Climate Change Impacts, Vulnerability and Adaptation Developing Country Perspective

1
Climate Change Impacts, Vulnerability and
Adaptation Developing Country Perspective

• K.S. Kavi Kumar
• Madras School of Economics, Chennai
• EMF Workshop on
• Critical Issues in Climate Change
• Snowmass, July 2005

2
Structure of Presentation

• Methodologies for Impact Assessment
• Applicability for Developing Countries
• Examples
• Moving towards Vulnerability Assessment
• Vulnerability Metric
• Indicator Based Approaches
• Other Approaches
• Moving towards Adaptation Assessment
• Conclusions

3
Impact Assessment - Agriculture
4
Agronomic-Economic Approach

• Agronomic models are used first to predict
  climate change impacts on crop yields and the
  estimated yield changes are then introduced into
  economic models to predict output and price
  changes.
• Feasible to model CO2 fertilization effects.
• Relatively difficult to include all possible farm
  level adaptation options.
• Adams et al. (1990, 1999), Rosenzweig and Paryy
  (1994), Kumar and Parikh (2001b)

5
Agroecological Zone Approach

• Assigns crops to agroecological zones and
  estimates potential crop yields. As climate
  changes, the extent of agroecological zones and
  the potential yields of crops assigned to those
  zones changes. These acreage and yield changes
  are then included in economic models to assess
  socio-economic impacts.
• Darwin et al. (1995, 2000), Kumar (1998), IIASA
  (2002)

6
Ricardian Approach

• Similar to Hedonic pricing approach of
  environmental valuation. The approach is based
  on the argument that, by examining two
  agricultural areas that are similar in all
  respects except that one has a climate on average
  (say) 3oC warmer than the other, one would be
  able to infer the willingness to pay in
  agriculture to avoid a 3oC temperature rise.
• Uses statistical analysis of data across
  geographic areas to separate climate from other
  factors (such as soil quality) that explain
  production differences across regions and uses
  the estimated statistical relationships to assess
  impacts of climate change.
• Main advantage of this method is that it
  incorporates all private adaptation measures.
• Assumes that relative prices do not change and
  hence biases the results.
• Not feasible to incorporate CO2 fertilization
  effects.
• Mendelsohn et al. (1994), Dinar et al. (1998),
  Kumar and Parikh (2001a)

7
Socio-economic Impacts of Climate Change
Source Kumar and Parikh (2001b)
8
Ricardian Model Specification

• The model is specified as follows
• R is net-revenue per hectare
• T and P are normal temperature and precipitation
  in level and square terms
• YVarT and DVarT represent the yearly and diurnal
  climate variation terms
• K represents the control variables such as soil
  characteristics, literacy, population density etc
• Analysis is carried out using pooled
  cross-sectional, time-series data of 271
  districts spread across India.

9
Impact Estimates with Climate Variation
An F-test comparing the models with and without
the climate variation terms showed that the
climate variation variables together are
significantly different from zero. The
t-statistic showed that barring a few all the
climate variation variables are significant in
improving the model specification.
Source Kumar (2003)
10
What are the lessons?

• In terms of wide-spread applicability the
  cross-sectional analysis appears promising
• However data constraints may limit its use in
  many developing countries
• Recent World Bank study addressed some of these
  issues

11
Cross-Sectional Analyses

• Can individual farm data be used?
• Agricultural census data not easily available in
  many developing countries
• Study on Sri Lanka agriculture (Kurukulasuriya
  and Ajwad, 2004) showed that farm level data
  collected through survey can be a good
  substitute
• Sri Lankan study also highlighted crucial role of
  (monsoon) precipitation along with temperature

12
Cross-Sectional Analyses

• What can be done in the absence of meteorological
  data from weather stations?
• Developing countries often have sparsely spread
  out weather stations giving little scope for
  meaningful interpolation
• Mendelsohn et al. (2004a) illustrate that
  satellite measures of climate (surface
  temperature and soil moisture as a proxy for
  precipitation) can explain more of the observed
  variation in farm performance than ground station
  data
• Ground station data would still be preferred
  option when precipitation is important

13
Cross-Sectional Analyses

• Is climate variation important?
• In developing countries it is often relatively
  easy to get climate long-run average data,
  compared to weather yearly data
• Mendelsohn et al. (2004b) show that climate
  normals explain large portion of variation in
  farm performance, but climate variance terms are
  also important
• Kumar (2003) also highlight the importance of
  inclusion of variance terms in Indian context

14
Cross-Sectional Analyses

• Other Potential Uses
• Rosenberg et al. (2000) apply the cross-sectional
  analysis to study climate change impacts on
  health (morbidity and infant mortality) in
  Brazil. Primarily the study demonstrates the
  usefulness of cross-sectional analyses in
  assessing health impacts
• Mendeloshn et al. 2004c) extend the notion of
  close dependence of rural incomes on agriculture
  to analyse impact of climate change on rural
  income using cross-sectional analysis and show
  that the method could be used for assessing
  impacts on rural income
• Other applications include amenity value of
  climate etc.

