Targeting attention on local vulnerabilities using an integrated … · 2011-02-28 · Water...

Water Science and Technology 51:5 Sullivan and Meigh, 2005 69

Targeting attention on local vulnerabilities using an integrated index approach: the example of the Climate Vulnerability Index

CAROLINE SULLIVAN and JEREMY MEIGH

Centre for Ecology & Hydrology, Wallingford, Oxfordshire OX10 8BB, United Kingdom Email: [email protected]

Abstract. It is known that climate impacts can have significant effects on the environment, societies and economies. For human populations, climate change impacts can be devastating, giving rise to economic disruption and mass migration as agricultural systems fail, either through drought or floods. Such events impact significantly, not only where they happen, but also in the neighbouring areas. Vulnerability to the impacts of climate change needs to be assessed, so that adaptation strategies can be developed and populations can be protected. In this paper, we address the issue of vulnerability assessment through the use of an indicator approach, the Climate Vulnerability Index (CVI). We show how this can overcome some of the difficulties of incommensurability associated with the combination of different types of data, and how the approach can be applied at a variety of scales. Through the development of nested index values, more reliable and robust coverage of large areas can be achieved, and we provide an indication of how this could be done. While further work is required to improve the methodology through wider application and component refinement, it seems likely that this approach will have useful application in the assessment of climate vulnerability. Through its application at sub-national and community scales, the CVI can help to identify those human populations most at risk from climate change impacts, and as a result, resources can be targeted towards those most in need. Key words: climate change, indices, vulnerability, water resources. Introduction The importance of water for people’s livelihoods is now widely recognised. In developing countries the time spent carrying domestic water, health problems related to water, the availability of and access to water for productive purposes and small-scale agriculture are all of great importance, especially to poor people. Because of these issues, poor people in developing countries tend to be vulnerable to short-term shocks as well as the longer term variation being brought about by climate change. In addition, other global changes due to increasing population densities, greater competition for resources, environmental degradation and loss of biodiversity are contributing to the degree of vulnerability faced by millions of people. In order to develop strategies for adaptation to these global changes, it is necessary as a first step to make reliable and consistent assessments of vulnerability. This needs to be holistic, at the appropriate scale, and to integrate a wide range of relevant factors. In this paper we discuss one approach to this problem, the Climate Vulnerability Index (CVI), an index-based approach which can encapsulate (into a single number) human vulnerability at any location. It includes a representative range of social and physical factors, and it can be calculated at a range of spatial scales. The CVI focuses mainly on water-related issues, but it is also to some degree relevant to vulnerability more generally. The objective of this paper is to demonstrate the potential for making vulnerability assessments using the CVI at a scale which is appropriate to targeting vulnerability of local populations. At the same time, we demonstrate that the approach is practicable for application over large areas. This means that while whole countries or regions can be


assessed, the significant local variation within countries can also be revealed. We discuss ways in which this can be achieved, and show that generally, sufficient data are likely to be available at the appropriate resolutions, and that the merging of social, economic, environmental and other physical data is a realistic prospect. Vulnerability and its assessment Interest in the concept of vulnerability has arisen from the need to address the increasing numbers of people who have been affected by natural or man-made disasters. Vulnerability was a concept used first by engineers to assess risk of building collapse and other problems, but today the concept is much more widely used, and can be taken to represent ‘A set of conditions and processes resulting from physical, social, economical and environmental factors which increase the susceptibility of a community to the impact of hazards’ (ISDR, 2002, p24). Taking this definition, any attempt to develop an appropriate framework for vulnerability assessment needs to address these diverse issues. While huge volumes of relevant data are available from many sources, the sorting, analysing and targeting of the information is extremely important if adaptation and mitigation is to be achieved. There is a need for the standardisation of data collection and storage, and the development of systematic methodological approaches, so that an accurate, reliable and politically neutral vulnerability assessment can be made. Such assessments are needed to address the problems arising from the increasing likelihood that human vulnerability is increasing as a result of climate change (IPCC, 2001; Lenton, 2002; Adger and Brooks, 2002). Indicators as policy tools To capture the essence of this definition of vulnerability, a composite index approach is proposed. This can explicitly incorporate indicators which represent the diverse dimensions of what makes a population vulnerable. Indicators have been used for policy making since the 1920s, (Fisher, 1922; Edgeworth, 1925). They are a statistical concept, providing an indirect way of measuring a given quantity or state, and allowing for comparison over time. The use of indicators has been recognised as important in attempts to understand the complexities associated with environmental monitoring and management (Hammond et al., 1995; Streeten, 1995; Rennings and Wiggering, 1997; Rogers et al., 1997; World Bank, 1998). An example of the index approach applied to water resource assessment is provided by the Water Poverty Index (WPI), outlined in Box 1.

