change came from substituting demand for ‘protein in general’ for demand for animal protein as one of the
proxies for food security. India’s overall vulnerability-resilience improved based
on this difference: India became relatively less vulnerable because of increased
food security (from an indicator value of −92 to −77). The newly calculated national water proxy value was scaled to the initial water proxy value and has therefore no impact on India’s national VRIP index while it allows for state-specific
impacts.
CASE STUDY OF INDIA AND INDIAN STATES
79
Figure 6. An overview of the range of the vulnerability-resilience indicators for the different states
in India, calculated with the VRIP methodology.
7.3.
VULNERABILITY OF INDIAN STATES
The resulting range of vulnerability-resilience indicator values for the different
states in India is shown in Figure 6. Six states are more vulnerable than India as a
whole, while the majority of states are more vulnerable than the global average. All
six most vulnerable states are coastal states, most with high population densities.
The small mountainous northern inland states show the highest resilience among
Indian states; however, these states represent only a low percentage (less than 3%)
of India’s total population.
To explore the components of vulnerability as reported in the model, we examine
the values for sectors and indicators in the states. These results are presented in
Figure 7 for sensitivity and Figure 8 for coping/adaptive capacity.
The resilient states (all mountainous and inland, as noted above) have high
values for environmental capacity, which can be attributed to low sulfur emissions
and a relatively large percentage of unmanaged land (Figure 8). Determining the
basis for higher or lower resilience is important, for the analyst can then evaluate
the adequacy and completeness of the set of indicators. For example, in the case
of mountainous states with relatively clean air, the indicator set does not account
for the risks of increased erosion, mudslides and other natural hazards that may be
impacts of climate change.
In contrast to the inland states, all coastal states show high vulnerability, especially Goa, for which it is reported that over 7% of the population would be
affected by 1-m sea level rise. This translates into a high sensitivity to sea-storm
80
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
Figure 7. Sensitivities of the different states in India and the participating sector values, calculated
with the VRIP methodology.
Figure 8. Adaptive and coping capacity of the different states in India and the participating sector
values, calculated with the VRIP methodology.
CASE STUDY OF INDIA AND INDIAN STATES
81
surges. Orissa and Tamil Nadu also show high sensitivity to sea-storm surges. For
comparison, Goa’s value for settlement sensitivity (which includes sensitivity to
sea-storm surges) is almost 100 times the value for mountainous Himachal Pradesh.
Goa, West Bengal and Kerala, among the coastal states, also have high sensitivity
in (fresh) water availability. Assam’s similarly high sensitivity in water availability
contributes to its lower resilience compared to other mountainous states.
States show a wide range of sensitivity to food security (Figure 7), with Kerala
and Sikkim being rather sensitive, but not Punjab. Ecosystem sensitivity is found to
be the largest in Punjab, mainly because of the polluting consequences of fertilizer
use.
With regard to economic capacity (Figure 8), all states rank lower than the
global representation (calculated in Moss et al., 2001). Kerala shows the highest
economic capacity among Indian states because of its relatively low inequality;
however, differences among the states are not large.
Interestingly, most of the variability in state-level sensitivity proxies results
from variability in settlement and food sensitivity, i.e. social and economic factors. On the other hand, most of the variability in state-level proxies for coping
and adaptive capacity results from environmental rather than economic or human
resource factors. If these results prove robust, they imply that social policies would
be more likely to be effective in reducing sensitivity, while environmental protection policies would likely be more effective in increasing coping and adaptive
capacity.
Redirecting the assessment from India as a whole to Indian states brings into
play a new level of detailed analysis that is possible. India’s states are diverse in
cultures, approaches to governance, natural resources, and emphases. They have
been differentially affected by climate, both because of their dependence or nondependence on agriculture and because of the climate events themselves.
Many analyses attempt to measure these kinds of differences by using GDP per
capita or some income measure as a summary proxy. However, Figure 9 clearly
shows that there is no meaningful correlation between net domestic product (NDP)
per capita in states5 and vulnerability of a state to climate impacts. Although annual
per capita NDP at current prices per capita ranges from 2000 Rs. for Bihar to 8797
Rs. for Goa and 11057 Rs. for Delhi, 6 most of the states have NDP per capita within
a very narrow range; however, states exhibit a wide range of vulnerabilities. Therefore, differences other than economic ones among states are of great importance.
For example, literacy rates in the states range from 44 to 91%, life expectancy from
58 to 70 years. And there are great disparities, as noted in the previous section, in
natural resources and climate hazards among Indian states.
In the next section, we first examine aspects of vulnerability in Orissa, comparing the model results to recent historical information. Then we present two brief
comparisons. Kerala, where the development emphasis has been on human capital, is compared with Maharashtra, which has focused on economic development.
Andhra Pradesh, a coastal state, is compared with Himachal Pradesh, an inland state;
82
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
Figure 9. No correlation can be shown between vulnerability and NDP per capita.
in this comparison, geography plays a large role, but other factors are important
as well.
7.4.
VULNERABILITY IN ORISSA
Compared to India as a whole, the state of Orissa is significantly poorer, less
industrialized, and lower in human development. Table IV shows comparisons
between India and Orissa. Orissa has a lower percentage of urban population than
India as a whole, a lower life expectancy, a similar birth rate for 1996, but a higher
death rate and a lower level of growth in productivity. Many factors may contribute
to Orissa’s lagging India, but environmental conditions is one likely factor.
The frequency of extreme weather events in India—for example, droughts,
heat waves, and floods—has increased over the past two decades. Orissa has been
reeling under contrasting extreme weather conditions for more than a decade: from
heat waves to cyclones and from droughts to floods. Since 1965, calamities are
not only becoming more frequent but striking areas that never had a vulnerability record. For instance, a heat wave in 1998 killed around 1,500 people. Most of
the casualties were in coastal Orissa, a region otherwise known for its moderate
temperatures.
Two very severe cyclonic storms occurred in October 1999, affecting 16 districts
of Orissa, 12,922,000 people, 14,643 villages, and 1,842,000 hectares of crop area.