15
Impacts or Vulnerability

• Sector-wise impact estimations in developing
  countries while being important may require
  significant resources and may not provide
  ground-level practical suggestions on adaptation
  process
• Motivation for impact assessment partly comes for
  justifying climate change mitigation policies
• However it may be argued that such motivation has
  outlived its purpose
• Vulnerability assessment helps in understanding
  the adaptation process

16
Impacts or Vulnerability

• Emphasis on vulnerability marks a shift away from
  traditional assessments, which limit analysis to
  the stressors (e.g., climate change) and the
  corresponding impacts, towards an examination of
  the system being stressed and its ability to
  respond
• By focusing on the mechanism that facilitates or
  constrain a systems ability to cope, adapt or
  recover from various disturbing forces,
  vulnerability assessments help in not only
  identifying who, but also why
• Such information is critical in prioritizing
  limited resources for most vulnerable and also
  for designing most effective vulnerability-reduc
  ing interventions

17
Vulnerability Characterization

• Three primitives must be identified for
  characterizing vulnerability appropriately
• The entity that is vulnerable - e.g., rice
  farmers
• The stimulus causing vulnerability e.g.,
  pressures such as climate change and
  globalization
• The (welfare) criteria with reference to which
  the entitys vulnerability is defined and on
  which preference order can be specified e.g.,
  break-even farm level yield level or minimum
  consumption level (poverty criteria)

18
Vulnerability Index

• The index presents a single-value measure of
  vulnerability based on meaningful criteria, which
  can be used when taking decisions regarding the
  allocation of financial and technical assistance.
• Basic methods for computing a vulnerability
  index
• Normalization procedure
• Mapping on a categorical scale
• Regression method
• Limitations of Index
• Subjective choice of variables
• Measurement problems
• Weighting
• Kumar and Tholkappian (2005), OBrien et al.
  (2004), Acosta-Michilik et al. (2004), Brenkert
  Malone (2004)

19
Coastal Vulnerability Index for India

• Index is constructed taking both climatic and
  non-climatic stresses into consideration, and
  focusing on sensitivity and adaptive capacity of
  units of analysis (namely, districts).
• Demographic (a) Population density (2001) (b)
  Annual growth rate of population (c) Population
  at risk due to sea level rise.
• Physical (a) Coast length (b) Insularity
  (defined as ratio of coastal length to the area
  of the district) (c) Frequency of cyclones
  (weighted to account for cyclones of different
  intensities) based on historic data (d) Probable
  maximum surge height (e) Area at risk of
  inundation due to sea level rise (f) Vulnerable
  houses both at the risk of damage and collapse
  (1991 census).
• Economic (a) Agricultural dependency (expressed
  in terms of population dependent on agriculture
  and other primary sectors) (b) Income and/or
  Infrastructure index.
• Social (a) Literacy (b) Spread of institutional
  set up.

20
Coastal Vulnerability Index
Source Kumar and Tholkappian (2005)
21
Vulnerability Resilience to Climate Change
Indian States

• Brenkert Malone (2004) applied VRIP methodology
  to assess vulnerability of Indian states
• Wide variety of sources of vulnerability across
  states
• Kerala and Sikkim are more sensitive than Punjab
  on food-security front
• Punjab is more sensitive on ecosystem front due
  to excessive resource use (fertilizers/pesticides)
  
• While social policies may be more effective in
  reducing sensitivity, policies aimed at
  environmental protection could be helpful in
  increasing coping capacity

22
Vulnerability Index Fuzzy Approach

• Viswanathan Kumar (2005) assessed vulnerability
  of Indian states based on IPCC conceptualization
  and constructing index using fuzzy logic
• Acosta-Michlik et al. (2004) also followed
  similar procedure for characterizing
  vulnerability of three countries (India,
  Portugal, Russia) over a period of 20 years
• Index is constructed based on fuzzy logic to make
  quantitative inference from linguistic statements