Box 1. A composite index: the Water Poverty Index The Water Poverty Index was developed in a participatory manner, following consultation with a wide range of stakeholders, policy-makers and scientists. Calculation of the WPI involves combining data which represent five key components. These are: • Resource The physical availability of surface and ground water, including variability and

quality of the resource as well as the total amount of water. • Access The extent of access to water for human use, including access to irrigation. • Capacity The effectiveness of people’s ability to manage water. • Use The efficiency of how water is used for domestic, agricultural and industrial

purposes. • Environment A measure of how human water use impacts on ecological integrity. Each component is made up of a number of sub-components, and there may be a weighting applied to indicate the importance of each variable. To avoid problems of subjectivity, a baseline value of the WPI should be first calculated with the weightings set equally. Each component is standardised to fall in the range 0 to 100; giving a final WPI value also between 0 and 100. The highest value, 100, is taken to be the best situation, that is the lowest possible level of water poverty, while 0 is the worst. Source: Sullivan et al. (2003).


Methodology to derive CVI values As with any indicator, it is essential to select appropriate variables for inclusion in the CVI framework. Clearly there is a need to represent the specific geographical attributes associated with any location, and for this purpose, we use a Geospatial component. This is made up of a suite of variables that serve to indicate some measure of the degree of risk arising as a result of the physical or geographical nature of the selected location. Examples of the types of variables that might be considered for inclusion in this category could include length of coastline, altitude, distance from nearest neighbour, and landscape topography. Considerable debate and consultation has been devoted to the identification of the variables used in the calculation of the WPI (see Box 1). Each of the major WPI components (Resource, Access, Capacity, Use and Environment) are highly relevant to human vulnerability, and those variables that are particularly relevant to climate impacts would be most valuable for inclusion in the CVI. A selection of possible variables for inclusion in a CVI assessment is shown in Table 1, but it is important to note that the number of variables used in any index influences its sensitivity, and that different variables are both available and relevant at different spatial scales. Table 1. Potential variables for inclusion as sub-components of the CVI

CVI component Sub-components / variables Resource (R) • assessment of surface water and groundwater availability

• evaluation of water storage capacity, and reliability of resources • assessment of water quality, and dependence on imported/desalinated water

Access (A) • access to clean water and sanitation • access to irrigation coverage adjusted by climate characteristics

Capacity (C ) • expenditure on consumer durables, or income • GDP as a proportion of GNP, and water investment as a % of total fixed capital

investment • educational level of the population, and the under-five mortality rate • existence of disaster warning systems, and strength of municipal institutions • percentage of people living in informal housing • access to a place of safety in the event of flooding or other disasters

Use (U) • domestic water consumption rate related to national or other standards • agricultural and industrial water use related to their respective contributions to GDP

Environment (E) • livestock and human population density • loss of habitats • flood frequency

Geospatial (G) • extent of land at risk from sea level rise, tidal waves, or land slips • degree of isolation from other water resources and/or food sources • deforestation, desertification and/or soil erosion rates • degree of land conversion from natural vegetation • deglaciation and risk of glacial lake outbursts

Combining components and sub-components of the CVI There are a number of different ways in which variables can be combined to form an index. Here we consider the composite index approach1, as used in the construction of the Human Development Index. The value of the CVI is a weighted average of all the components, although in the examples given here, the weights are all assigned a value of one. In practice, determination of the value of weights to be applied to an index such as the CVI should be achieved through participatory consultation and expert opinion.