The Natural Disaster Management Division of the Ministry of Agriculture coordinated relief and rehabilitation efforts to mitigate problems of affected people. The
Prime Minister released Rs. 5,000,000,000 from the National Fund for Calamity
Relief and Rs. 4,500,000,000 were advanced as plan assistance.7
CASE STUDY OF INDIA AND INDIAN STATES
83
TABLE IV
Comparison of India and Orissa in human and economic development
Basic statistics
Area (1000 sq km)a
Population (millions) for 1991
(1996)a
Population density, 1991 (persons/
sq km)a
Total population, 1991 (%)
Urban population for 1991 (%)a
Human development
Literacy (%) (for 1991 and 2001)b
Birth rate (for 1991 & 1996, per 1000
persons)c
Death rate (1991 & 1996, per 1000
persons)c
Infant mortality rate (1996, per 1000
births)c
Life expectancy (1990–1992)
Economic development
Net state domestic product (1990–
1991) at current (2001–2002) prices
at factor cost (and annual growth
rate: % yoy)a
Net state domestic product (1998–
1999) at current (2001–2002) prices
at factor cost (and annual growth
rate: % yoy)a
Per capita net state domestic product
at current prices (2001–2002 Rs)
(1990–1991)a
Per capita net state domestic product
at current prices (2001–2002 Rs)
(1998–1999)a
Installed generating capacity (MW)
(March 1998)a
a
India
Orissa
846.0 (934.2)
156
31.7 (34.4)
273
203
100
25.7
3.9
13.4
52 (1991)
65 (2001)
(76% male; 54% female)
27.5 (29.3 rural; 21.6 urban, 1996) [30.9 rural; 24.3
urban (1991)]a
9.0 (9.7 rural; 6.5 urban,
1996) [10.6 rural; 7.1 urban, 1991]a
72 (77 rural; 46 urban)
49 (1991)
64 (2001)
(76% male; 51% female)
27.0 (27.7 rural; 21.3 urban, 1996)
57.7 for males; 58.7 for
femalesa
55.9 for males; 54.8 for
femalesd
4181 (17.0%)
96.6 (unknown)
14150 (15.5%)
295 (7.4%)
4983 (14.4%)
3077
14396 (13.3%)
8324
89167 MW
1693 MW
http://www.indiainfoline.com/econ/andb/contents.html.
http://www.censusindia.net/results/provindia3.html.
c
http://www.censusindia.net/srs1.html.
d
http://www.undp.org.in/report/preidf98/TABLE02.htm.
b
10.8 (11.2 rural; 7.5 urban)
96 (99 rural; 65 urban)
84
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
Almost 490,000 ha of fertile lands have been waterlogged, salinated and sandcasted in coastal Orissa from cyclones and floods. The devastating floods in 2001
(15 floods were reported between July 8 and August 10) induced crop failure worth
Rs. 150 billion. This was the worst flood recorded in the last century. Orissa was
provided Rs. 5,350,000,000 and 3 LT of rice worth Rs. 2,100,000,000 under the
Food for Work Scheme and other aids.8 In 2002, while drought conditions prevailed
in Vidarbha during June and July, heavy downpours in August amounted to 80 cms
of rainfall (95 cms is normal for the season), resulting in inundation of hundreds of
villages in the region. Such intense rainfall events have become more frequent in
recent years in many parts of India, Nepal and Bangladesh.
In contrast, west Orissa experiences drought that is chronic, both in recurrence
and in expanse. The drought in the year 2001, the worst drought since 1866, affected 25 of the State’s 30 districts and engulfed districts like Sundergarh and the
Kendrapada that have been historically drought-free. By February 2001, people
started migrating. The worst affected districts like Kalahandi and Balangir reported
60% less rainfall than normal. The situation in nine western districts was severe
because it was the second consecutive drought. By May 2001, 61 starvation deaths
had already been reported. The state government put the economic loss due to crop
damage at Rs. 64.289 billion. The Orissa drought in 2001 affected the lives of 11
million people.
Orissa’s vulnerability, as assessed in the VRIP model, is high. Four sensitivity
sectors have slightly lower values than the world values, but three (water availability, cereal production, and and access to safe water) have significantly higher
(−) values than India as a whole and the global value of −100 (see Figure 10).
These sensitivities accord with the recent history of storms, floods, and droughts.
Economic capacity in Orissa is relatively low; the high sulfur emission proxy indicates low economic development but relatively clean air compared to India as a
whole. Ways of building resilience to current climate variability, beyond immediate aid, will allow Orissa to also enhance economic and human resource capacity.
Resilience-building policy needs to include ways to cope with both flooding and
drought.
7.5.
COMPARISON BETWEEN MAHARASHTRA AND KERALA
The nature and severity of direct climate impacts, and the type and extent of government intervention affect vulnerability in any state. So, too, does the state’s approach
to economic and human development generally. Within India, we may discern several different approaches. A brief discussion of two states, Kerala and Maharashtra,
will illustrate a fundamental difference. Kerala has focused on human development;
in contrast, Maharashtra has emphasized industrial/economic development. The results are evident in the statistics presented in Table V. Kerala has a much higher
literacy rate (particularly for females), lower birth and infant mortality rates, and
CASE STUDY OF INDIA AND INDIAN STATES
85
Figure 10. India’s and Orissa’s scaled proxy values determining vulnerability.
longer life expectancies than Maharashtra. Maharashtra has a much higher growth
rate, higher per capita income, almost seven times the installed generating capacity, and almost five times the state domestic product (for less than three times the
population).
Kerala has made specific choices to achieve its high levels of human development, although these are partly attributed to the enlightened nineteenth century
maharajas and the work of Christian missionaries. The work of human development
has been continued by the activism of the ordinary people. Several Left Front and
Communist party elected governments have carried out the demands of popular
progressive movements. Franke and Chasin9 define the “Kerala Model of Development” as having three principal characteristics:
• A set of high material quality-of-life indicators coinciding with low per-capita
incomes, both distributed across nearly the entire population of Kerala
• A set of wealth and resource redistribution programs that have largely brought
about the high material quality-of-life indicators
• High levels of political participation and activism among ordinary people
along with substantial numbers of dedicated leaders at all levels. Kerala’s
mass activism and committed cadre were able to function within a largely
democratic structure, which their activism has served to reinforce.
However, the Kerala model has been criticized for its failure to strengthen its economic base, particularly industry. In addition, strong labor unions and high wages
86
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
TABLE V
Comparison of Kerala and Maharashtra in human and economic development
Basic statistics
Area (1000 sq km)a
Population (millions) for 1991
(1996)a
Population density, 1991 (persons/
sq km)a
Percent of total population, 1991 (%)
Urban population 1991 (%)a
Human development
Literacy (%) (for 1991 and 2001)b
Birth rate (per 1000 persons)c
Death rate (per 1000 persons)c
Infant mortality rate (1996, per 1000
births)c
Life expectancy (1990–1992)d
Economic development
Net State Domestic Product at
factor cost (1990–1991) at current
(2001–2002) prices (and annual
growth rate: % yoy)a
Net state domestic product at
factor cost (1998–1999) at current
(2001–2002) prices (and annual
growth rate: % yoy)a
Per capita net state domestic product
at current prices (2001–2002 Rs.)