23
Why Fuzzy?

• Fuzzy sets allow for gradual transition from one
  state to another while also allowing one to
  incorporate rules and goals, and hence are more
  suitable for modeling preferences and outcome
  that are ambiguous
• While the use of two-valued logic would be
  limited to determining only whether vulnerability
  exists or not, a multi-valued logic can be use to
  assess the degree of vulnerability
• That is, it is also possible to attach linguistic
  values such as low, moderate, and high to
  certain index value ranges

24
How Fuzzy?

• Cerioli Zani (1990), Cheli Lemmi (1995),
  Qizilbash (2001) all use fuzzy set theory based
  notions in poverty and vulnerability analysis,
  typically focusing on fuzzification alone
• It may be noted that their notion of
  vulnerability is not the mainstream understanding
  of vulnerability in Economics
• Vulnerability in these exercises relates to the
  possibility of being classified as poor, rather
  than risk of becoming poor
• Like in many engineering and science applications
  of Fuzzy set theory, study by Viswanahan Kumar
  (2005) focuses on fuzzification, fuzzy inference
  and defuzzification for assessing vulnerability

25
Framework for Vulnerability Analysis
26
Moving Beyond Indices
Source Luers et al. (2003)
27
Moving Beyond Indices (contd.)

• Luers et al. (2003) in their study on Yaqui
  Valley, Mexico define vulnerability a function of
  ratio of sensitivity (of well being) to the
  proximity of well being to the threshold level of
  well being, and exposure of the system (captured
  through expected value)
• Taking cue from poverty dynamics literature
  (e.g., Chaudhuri et al., 2002) one may further
  refine and bring-in probabilistic notion in
  vulnerability characterization

28
Moving Towards Adaptation Assessment

• End-point characterization of vulnerability in
  climate change literature emphasizes on
  regional/national level adaptation strategies
• In contrast vulnerability assessments practiced
  by poverty and disaster management communities
  depend directly on the vulnerable community
  itself to make use of wider-range of social,
  cultural, economic and institutional factors and
  also characterize vulnerability as
  starting-point of their analysis
• These aspects make vulnerability assessment
  conducive for providing local-scale guidance on
  adaptation

29
Moving Towards Adaptation Assessment

• As Klein (2004) argues even the most recent
  sophisticated scenario-based assessments of
  impacts and vulnerabilities (e.g., DINAS-COAST)
  may increase awareness for adaptation but give
  little information to the local decision makers
  on most efficient or effective adaptation
  strategies
• Such information may come only from local
  knowledge and one needs different tools/methods
  to comprehend the same

30
Intervulnerability Use of ABMs

• Focus on adaptive capacity of agricultural
  farmers and its link with poverty in the context
  of composite pressure of globalization and
  climate change and policies for enhancing the
  same.
• Adaptation strategies geared to cope with large
  climate anomalies are assumed to embrace a large
  proportion of the envelope of adjustments
  expected under long-term climate change. In a
  similar vein adaptation strategies addressing
  current day market fluctuations could help in
  coping with globalization.
• Thus the focus would be on current day observed
  strategies (along with potential new strategies)
  and assessing their effectiveness with and
  without the consideration of social interaction
  among the agents

31
Modeling Vulnerability

• Vulnerability is a function of exposure,
  sensitivity, and adaptive capacity plus entitys
  cognition
• Cognition allows the agents to receive and
  exchange information, to perceive and evaluate
  risks, to identify and weight options, to make
  decisions and perform actions, and modify and
  update profile based on the outcomes
• Various cognitive strategies include
  deliberation (maximization), repetition both of
  which deal with low uncertain situations
  comparison and imitation both of which deal
  with high uncertain situations (Jager et al.,
  2000)

32
Modeling Vulnerability

• Social interactions of the agents could be based
  on social psychological consistency principle
  i.e. individuals tend to agree most with those
  whom they like the best and tend to like best
  those with whom they agree the most
• Moss et al. (2001) followed such approach in
  their modeling of water demand in Thames region
• Based on common finding in social psychology that
  there exists a strong correlation between shared
  attitudes and attractiveness

33
(No Transcript)
34
Developing Adaptation Efforts

• Effective adaptation strategies require
  understanding of regional / local dimensions of
  vulnerability
• Climate change does not occur in isolation
  multiple stresses
• Domestic policies can enhance or constrain
  farmers ability to adapt to climate change

35

• Thank You for Your Attention!