1 See Sullivan (2002) and Sullivan et al. (2002) for a discussion of other ways of creating indices from variables.


Calculating CVI values Once appropriate variables have been identified, and data assembled, the CVI values can be determined. The structure of the CVI can be written mathematically as:

geucar

geucar

rrrrrrGrErUrCrArRr

CVI+++++

+++++=

where R, A, C, U, E and G are the Resource, Access, Capacity, Use, Environment and Geospatial components. Here, the weight given for each component is determined by a factor r representing the relevance of the component in a specific place. Initially, the risk factors r are all set to one, providing the opportunity for the establishment of a ‘base rate CVI’, but these risk factors could be used to focus on specific locally important issues. In practice, risk factors should be estimated from a detailed consultative exercise, and their application should be made explicit, to avoid manipulation of the values for political ends. The resulting CVI scores give a measure of vulnerability to existing climate variability at the present time, and allow comparisons to be made between different locations. The index values range from 0 to 100, with high values indicating high vulnerability. By applying scenarios of future conditions, the change in the CVI scores from the present values indicate how different components of the CVI will impact on the selected aspects of human lives represented by the CVI components. This approach therefore provides a consistent and transparent methodology for the comparative assessment over both time and space of the vulnerability of human populations to global change. Implementation of the CVI The CVI provides a globally generic approach which can be applied to broadly homogeneous zones representing different geographical or ecosystem types. Table 2 shows possible criteria for zone identification, and gives examples of locations to which they could be applicable. Table 2. Possible geographical types for identification of areas of high vulnerability Geographical type Example issues and locations Small islands Sea-level rise, salt water intrusion; Maldives, Caribbean islands, Pacific

atolls Developing cities Inadequate infrastructure, social exclusion, squatter communities; especially

mega-cities Mountainous regions Glacier loss, land slides, soil degradation, loss of forest cover; Nepal,

Bolivia, Ethiopia Semi-arid regions High rainfall variability, desertification; Sahel, north-east Brazil Over-abstracted or degraded catchments

Aral Sea basin, Jordan basin

Low-lying coastal zones Reduced flows from upstream, sea-level rise; Indus, Ganges and Nile deltas Source: Sullivan et al. (2002) For each of the variables used in all of the components, their selection for use would be made on the basis of how effective they are in expressing key aspects of vulnerability that are relevant locally, as well as their practicality and data availability. Most of the data used for the calculation of the CVI would be available from existing sources. Water and environmental data can be derived from a range of international sources and local agencies. Economic data are available from a number of sources, both within a country and internationally. Social data can be taken from census data, health surveys, school enrolment rates, crime statistics and other national data sources. On the basis of the


collated data and the structure of the integrating framework, CVI values can be calculated for specific locations, Within each geographical type, patterns of vulnerability can be examined, and generalised to give broader coverage. This procedure is summarised in Box 2.

Applying scenarios of climate change Through the application of scenarios of future changes, the CVI can become a dynamic modelling tool. Changes in the CVI values need to be based on assumptions about future social, economic and technological driving forces, as well as climate change. For this purpose, we have used four possible future world scenarios: Markets First, Policy First, Security First, and Sustainability First (UNEP, 2002). One of the challenges of making assessments for policy development is the fact that time-scale priorities are different for different stakeholder groups. While climate scientists are often concerned with periods of up to 100 years ahead, for the purpose of estimating climate vulnerability at a policy-relevant time scale, shorter time periods are more appropriate. Here, we consider changes to 2030 by estimating the possible changes in each component arising from projected levels of global change for that period. Interpreting the meaning of CVI scores In order to get a holistic understanding of how changes in the CVI scores actually impact on people, the sustainable livelihoods framework (Scoones, 1998; Carney, 1998) can be used. In practice, this would mean that the CVI components would be taken individually, and considered in terms of the five livelihood asset types used in that framework. By applying this method to specific geographical types (as in Table 2), we will be better able to get a more generic understanding of how climate change will affect human systems. As an example, CVI values for selected small island states were generated using national data from public sources. It is important to note that the results are going to be subject to a degree of uncertainty, and there is a need to indicate the extent of this when information is provided to policy makers. Application of the CVI to small island states The majority of the world’s small island states are concentrated in tropical regions and are developing countries. Many have small land areas with high population densities, a fragile resource base, low economic resilience, and are susceptible to hurricanes and tsunamis (Lal et al., 2002). They are vulnerable because of poverty, lack of skills and weak institutions, but there are other specific geospatial attributes that increase the severity of climate change impacts. These include the extent of land and populations at risk from sea level rise, degree of isolation and dependence on water storage. The potentially devastating effects of sea level rise on small islands are well known. Amadore et al. (1996) identify sea level rise as the greatest threat to small island nations, some of whose very existence is threatened. The threats to water resources are also significant, as reductions