(1990–1991)a
Per capita net state domestic product
at current prices (2001–2002 Rs)
(1998–1999)a
Installed generating capacity (MW)
(March 1998)a
Agriculture as the occupation (%)a
a
Kerala
Maharashtra
38.9
29.10 (31.0)
307.7
78.90 (86.6)
748
256
1.1
26.4%
9.7
38.7
90 (1991)
91 (2002)
(94% male; 88% female)
18.0 (18.0 rural; 17.9
urban)
6.2 (6.3 rural; 6.0 urban)
14 (13 rural; 16 urban)
65 (1991)
77 (2002)
(86% male; 68% female)
23.4 (24.9 rural; 21.0
urban)
7.4 (8.7 rural; 5.4 urban)
48 (58 rural; 31 urban)
68.8 for males; 74.4 for
females
63.1 for males; 64.7 for
females
122 (16.0%)
582 (16%)
511 (13.8%)
1883 (9.5%)
4200
7439
16029
20356
1766
10546
Nearly 50
70
http://www.indiainfoline.com/econ/andb/contents.html.
http://www.censusindia.net/results/provindia3.html.
c
http://www.censusindia.net/srs1.html.
d
http://www.undp.org.in/report/preidf98/TABLE02.htm.
b
CASE STUDY OF INDIA AND INDIAN STATES
87
Figure 11. Maharashtra and Kerala’s scaled proxy values.
have kept industries and investment away from Kerala. The state depends, perhaps
too much, on remittances from local people who are working abroad (mainly in the
Gulf region).
Maharashtra describes itself as “the industrial powerhouse of India,” with Mumbai (Bombay), its capital, as the commercial center of the state. Major international
banks operate in Mumbai, and the two largest domestic banks are headquartered
there. Its port, built by the British, remains a trade center.
Is what we know about the descriptive statistics of Maharashtra and Kerala
reflected in their vulnerability-resilience indicator differences and/or in the more
detailed indicator value differences?
Despite its higher human development levels, Kerala has a much higher vulnerability index than Maharashtra (−101 versus −41), attributable mainly to higher
sensitivities in water availability (Figure 11). Maharashtra has abundant developed water resources. In this comparison, geography, including available natural resources, has more effect on the differing vulnerabilities of the states than
their different development paths. Kerala’s low score in available water is due
to its low scores for the level of groundwater development; total replenishable
ground water per capita; river length per capita (it hugs the coast); and built
canals, reservoirs, tanks, and ponds per capita10 compared to the other states.
However, Kerala is rich in natural waterways—but a great deal of that water is
brackish.11
88
7.6.
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
COMPARISON BETWEEN ANDHRA PRADESH AND HIMACHAL PRADESH
Kerala and Maharashtra exemplify the contrasts in development styles in India;
Andhra Pradesh and Himachal Pradesh demonstrate major differences in geography
as well: coastal versus mountainous. The two states are dissimilar in many ways,
illustrating the great diversity of conditions and vulnerabilities at the sub-national
level (see Table VI). However, the two states have comparable rates of growth in net
state domestic product (21.1% for Andhra Pradesh, 21.8% for Himachal Pradesh
for 1991–1998), comparable low income, and comparable infant mortality rates.
Their differences include low (Himachal Pradesh) versus higher (Andhra Pradesh)
population density and associated low versus higher percentage of urban population.
Himachal Pradesh has much higher literacy rates despite its overwhelmingly rural
population.
Andhra Pradesh is the fifth largest state in India, both in area (275,048 sq km)
and population (66,508,008 in 1991). The literacy rate is 71% for males and 51%
for females. Andhra Pradesh is following a mixed strategy, still overwhelmingly
agricultural but with great efforts being made in developing a high-technology
sector. About 70 percent of the population works in agriculture, and the state is one
of India’s main rice-producing areas.12 Although poor, Andhra Pradesh “has been
a supplier of skilled software labor for more than a decade and is one of the most
innovative sites in applying IT for social development” (Eischen, 2003).
Andhra Pradesh’s Annual Plan developed for the National Planning Commission
for the year 2002–2003 was finalized in July 2002. The Deputy Chairman of the
Planning Commission complimented the state for initializing economic reforms,
achieving reductions in population growth in spite of low per capita income and
literacy, using technology extensively in government, and initiating power sector
reforms. Rural poverty is lower than urban poverty, said the Deputy Chairman,
reflecting positive reforms. However, inequality, poor revenue generation by the
road transport corporation, and the need to encourage self-help groups were also
cited.
In 2002, under drought conditions in Andhra Pradesh, the state released dam
water for electricity generation but not for irrigation. Poor farmers respond to such
policies by smashing the pumps of their richer neighbors. Quarrels over water
rights between states can be bitter, too. Tamil Nadu claims that its neighbor state,
Karnataka is violating agreements about sharing water from the Cauvery River
(“Water in India”, 2002).
Himachal Pradesh, in contrast, is much smaller (55,673 sq km) and has less
than a tenth of Andhra Pradesh’s population (5,170,877 in 1991). It is situated
in the heart of the western Himalayas, with elevations ranging from 350–7000 m
above sea level. Most people depend on agriculture, especially income derived
from sheep, goats, and other grazing animals. However, the state has experienced
climate variability. For example, unprecedented flooding occurred in the river Sutlej
on August 1, 2000, causing widespread damage in Himachal Pradesh.
CASE STUDY OF INDIA AND INDIAN STATES
89
TABLE VI
Comparison of Andhra Pradesh and Himachal Pradesh in human and economic development
Andhra Pradesh
Himachal Pradesh
Area (1000 sq km)a
275.1
55.7
Population (millions) for 1991
(1996)a
Population density, 1991 (persons/
sq km)a
Percent of total population, 1991 (%)
66.50 (72.2)
5.20 (6.0)
241
92
8.2
0.6
Urban population 1991 (%)a
26.9
8.7
44% (1991)
64% (1991)
61% (2001)
77% (2001)
(71% male; 51% female)
(86% male; 68% female)
Birth rate (1996, per 1000 persons)
22.8 (23.5 rural; 20.6
urban)
12.0 (23.5 rural; 17.1
urban)
Death rate (1996, per 1000 persons)c
8.4 (9.2 rural; 5.9 urban)
8.0 (8.2 rural; 6.0 urban)
Infant mortality rate (1996, per 1000
births)c
Life expectancy (1990–1992)d
65 (73 rural; 38 urban)
62 (63 rural; 40 urban)
59.0 for males; 61.5 for
females
63.8 for males; 64.2 for
females
312 (22.6%)
25.2 (15.4%)
1039 (21.1%)
95.1 (21.8%)
4531
4910
13993
16144
6286
299
70
about 87
Basic statistics
Human development
Literacy (%) (for 1991 and 2001)b
c
Economic development
Net state domestic product at factor
cost (1990–1991) at current (2001–
2002) prices (and annual growth
rate: % yoy)a
Net state domestic product at factor
cost (1998–1999) at current (200120-02) prices (and annual growth
rate: % yoy)a
Per capita net state domestic product
at current prices (2001–2002 Rs)
(1990–1991)a
Per capita net state domestic product
at current prices (2001–2002 Rs)
(1998–1999)a
Installed generating capacity (MW)
(March 1998)a
Agriculture as the occupation (%)a
a
http://www.indiainfoline.com/econ/andb/contents.html
http://www.censusindia.net/results/provindia3.html
c
http://www.censusindia.net/srs1.html
d
http://www.undp.org.in/report/preidf98/TABLE02.htm
b
90
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
Figure 12. Andhra Pradesh and Himachal Pradesh’s scaled proxy values.
Is what we know about the descriptive statistics of Andhra Pradesh reflected
in its vulnerability-resilience indicator set and how does this compare with the
indicator set values from a northern mountainous state like Himachal Pradesh?
In the vulnerability assessment, Andhra Pradesh is similar to India as a whole.
Several illuminating contrasts are apparent when we examine Andhra Pradesh
and Himachal Pradesh side by side (Figure 12). Andhra Pradesh is vulnerable; Himachal Pradesh is resilient (although modestly). The largest contrasts
are in three indicators. Andhra Pradesh has much higher sensitivities in safe
water, managed land as percent of total land, and fertilizer use/land area. Himachal Pradesh, like most of the mountainous states, has relatively high environmental capacity—less managed land, clean air, and lower population density. This accords with the finding that the variation in sensitivity seems to
result primarily from variability in human systems, while variation in coping
and adaptive capacity seems to result primarily from variation in environmental
capacity.
8. Conclusions
The results of this study both increase our understanding of the sources of vulnerability in India and demonstrate the viability of approaching vulnerability assessment
using an indicator-based modeling approach.
CASE STUDY OF INDIA AND INDIAN STATES
91
9. Conclusions for India and Indian States
This methodology, like other indexing methodologies, ranks India as very
vulnerable compared to other countries in the world. India’s vulnerability stems
from certain governance and development issues that are not captured in the VRIP
analysis, and from environmental and social conditions that are represented in the
VRIP by indicators.
The largest contributions to vulnerability of India in the VRIP analysis are total
protein intake and water availability. Water availability was shown as an important
sensitivity sector. And, although India is listed under the water-abundant category
of countries with 2,464 m3 of renewable annual fresh water available per person in
199013 or an annual total of 1850 km3 ,14 water availability is a regional phenomena
and, “According to long-range United Nations population projections, India’s population could, under the low-growth projection, stabilize and eventually shrink in
size after 2025, which would actually increase per capita water availability. Under
the medium and high population projections, however, the country would move
into conditions of water stress and water scarcity, respectively.” India gets most of
its water during the four-month monsoon; thus, dryland agricultural production and
replenishment of reservoirs depends on the monsoon. Water is mainly a state, not
the central government, issue; water is, at the same time, considered a social good.15
The India water resources management report16 states that groundwater is an important source of drinking water and food security for India’s 1 billion inhabitants,
and that it supplies 80% of water for domestic use in rural areas and perhaps 50% of
water for urban and industrial uses. It states that rapid expansion in the use of groundwater, primarily for irrigation, has contributed significantly to agricultural and overall economic development in India but that, in many arid and hard-rock zones,
increases in overdraft areas and associated water-quality problems are emerging.
The report identifies two broad issues that need to be addressed and proposes an
action plan. First, solutions must be found for competing inter-sectoral demands.
Mechanisms must be developed for allocating scarce water resources between competing uses such as irrigation, rapidly expanding domestic and industrial needs,
hydropower, and environmental requirements. Second, water must be managed on
a river basin basis, including states sharing the same river basin.
On the positive side, India’s democracy results in equity scoring slightly higher
than the global average. The dependency ratio—the percentage of the population dependent on the percentage of the population in the work force is relatively high. This
indicates that many people are available for the workforce, supporting relatively
few people other than themselves. And the sulfur emission proxy (representing
lower emissions than the global average) adds to adaptive capacity as well.
The state-level results demonstrate the wide variety in sources of vulnerability within India and reinforce the need to formulate differentiated strategies to
reduce vulnerability and build resilience. For example, states show a wide range
of sensitivity to food security; Kerala and Sikkim are more sensitive than Punjab.
92
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
However, Punjab has the highest ecosystem sensitivity, mainly because of the polluting consequences of fertilizer use. Therefore, policies aimed at building food
security in states like Kerala and Sikkim should also be designed to provide incentives to limit fertilizer use.
All states rank lower than the global representation of economic capacity. Kerala
shows the highest economic capacity among Indian states because of its relatively
low inequality; however, differences among the states are not large. This finding
reinforces India’s current general policy that emphasizes economic development.
As was discussed previously, most of the variability in state-level sensitivity
proxies results from variability in settlement and food sensitivity, i.e. social
and economic factors. On the other hand, most of the variability in state-level
proxies for coping and adaptive capacity results from environmental rather than
economic or human resource factors. If these results prove robust, they imply that
social policies would be more likely to be effective in reducing sensitivity, while
environmental protection policies would likely be more effective in increasing
coping and adaptive capacity.
For the resilient states (all mountainous, inland, and sparsely populated), development policies should protect unmanaged land, but also account for factors not
covered in the quantitative analysis, such as risks of increased erosion, mudslides
and other natural hazards that may be impacts of climate change.
In contrast to the inland states, all coastal states are candidates for national
and state policies to prepare them for sea level rise and increased storm surges.
However, national and local governments must also deal with freshwater issues
and build resilience to drought as well as flooding.
10. Conclusions for the Indicator-Based Modeling Approach
The results show that a vulnerability analysis using this framework provides a means
to ask structured questions about the elements of vulnerability in particular places.
The modeling framework allows analysts and stakeholders to systematically evaluate individual and sets of indicators and to compare geographical or political units
(for example, river basins or states). The transparency of the framework provides
ready explanations for differences in the result. Thus, the framework indicates, to a
first approximation, where the likely vulnerabilities are in the area being assessed;
these areas are then candidates for a more in-depth, qualitative examination.
Ideally, looking at the sub-national level would help us to better understand the
tradeoffs among different assets for coping/enhancing resilience, and between different environment and development policy choices. We know that, in the abstract,
these tradeoffs involve many factors, including the rate of economic growth, the
quality of the resulting wealth (its sectoral composition and degree of vulnerability
to future climate changes), and tradeoffs across sectors (e.g. increasing agricultural
resilience through increased irrigation, but corresponding increases in vulnerability
CASE STUDY OF INDIA AND INDIAN STATES
93
in water resources and ecological systems). These different types of assets can be
used to cope with climatic variation or increase resilience (adapt in the longer term).