Box 2. Procedure for estimating the CVI • Identify zones of present and likely future water stress. • Identify geographical types likely to be vulnerable, select variables in the Geospatial

component for each, and select sample locations within the geographical types. • Collect and collate all relevant data for the sample locations. • Construct scenarios of change in social, economic and environmental conditions to combine

with estimates of change in water resources derived from climate impact assessments using Global or Regional Climate Model outputs.

• Calculate CVI scores for the present situation, and then under the combined change scenarios. • Interpret the meaning of the CVI scores in terms of the impacts that they would have on people. • Provide results at a range of spatial scales, incorporating some indication of uncertainty.


in precipitation and sea level rise could drastically diminish freshwater lenses. Increasing storage capacity would be one solution to this problem, but one that is not always feasible in small islands. The small scale of most islands in relation to GCM resolutions hampers research on climate trends and impacts, so integrated approaches like the CVI could be of use. Preliminary component and CVI values for four small islands were calculated, with a base year of 2000, and projections for 2030, and these are shown in Table 3. In this first iteration, a number of approximations and simplifying assumptions have been made to generate these data, so the values must be treated as an illustration only. For simplicity, only the Policy First scenario is illustrated here. Table 3. Preliminary CVI values from national data for some small island states

Component values Country Year Res-

ources Access Cap-acity Use Environ

-ment Geo-

spatial CVI

Bahrain 2000 94.2 3.0 13.1 63.6 45.6 38.5 43.0 2030 107.8 0.8 10.2 82.6 58.2 38.5 49.7 Barbados 2000 67.8 0.0 10.2 46.3 45.6 35.8 34.3 2030 78.5 0.0 8.8 53.6 53.2 35.8 38.3 Comoros 2000 69.6 62.0 43.7 57.2 45.6 29.1 51.2 2030 84.2 15.5 31.1 86.5 58.6 29.1 50.8

2000 58.0 12.0 23.0 58.4 53.8 22.8 38.0 Trinidad and Tobago 2030 68.0 3.0 19.7 67.5 62.8 22.8 40.6 Note: these values are included for illustration of the CVI methodology only; they do not indicate any statement on the part of the authors about the countries concerned. Under the Policy First scenario, we can assume that the Millennium Goals on access to safe water and on health (massive reductions in under-5 mortality) are met. This means that the Access and Capacity components improve enormously in the poorer countries (Comoros and Trinidad), while in the others they do not change a great deal as they are already good. Overall, the results show that the poorest country, Comoros, is the most vulnerable under present conditions. In the future, the results indicate that it will remain the most vulnerable, but the increase in vulnerability for the other countries will be much larger. This is especially so for Bahrain, where there is a relatively large extent of land at risk from sea level rise, declining resources (due to population growth) and a worsening environmental situation. With other scenarios, where it would be appropriate to assume that Millennium Goals would not be met, vulnerabilities would show a larger increase in all cases. These examples indicate how this approach can be used for focused, small-scale coverage, but the values are based on national level data. In larger countries, and indeed in some smaller ones, sub-national scale data would provide a much more accurate picture of the degree of vulnerability faced by local populations. Application of the CVI over larger areas and at varied spatial resolutions Figure 1 compares national level application of the CVI approach to the outputs from a finer resolution, grid-based water resources model. This demonstrates how the utilisation of such fine resolution models could generate more accurate site-specific results, and the fact that considerable amounts of data are already available at sub-national scales. The resolution of index-based vulnerability assessment can be further refined by combining outputs from fine resolution grid models, such as that shown in Figure 1, with district (sub-national) social and economic data which is available in many countries. Such a combination can be used for an intermediate level of vulnerability assessment, as illustrated in Figure 2. In this example, CVI values at the department (county) level are mapped for selected departments in Peru, but much social and other data are also available from district sources, a much finer level of resolution. In this example, the percentage of