One way to examine this issue, a step beyond what has been accomplished in the
current study, would be to survey actual events and trying to compare how different states reacted to them, using the framework of indicators to see whether the
indicators capture the strengths and weaknesses of the different actual responses.
The use of the VRIP analysis, at the national and state levels, points the way
toward such useful, sector-level or place-based analyses. Another strategy for extending vulnerability assessment to these levels is to engage stakeholders in choosing or modifying indicators (and, perhaps, sectors) to investigate questions that are
important to those who may need to adapt to climate change.
A few caveats are in order. To achieve a functional framework, we have used the
existing structure of the VRIP and somewhat modified requirements for indicators;
thus we are at a preliminary scoping and surveying stage. The framework and results
presented here represent a first-order modeling effort, which needs to be refined
and developed. There are issues of interactions among indicators and projections
into the future that lend themselves to exploration.
We did not attempt to extend our VRIP modeling exercise with scenario building
and projections. In our previous experience with the VRIP model we calculated
future vulnerability-resilience indicator values by using the methods and values
of the IPCC’s Special Report on Emissions Scenarios (SRES) (Nakicenovic et
al., 2000) scaled to the regional model outputs of PNNL’s integrated energy and
economics model, MiniCAM and its post-processor Sustain (Pitcher, 1999). For the
global and national projections, scaling down from aggregate, established scenarios
was an adequate approach.
Attempting to make projections into the future for Indian states would require
a more detailed level of approach. Not only would we need a re-evaluation of
baseline data used, with regard to initial values for the calculations, we also would
need to look at state-by-state historical development pathways, e.g. income growth
in addition to state plans for future development. Water availability, as discussed
previously for India, is just one of the issues that would need to be tackled in
attempting vulnerability-resilience projections.
11. Toward an Improved Methodology
Several key issues emerged in using the VRIP for this study. The VRIP was designed
to be straightforward and transparent. The structure, mathematical process, and
method merited examination and evaluation of possible improvements, however.
First, the hierarchical structure of the VRIP introduces implicit weighing of
the participating proxies given that each layer in the hierarchy is comprised of a
different number of values.
Second, the VRIP indexing method, which is formulated as dividing a country’s
proxy value by the world’s proxy value, results in a range of indexed (scaled) proxy
94
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
values that has mathematical consequences: (a) proxies with a large range will have
a relatively larger impact on the index comparisons than proxies with a small range
and (b) high sensitivities are weighed less than low sensitivities when a geometric
mean is used in the aggregation process because, in the VRIP, sensitivities are
valued as negative numbers.
These issues led to the development of a VRIP2 method where exposure remains defined as the nature and extent of the changes a location’s climate may be
subjected to, and sensitivity and capacity to cope/adapt are expressed as indicators
of the response of socioeconomic and natural systems to exposure. Any assumed
or forecasted change in exposure will lead to changes in indicator values given
that all aspects of vulnerability and resilience depend on location-, sensitivity- and
capacity to respond.
The hierarchical structure was retained, because altering the structure (even
complete elimination of the structure) did not have a substantial effect on the results.
At the same time, the hierarchical structure helps lead to transparency of model
results. Hierarchical aggregation of values remains based on geometric means and
is also implemented when the vulnerability-resilience index is calculated, since we
calculated sensitivities as positive numbers instead of negative ones.17
Addressing the second set of issues raised, we implemented a range method
of proxy indexing. Analogous to the HDI method, our range method standardizes
scales and prevents variables from being unintentionally weighed. Each proxy variable is scored on the same scale by calculating its score within the range of values
of the specific proxy participating in the analysis. The potential bias of large ranges
and the impact of outliers are thus eliminated.
Logarithmic transformations of certain proxy values were used where deemed
necessary to better represent the gradation in differences in a proxy value. Certain
processes may be best explained by using a proxy after a log transform, since the
value of an additional increase in a variable may decrease as the variable gets larger.
Through a log transform of such a variable each additional increase in the indexed
proxy becomes equally meaningful.
Using both a log transform and the geometric mean to emphasize the lower
end of the scale may appear to be “double-counting.” They are, however, used
for two different purposes and accomplish two separate effects. The log transform
corrects the meaning of individual indicators (e.g. an increase in annual income
from $100 to $500 is more meaningful than an increase from $20,100 to $20,500).
The aggregate score emphasizes the lower end of the scale for all variables when
the geometric mean is used.
The effect of these changes in the VRIP2 methodology—scaling method, positive numbers for sensitivity values, and selected use of log transforms—is that
interpretation of proxies and indicator values remains transparent, that outliers are
appropriately taken care of, and that indicators representing high sensitivity are
more appropriately weighed. Locations (regions, countries, states, or smaller entities) with one or more highly vulnerable sectors will thus not score well when
CASE STUDY OF INDIA AND INDIAN STATES
95
TABLE VII
Comparisons between VRIP and VRIP2
Steps in the
hierarchy
Sensitivity index
Indexing
Log transforms
Determination of
the scaled proxy
value; alternatives
World indexing method (VRIP)
Range method (VRIP2)
geometric mean of proxies >
sector indices
geometric means of indices >
sensitivity or coping/adaptive
capacity
addition of negative sensitivity
and positive coping/adaptive
capacity > VRIP index
after calculation > negative value
based on global values
None
geometric mean of proxies > sector
indices
geometric means of indices > sensitivity
or coping/adaptive capacity
100 × P
Pworld
100 × Pworld
P
depend on its value
geometric mean of sensitivity and
coping/adaptive capacity indices >
VRIP2 index
kept as positive value
based on the range of values
income (per capita GDP)
sulfur deposition
water sensitivity
100 × (P − Pmin )
(Pmax − Pmin )
100 × (Pmax − P)
Shift +
(Pmax − Pmin )
Shift +
P is the country or state’s proxy or indicator value; Pworld is the value of that indicator for the entire
world; Pmin and Pmax are the minimum and maximum proxy values seen in any country in the world;
“Shift” is a shift value that must be used to avoid scores very close to 0. Results were found to be not
very sensitive to the shift value as long as it is larger than one.
TABLE VIII
India’s ranking in different indexing methodologies
Method
India ranking
VRIP world indexing
VRIP2 range method
101/103
74/103
comparisons are made at the country or state levels or among any other geopolitical
entities. Table VII summarizes the comparison in methodologies.
Table VIII illustrates the impact of the revised method on India’s ranking
among 103 countries around the world. The change in ranking is due to the change
in method that affects the ranking of all countries, including India. India’s relative
ranking improved because several countries rank lower in the VRIP2 methodology
compared to India; India’s indicator values remained relatively consistent in both
methods.
96
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
12. A Final Word
Our main purpose has been to develop a robust tool for use in vulnerability
assessment. Such a tool allows for comparative analysis and enables more in-depth
exploration of the qualitative conditions that contribute to the quantitative results.
Exercising any developing tool and critically examining the results are necessary
to improve such a tool. This study of India and Indian states demonstrates the value
of the VRIP methodology given that it provides insights into vulnerability at country
and state levels, enabling analysts to ask the next level of questions and explore
directions for specific policy options that may mitigate climate impacts.
The study also led to improvements in the methodology. The method, as
presently developed, integrates potential response of the natural system, the managed landscape and the socioeconomic system; and goes beyond F¨ussel and Klein’s
(in press) second-generation vulnerability assessment which describes the inclusion of the evaluation of impacts on goods and services. The model represents
a comprehensive framework that can be elaborated on or reduced when used for
specific purposes and provides for a means to go beyond evaluating “exposure to
hazard” while retaining Cutter’s (1996) argument that location with all its specifics
of geography and climate forms the fundamental unit of analysis for vulnerability.
Because we have a robust framework and methodology to evaluate current vulnerability and resilience, we have a basis on which to begin developing credible
scenarios and analyses for projections of future vulnerability to climate changes.
Such scenarios must integrate information about current climate dependencies and
socioeconomic well-being, accounting for geographic differences, then project this
integrated and differentiated information into the future. It may be precisely in this
area that the most progress could be made in combining the more quantitative and
more qualitative approaches to vulnerability.
Acknowledgements
Many thanks to Joel Smith for giving us the opportunity to perform this analysis
of India’s vulnerability-resilience to climate change. We also appreciate very much
the material contributions of Sangamitra Ramachander in gathering the data necessary for the modeling exercise described in this report. Ramachander, a graduate
student in economics at the University of Maryland, assisted in locating, preparing, and interpreting available datasets. Greg Brinkman, graduate student in public
affairs at the University of Maryland helped us nail down the VRIP2 methodology
and his systematic approach to pursuing answers to questions greatly enriched the
understanding of the impact of methodology. Richard Moss initiated the quest for
quantification of vulnerability to climate change. Working with him on the initial
formulation of the VRIP inspired us to pursue the work at the Indian states level.
We thank both Richard and Greg very much. Finally, the thoughtful comments and
critiques of the reviewers helped us improve the analysis in important ways.
CASE STUDY OF INDIA AND INDIAN STATES
97
Appendix
Coping/
adaptive
capacity (+)
Sector
Proxies
Sources
Economic
capacity
GDP per capita
http://economywatch.com/
database/income7.htm & http://
www.indiainfoline.com/econ/andb/
nia/nia3.html
http://www.wws.princeton.edu/
∼rpds/Downloads/povertyinequality-india-july8-2002.pdf.18
Deininger and Squire 1996, 1998
& www.worldbank.org/research/
growth/dddeisqu.htm
www.indiastat.com: Table: Statewise
workforce participation rate by
rural/urban (93-94)
http://www.indiainfoline.com/
http://www.cs.colostate.edu/
∼malaiya/india.html#States
http://agricoop.nic.in/statistics/
st3.htm
GEIA
http://weather.engin.umich.edu/
geia/ & CDIAC http://
cdiac.esd.ornl.gov/home.html
http://www.economywatch.com/
database/population4.htm &
http://www.cs.colostate.edu/
malaiya/india.html & http://
www.cyberjournalist.org.in/
census/cenden0.html
Climate Change in Asia:
http://www.ccasia.teri.res.in/
country/india/impacts/tables.htm &
http://www.envfor.nic.in/cc/
adaptation/coastal.htm & Delft
Hydraulics, 1993
http://www.rainwaterharvesting.org/
resources/statistics/stat531.htm
Income distribution
equity
Human and civic
resources
Environmental
capacity
Dependence ratio
(calculated from
% population in
the workforce)
Illiteracy (calculated
form literacy data)
% Non-managed
land
SO2 emissions
(calculated)
Population density
Sensitivity(−)
Settlement/
Infrastructure
Sea level rise
resulting in
number of people
at risk
% Population with
access to safe
water
% Population with
access to
sanitation
www.indiastat.com: Table: State wise
Per Capita (Public Sector)
Expenditure on Health (Medical
and Public Health) including Water
Supply and Sanitation and Family
Welfare (During 1985–1986 to
1989–1990).
(Continued on next page)
98
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
(Continued)
Sector
Proxies
Sources
Food security
Cereal production/
agricultural land
http://agricoop.nic.in/statistics/ stpart3d.htm
for food crops &
http://www.andhrapradesh.com/
apwebsite/tables/statisticsstate/table-4.htm
for production and areas & FAOSTAT98
http://www.unu.edu/unupress/food/
8F173e/8F173E08.htm & FAOSTAT98
Human health
Total protein demand
per capita (animal +
vegetable + pulses)
Fertility rate
Life expectancy
Ecosystems
% Managed land
Water resources
Fertilizer use/area
cropland
Water sensitivity,
based on availability
and consumption
Additional
information
http://www.censusindia.net/srs1.html
http://www3.who.int/whosis/hale/
hale.cfm?path=whosis,hale&
language=english
http://www3.who.int/whosis/hale/
hale.cfm?path=whosis,hale&
language=english
http:/agricoop.nic.in/statistics/
consum4a.htm
www.indiastat.com: Table: Statewise
groundwater resources and irrigation
potential
http://www.grida.no/climate/ipcc/
regional/293.htm: for annual renewable
World Bank, 1998
Notes
1
“Delhi’s CNG success inspiring many countries: Naik”, December 11, 2002. OutlookIndia.com.
Available at: http://www.outlookindia.com/pti 11 news.asp?id=103516.
2
http://www.bmtpc.org/disaster.htm.
3
data from: www.worldbank.org/research/growth/dddeisqu.htm.
4
See http://www3.who.int/whosis/hale/hale.cfm?path=whosis,hale&language=english.
5
Net Domestic Product of States at current prices 1980–1981 to 1997–1998 (Rs.
bn): http://www.economywatch.com/database/income7.htm population estimates: http://www.cs.
colostate.edu/∼malaiya/india.html#States: Literacy & Population by Religion.
6
http://www.indiainfoline.com/econ/ andb/nia/nia10.html).
7
Example from the Press Information Bureau, Government of India, 2001, http://pib.nic.
in/archieve/ppinti/ppioct2001/agriculture.html.
8
Example from the Press Information Bureau, Government of India, 2001,
http://pib.nic.in/archieve/ppinti/ppioct2001/agriculture.html.
9
Cited in Sofia Checa, Inequality in Three Indian States, unpublished paper, 2001.
10
Calculated from data in the Annual Report 1999–2000, Ministry of Water Resources, Govt. of
India.
11
www.indiastat.com.