households having access to piped water are mapped at the district level, and compared to a 0.5o by 0.5o grid such as is typically used for water resources modelling (as in Figure 1). This shows that the two scales are broadly comparable, indicating that combining global-scale modelling of water resources on a grid basis with the CVI could be a very effective approach to make spatially distributed vulnerability assessments over large areas (whole countries or regions).

Figure1. National scale CVI assessment compared to water resources modelling assessment in West Africa. The left diagram shows CVI vulnerability assessments at the national level (green is medium-low, yellow medium, brown medium-high, and red high); the right diagram shows a comparison of water availability to demand (taking account of temporal variation), for the same area (blue indicates the highest level of availability in relation to demand, and red the least). (Source of right diagram: Meigh et al., in press).

Figure 3. Comparison of spatial scales of social and water information. The left map shows the CVI values (for 2000) for some of the departments of Peru. The right map shows finer resolution data, illustrating the percentage of households with a piped water supply at the district level, and this is compared to a 0.5o by 0.5o modelling grid. (Source of water supply information: Instituto Nacional de Estadística e Informática, Perú, www.inei.gob.pe).


Further spatial refinement can be achieved by applying this methodology to specific communities or locations, within a country. An example of this is provided by examining how the CVI can provide insights into the situation of vulnerability at many different scales. This is illustrated in Figure 3 which shows how vulnerability assessment can be applied at a range of scales, highlighting a comparison of national and local assessments for South Africa. By using a standardised structure, definition and measurement of variables, CVI values can be nested at different scales, providing a robust and detailed picture of how human vulnerability varies both spatially and temporally as a result of climate change impacts. In this figure, the CVI scores shown for the two local communities differ significantly, highlighting the need to move away from large scale or national averages which are unrepresentative of the high degree of variability found within most countries of the world.

Figure 3. How CVI values can be nested to generate better coverage Discussion The application of the Climate Vulnerability Index provides a novel approach to the assessment of vulnerability to climate change and global change more generally. This is an important and urgent issue, given the likelihood of increasing extreme events. Estimates of the financial costs of such events are likely to far outweigh the costs of their prevention, and the human and social costs of climate change impacts are in many ways, unaccountable. As with any other technique, the reliability and accuracy of the CVI approach are influenced by the available data and the method of analysis. We have shown that there is a wealth of data at many scales relevant to vulnerability assessment, but currently the data are often both fragmented and inconsistent. There is an urgent need to develop systematic and robust approaches to data collection, storage and analysis, and global efforts should be made to achieve this. A parallel can be drawn with the process of developing standardised national financial accounting systems, and as a result of efforts made by key institutions in the 1980s and 1990s, most countries have now implemented the UN System of National Accounts, enabling international comparisons of economies to be made, and ensuring some measure of integrity and validity of financial data. While there is still some way to go in refinement of the structure of the CVI, and its application to different types of ecosystems, it seems clear that an indicator-based approach can make a useful contribution to the process of vulnerability assessment. The use of scenarios to build in a more dynamic dimension to the CVI approach enables the impacts of climate and other aspects of global


change to be investigated for different time scales, and different spatial scales can be investigated through the integration of different types of data from different sources. Through the use of the indicator structure, the problem of incommensurability of variables is much reduced. Conclusion Most work on climate impacts has evolved from the disciplines of meteorology, hydrology and physics. The CVI provides an approach to vulnerability assessment which goes beyond this, being more holistic, and applicable at a variety of scales. It explicitly highlights the fact that much vulnerability is due to human, social and political factors rather than purely physical ones. By facilitating more site-specific approaches, and being less deterministic than other models, the CVI can contribute to more effective adaptation strategies in response to the impacts of climate change. Acknowledgements: This paper has been generated from work funded by the Climate and Water Dialogue, the UK Department for International Development, and the UK Natural Environment Research Council. The authors are indebted to all of those who have contributed to the development of the ideas in this paper. References Adger, N. and Brooks, N. 2002. Does global environmental change cause vulnerability to natural

disaster? In Pelling, M. (ed.), Natural Disaster and Development in a Globalising World, Routledge, 19-42