CASE STUDY OF INDIA AND INDIAN STATES
99
12
http://www.e-greenstar.com/India/Andhra-Pradesh-info.htm. Information excerpted from the Microsoft Encarta Encyclopedia 2000, http://www.encarta.msn.com/products/.
13
http://www.cnie.org/pop/pai/water-30.html.
14
http://www.grida.no/climate/ipcc/regional/293.htm.
15
“Water in India: nor any drop to drink.” August 24, 2002. The Economist, volume 364, issue
8287: 31–32.
16
http://www.andhrapradesh.com&sol.
17
For coping and adaptive capacity the geometric mean responds well to its mathematical
consequences with high proxy or sub-index values weighing less than low values, resulting in a
conservative estimate. The same does not hold for sensitivity, however, which in the VRIP method
is based on large negative values representing high sensitivities and low negative values representing
sectors of low sensitivity. The impact of using high proxy values for highly sensitive sectors may
have underestimated the final vulnerability-resilience index value given that a geometric mean
weighs large values less than small values. Arithmetic means are not an answer, however, given
that they imply perfect substitution between sectors (see also Sagar and Najam 1998), which is
contrary to common sense; deficiency in one sector cannot always be offset by another sector. The
arithmetic mean and the geometric mean are two ways to average a set of numbers. Arithmetic mean:
(X 1 + X 2 + X 3 + · · · + X n)/n. The arithmetic mean is an additive method since sector values
are added to determine the total, which is divided by the number of sectors. Geometric mean: (X 1
× X 2 × X 3 × · · · × X n)(1/n). The geometric mean is a multiplicative or product method because
sector values are multiplied by each other to determine their product, which is taken to the 1/n
power.
18
Poverty and Inequality in India: A Reexamination, by Angus Deaton and Jean Dreze, Research
Program in Development Studies, Woodrow Wilson School, Princeton University and Delhi School
of Economics. (July, 2002). Table 5 — 55th Round adjusted values. (Corresponding to the National
Sample Survey: 1999–2000).
References
Aandahl, G. and Karen, O’Brien: 2001, ‘Vulnerability to climate changes and economic changes
in Indian agriculture’, Paper presented at Waters of Hope. The Role of Water in South Asian
Development, Biannual Conference of the Nordic Association for South Asian Studies, 20–22
September 2001.
Acosta Moreno, A., Skea, J., Gacuhi, A., Greene, D. L., Moomaw, W., Okita, T., Riedacker, A.,
and Tran Viet, L.: 1995, ‘Industry, energy, and transportation: Impacts and adaptation’, Climate
Change, op.cit.
Adger, W. N., Khan, S. R., and Brooks, N.: 2003, Measuring and enhancing adaptive capacity, Technical Paper 7, UNDP Adaptation Policy Framework, http://www.undp.org/cc/apf outline.htm.
Agarwal, A. and Narain, S.: 1991, Global Warming in an Unequal World, Centre for Science and
Environment, New Delhi.
Agarwal, A. and Narain, S.: 1992, Towards a Green World: Should Global Environmental Management be Built on Legal Conventions or Human Rights? Centre for Science and Environment,
New Delhi.
Ahmad, Q. K., Warrick, R. A., Downing, T. E., Nishioka, S., Parikh, K. S., Parmesan, C., Schneider,
S. H., Toth, F., and Yohe, G.: ‘Methods and tools’, in McCarthy, J. J., Canziani, O. F., Leary,
N. A., Dokken, D. J., and White, K. S. (eds.), Climate Change 2001: Impacts, Adaptation and
Vulnerability, Cambridge University Press, Cambridge.
Allen-Diaz, B., Chapin, F. S., Diaz, S., Howden, M., Puigdef´abregas, J., and Smith, M. S.: 1995,
‘Rangelands in a changing climate: Impacts, adaptations, and mitigation’, Climate Change, op.
100
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
cit.; Everett, J. T., Krovnin, A., Lluch-Belda, D., Okemwa, E., Regier, H. A., and Troadec, J.-P.:
1995, ‘Fisheries’, Climate Change, op. cit.
Bohle, H. G., Downing, T. E., and Watts M. J.: 1994, ‘Climate change and vulnerability: Toward a
sociology and geography of food insecurity’, Global Environ. Change 4, 37–48.
Brenkert, A. L. and Malone, E. L.: Vulnerability and Resilience of India and Indian States to Climate
Change: A First-Order Approximation, Washington, DC, Joint Global Change Research Institute,
http://www.pnl.gov/globalchange/projects/vul/index.htm.
Cannon, T.: 1994, ‘Vulnerability analysis and the explanation of natural disasters’, in Varley, A. (ed.),
Disasters, Development and Environment, Wiley, Chichester, pp. 13–30.
Chambers, R.: 1989, ‘Editorial introduction: Vulnerability, coping and policy’, IDS Bulletin 21, 1–7.
Chandler, W., Schaeffer, R., Dadi, Z., Shukla, P. R., Tudela, F., Davidson, O., and Alpan-Atamer, S.:
2002, Climate Change Mitigation in Developing Countries: Brazil, China, India, Mexico, South
Africa, and Turkey, Pew Center on Global Climate Change, Washington, DC.
CIA World Factbook: July 2001, http://www.odci.gov/cia/publications/factbook/geos/in.html.
Cohen, S., Demeritt, D., Robinson, J. and Rothman, D.: 1998, ‘Climate change and sustainable
development: Towards dialogue’, Global Environ. Change 8(4), 341–371.
Comfort, L., Wisner, B., Cutter, S., Pulwarty, R., Hewitt, K., Oliver-Smith, A., Wiener, J., Fordham,
M., Peacock W., and Krimgold, F.: 1999, ‘Reframing disaster policy: the global evolution of
vulnerable communities’, Environ. Hazards 1, 39–44.
Cutter, S. L.: 1996, ‘Vulnerability to environmental hazards’, Prog. Hum. Geogr. 20, 529–539.
Deininger, K. and Squire, L.: 1996, ‘A new data set measuring income inequality’, World Bank Econ.
Rev. 10(3) 565–591.
Deininger, K. and Squire, L.: 1998, ‘New ways of looking at old issues: Inequality and growth’, J.
Dev. Econ. 57, 259–287.
Delft Hydraulics, The Netherlands: 1993, Sea Level Rise: A Global Vulnerability Assessment. Tidal
Water Division, Rijkswaterstaat, Ministry of Transport, Public Works and Water Management,
the Netherlands, Table B-3.
Dow, K.: 1992, ‘Exploring differences in our common future(s): The meaning of vulnerability to
global environmental change’, Geoforum 23, 417–436.
Downing, T. E.: 1991, ‘Vulnerability to hunger in Africa: A climate change perspective’, Global
Environ. Change 1, 365–380.
Downing, T. E. and Patwardhan, A.: 2003, ‘Vulnerability assessment for climate change adaptation’, Technical Paper 3, Adaptation Policy Framework, United Nations Development Program.
http://www.undp.org/cc/apf outline.htm.
Edmonds, J., Wise, M., Pitcher, H., Richels, R., Wigley, T., and MacCracken, C.: 1997, ‘An integrated
assessment of climate change and the accelerated introduction of advanced energy technologies:
an application of MiniCAM 1.0’, Mitigation and Adaptation Strategies for Global Change 1,
311–339.
Eischen, K.: 2003, ‘Andhra Pradesh: Lessons for global software development’, Computer 36(6),
31–37.
FAO [Food and Agriculture Organization]: 1999, FAOSTAT98 [CD-ROM], Sales and Marketing
Group, FAO, Rome.
Freedom House: 2003, Freedom in the World: The Annual Survey of Political Rights and Civil Liberties
2003, Freedom House, http://www.freedomhouse.org/research/freeworld/2000/index.htm.
F¨ussel, H.-M. and Klein, R. J. T.: in press, ‘Climate change vulnerability assessments: An evolution
of conceptual thinking’, Clim. Change.
Gannon, M. J., and Associates: 1994, Understanding Global Cultures, Sage, Thousand Oaks, California.
Kasperson, J. X., Kasperson, R. E. and Turner II, B. L. (eds.): 1995, Regions at risk: International
Comparisons of Threatened Environments, UNU Press, Tokyo.
CASE STUDY OF INDIA AND INDIAN STATES
101
Kelly, P. M. and Adger, W. N.: 2000, ‘Theory and practice in assessing vulnerability to climate change
and facilitating adaptation’, Clim. Change 47, 325–352.
Lal, M.: 2001, ‘Global climate change-India’s monsoon and its variability: Vulnerability and adaptation issues’, Report on Country Studies Vulnerability and Adaptation, Work Assignment
402 – Task 11 under Stratus Consulting Contract 68-W6-0055, Washington, DC, U.S. Environmental Protection Agency.
Lal, M., Cubasch, U., Voss, R., and Waszkewitz, J.: 1995, ‘Effect of transient increases in greenhouse
gases and sulphate aerosols on monsoon climate’, Curr. Sci. 69(9), 752–763.
Moss, R. H., Brenkert, A. L., and Malone, E. L.: 2001, Vulnerability to Climate Change: A Quantitative Approach, PNNL-SA-33642, Pacific Northwest National Laboratory, Washington, DC,
http://www.pnl.gov/globalchange/projects/vul/index.htm.
Nakicenovic, N., et al.: 2000, Special Report on Emissions Scenarios, Cambridge, Cambridge University Press.
Parikh, J. K. and Parikh, K.: 2002, Climate Change: India’s Perceptions, Positions, Policies and
Possibilities, Organization for Economic Cooperation and Development, Paris.
Prescott-Allen, R.: 2001, The Wellbeing of Nations: A Country-by-Country Index of Quality of Life
and the Environment, Washington, DC, Island Press.
Reilly, J., et al.: 1995, ‘Agriculture in a changing climate: Impacts and Adaptation’, Climate Change,
op. cit.
Ribot, J. C.: 1996, ‘Introduction’, in Ribot, J. C., Magalh˜aes, A. R., and Panagides, S. S. (eds.),
Climate Variability, Climate Change and Social Vulnerability in the Semi-Arid Tropics, Cambridge
University Press, Cambridge.
Ribot, J. C., Magalh˜aes, A. R., and Panagides, S. S. (eds): 1996, Climate Variability, Climate Change and Social Vulnerability in the Semi-Arid Tropics, Cambridge University Press,
Cambridge.
Robinson, J. and Herbert, D.: 2000, ‘Integrating climate change and sustainable development’, in
Munasinghe, M. and Swart, R. (eds.), Climate Change and Its Linkages with Development, Equity,
and Sustainability, LIFE, Columbo, Sri Lanka; RIVM, Bilthoven, Netherlands; and World Bank,
Washington, DC, pp. 143–162.
Sagar, A. D. and Najam, A.: 1998, ‘The human development index: A critical review’, Ecol. Econ.
25, 249–264.
Scott, M. J., Aguilar, A. G., Douglas, I., Epstein, P. R., Liverman, D., Mailu, G. M., and Shove, E.:
1995, ‘Human Settlements in a change climate: Impacts and adaptation’, Climate Change, op.
cit.
Smith, J. B., Schellnhuber, H. J. and Monirul Qader Mirza, M.: 2001, ‘Vulnerability to climate change
and reasons for concern: A synthesis’, in McCarthy, J. J., Canziani, O. F., Leary, N. A., Dokken,
D. J., and White, K. S. (eds.), Climate Change 2001: Impacts, Adaptation and Vulnerability,
Cambridge University Press, Cambridge.
Stern, P. C. and Easterling, W. E. (eds.): 1999, Making Climate Forecasts Matter: Panel on the
Human Dimensions of Seasonal-to-Interannual Climate Variability, National Academy Press,
Washington, DC.
Tata Energy Research Institute, http://www.ccasia.teri.res.in/country/india/impacts/impacts.htm.
Vogel, C.: 1997. ‘Vulnerability and global environmental change’, LUCC Newsletter 3, http://www.
lucc.vogel.vuln.review.article.html.
“Water in India: nor any drop to drink.” August 24, 2002. The Economist 364(8287), 31–32.
Watson, R., Zinyowera, M. C., Moss, R. H., and Dokken, D. J.: 1998, The Regional Impacts of Climate
Change, an Assessment of Vulnerability. A special report of IPCC working group II, Cambridge
University Press, Cambridge.
Watts, M. J. and Bohle, H. G.: 1993, ‘The space of vulnerability: The causal structure of hunger and
famine’, Prog. Hum. Geogr. 17, 43–67.
102
ANTOINETTE L. BRENKERT AND ELIZABETH L. MALONE
World Bank: 1998, World Development Indicators 1998 [CD-ROM], International Bank for Reconstruction and Development/The World Bank, Washington, DC.
World Bank: 2002, Country Brief: South Asia Region (SAR)–India, http://www.Worldbank.
org.
World Resources 1994–1995: 1996, A guide to the global environment. A report by The World
Resources Institute in collaboration with The United Nations Environment Programme and The
United Nations Development Programme, Oxford University Press.
[YCELP and CIESIN]: 2002, Yale Center for Environmental Law and Policy and Center for International Earth Science Information Network, Environmental Sustainability Index: An Initiative of
the Global Leaders of Tomorrow Environment Task Force, World Economic Forum, New Haven,
CT, Yale Center for Environmental Law and Policy.
(Received 31 March 2003; in revised form 22 July 2004)