Amadore L., Bolhofer W. C., Cruz R. V., Feir R. B., Freysinger C. A., Guill S., Jalal K. F., Iglesias A., Jose A., Leatherman S., Lenhart S., Mukherjee S., Smith J. B. and Wisniewski J. (1996). Climate change vulnerability and adaptation in Asia and the Pacific: workshop summary. Water, Air, and Soil Pollution, 92, 1-12.

Carney D. (ed.) (1998). Sustainable Rural Livelihoods: What Contribution Can We Make? Department for International Development, London.

Edgeworth F.Y. (1925). The plurality of index numbers. The Economic Journal, 35(139), 379-388. Fisher, I. (1922) The making of index numbers. Journal of the American Statistical Association Hammond, A., A. Adriaanse, E. Rodenburg, Bryant D. and Woodward R. (1995), Environmental

Indicators: A Systematic Approach to Measuring and Reporting on Environmental Policy Performance in the Context of Sustainable Development. World Resources Institute, Baltimore.

Intergovernmental Panel on Climate Change (IPCC) (2001). Climate Change 2001: Impacts, Adaptation and Vulnerability. Cambridge University Press, Cambridge and New York.

International Strategy for Disaster Reduction (ISDR) (2002). Living with Risk. ISDR, Geneva. Lal M., Harasawa H. and Takahashi K., 2002. Future climate change and its impacts over small

island states. Climate Research, 19, 179-202. Lenton, T.M. (2002) Testing Gaia: the effect of life on Earth's habitability and

regulation. Climatic Change 52: 409-422 Meigh J. R., Folwell S. S. and Sullivan C. A. (in press). Linking water resources and global change

in West Africa: options for assessment. International Association of Hydrological Sciences, VIIth Scientific Assembly.

Rennings K. and Wiggering H. (1997). Steps towards indicators of sustainable development: Linking economic and ecological concepts. Ecological Economics, 20, 25-36.

Rogers P. R., Jalal K. F., Lohani B. N., Owens G. M., Chang-Ching Yu, Dufournaud C. M and Bi, J. (1997). Measuring Environmental Quality in Asia. Harvard University Press, Harvard.

Scoones I. (1998). Sustainable Rural Livelihoods: A Framework for Analysis. IDS Working Paper No. 72, Institute of Development Studies, University of Sussex, UK.

Streeten P. P. (1995) Human development – the debate about the index. International Social Science Journal, 47, 25-37.

Sullivan C.A. (2002). Calculating a Water Poverty Index. World Development, 30, 1195-1210.


Sullivan C. A., Meigh J. R., Giacomello A. M., Fediw T., Lawrence P., Samad M., Mlote S., Hutton

C., Allan J. A., Schulze R. E., Dlamini D. J. M., Cosgrove W., Delli Priscoli J., Gleick P., Smout I., Cobbing J., Calow R., Hunt C., Hussain A., Acreman M. C., King J., Malomo S., Tate E.L., O’Regan D., Milner S. and Steyl I. (2003). The Water Poverty Index: development and application at the community scale. Natural Resources Forum, 27, 189-199.

Sullivan C. A. Meigh J. R. and Acreman M. C. (2002). Scoping Study on the Identification of Hot Spots – Areas of high vulnerability to climatic variability and change identified using a Climate Vulnerability Index. Report to Dialogue on Water and Climate, Centre for Ecology & Hydrology, Wallingford, UK.

United Nations Environment Programme (UNEP) (2002). Global Environment Outlook 3: Past, present and future perspectives (GEO-3). Earthscan Publications, London and Sterling, VA.

World Bank (1998). Standardised Welfare Indicators. World Bank, Washington, DC.

Targeting attention on local vulnerabilities using an integrated … · 2011-02-28 · Water...

Documents

Transcript of Targeting attention on local vulnerabilities using an integrated … · 2011-02-28 · Water...