sustainability and vulnerability indicators for decision making: lessons learned from honduras

13
Sustainability and vulnerability indicators for decision making: lessons learned from Honduras Manuel Winograd European Environment Agency, Kongens Nytorv 6, DK-1050 Copenhagen K, Denmark E-mail: [email protected] Abstract: The goal of this paper is to show, in practice, how to define and use indicators to provide and communicate information for policy-making and decision-making to reduce environmental, social, and economic vulnerability and increase sustainability. The information should allow the definition of strategies regionally, the elaboration of policies nationally, and the application of actions locally to move from blaming ‘climatic uncertainty’ to ‘plan’ possible impacts, ‘adapt’ to adverse consequences, ‘prevent’ negative effects and ‘mitigate’ direct and underling causes. Analysis from Honduras on pre- and post- Hurricane Mitch (October 1998) environmental, social and economic conditions and vulnerability, and evaluations of optional response strategies, are presented as examples of how to produce meaningful information to close the gap between research and action. Keywords: assessment; Central America; climate; decision-making; Honduras; indicators; sustainability; vulnerability. Reference to this paper should be made as follows: Winograd, M. (2007) ‘Sustainability and vulnerability indicators for decision making: lessons learned from Honduras’, Int. J. Sustainable Development, Vol. 10, Nos. 1/2, pp.93–105. Biographical notes: Manuel Winograd was a researcher and director in the Ecological Systems Analysis Group at Fundacion Bariloche (Argentina) and international fellow at the World Resources Institute (USA). During the last 12 years he was senior researcher at the International Center for Tropical Agriculture (Colombia) and visiting scientist in CIRAD (France). He is the Project Manager on integrated assessment at the European Environment Agency ((EEA, Copenhagen, Denmark). He has served as consultant and advisor for many international, regional and national institutions (World Bank, World Bank Institute, United Nations Environment Program, Inter-American Institute for Agriculture Cooperation, United Nations Development Program and FONTAGRO). He was coordinator of different projects on sustainability indicators, vulnerability assessment, rural development and environmental assessment in Latin America and Africa. He works on sustainability and vulnerability indicators, the production of decision-making tools and information, land-use models and scenarios, and capacity strengthening. Int. J. Sustainable Development, Vol. 10, Nos. 1/2, 2007 93 Copyright # 2007 Inderscience Enterprises Ltd.

Upload: independent

Post on 03-Dec-2023

0 views

Category:

Documents


0 download

TRANSCRIPT

Sustainability and vulnerability indicators fordecision making: lessons learned from Honduras

Manuel Winograd

European Environment Agency, Kongens Nytorv 6,DK-1050 Copenhagen K, DenmarkE-mail: [email protected]

Abstract: The goal of this paper is to show, in practice, how to define anduse indicators to provide and communicate information for policy-makingand decision-making to reduce environmental, social, and economicvulnerability and increase sustainability. The information should allowthe definition of strategies regionally, the elaboration of policies nationally,and the application of actions locally to move from blaming `climaticuncertainty' to `plan' possible impacts, `adapt' to adverse consequences,`prevent' negative effects and `mitigate' direct and underling causes.Analysis from Honduras on pre- and post- Hurricane Mitch (October1998) environmental, social and economic conditions and vulnerability, andevaluations of optional response strategies, are presented as examples ofhow to produce meaningful information to close the gap between researchand action.

Keywords: assessment; Central America; climate; decision-making;Honduras; indicators; sustainability; vulnerability.

Reference to this paper should be made as follows: Winograd, M. (2007)`Sustainability and vulnerability indicators for decision making: lessonslearned from Honduras', Int. J. Sustainable Development, Vol. 10, Nos. 1/2,pp.93±105.

Biographical notes: Manuel Winograd was a researcher and director in theEcological Systems Analysis Group at Fundacion Bariloche (Argentina)and international fellow at the World Resources Institute (USA). Duringthe last 12 years he was senior researcher at the International Center forTropical Agriculture (Colombia) and visiting scientist in CIRAD (France).He is the Project Manager on integrated assessment at the EuropeanEnvironment Agency ((EEA, Copenhagen, Denmark). He has served asconsultant and advisor for many international, regional and nationalinstitutions (World Bank, World Bank Institute, United NationsEnvironment Program, Inter-American Institute for AgricultureCooperation, United Nations Development Program and FONTAGRO).He was coordinator of different projects on sustainability indicators,vulnerability assessment, rural development and environmental assessmentin Latin America and Africa. He works on sustainability and vulnerabilityindicators, the production of decision-making tools and information,land-use models and scenarios, and capacity strengthening.

Int. J. Sustainable Development, Vol. 10, Nos. 1/2, 2007 93

Copyright # 2007 Inderscience Enterprises Ltd.

1 Introduction

The goal of sustainable development is to create and maintain prosperous and viablesocial, economic, institutional and ecological systems (Carpenter et al., 2002).Sustainable development has also been described as fostering adaptive capabilitiesand creating opportunities. This definition comes from combining sustainability, asthe capacity to create, test, maintain and improve adaptive capability and development,as the process of creating, testing, maintaining and improving opportunities (Holling,2001). At the end, the challenges for sustainable development are related to theimprovement of resilience and adaptive capacities, take advantage of emergingopportunities and cope with the consequences of different processes of change. Inthis context, vulnerability is emerging as a critical component of any sustainabledevelopment strategy (Kasperson and Kasperson, 2001). Vulnerability could bedescribed as a function of risks and threats minus adaptive options and copingresponses (Downing et al., 2003; IPCC, 2001). At the end, from a pragmaticperspective, vulnerability and sustainability can be seen as two sides of the same coin.

In the last few decades, reducing vulnerability to climate variability and climatechange has increasingly become an urgent issue for developing countries, likeHonduras, for at least two reasons. Firstly, because there are significant omissions inthe way in which they face up to the impacts and economic and social effects of thesechanges; secondly, because the economies of these countries are heavily dependent onsectors that are highly sensitive to natural disasters and climate variations andchanges, such as agriculture, coastal resources, water resources and infrastructures.Thus, for countries like Honduras, adapting to natural disasters, climate variabilityand climate change is a major concern of any policy agenda toward sustainabledevelopment.

The results described in this paper were developed under a collaborative projectbetween CIAT (Land use unit), the World Bank (Disaster Management FacilityUnit) and the Government of Honduras (Secretaria de Recursos Naturales y delAmbiente, SERNA). In this context, at international and regional level, the mainusers were the World Bank and CIAT with the goal to define actions to evaluate andnegotiate mitigation and reconstructions options at regional and national level aswell explore options for adaptation face to climate variability and climate change. Atnational and local level, the main users were the World Bank, the government ofHonduras, the national institutions and the federation of municipalities in Honduras(AMHON) with the goal to determine, explore and negotiate priorities onreconstruction, rehabilitation, mitigation and adaptation options, actions andresponses related to land use and urban planning and early warning systems.

2 The assessment framework

Decision and policy making are not random processes and are based on the use ofknowledge, insight and information at different levels (Bossel, 1999). In consequenceassess sustainability and vulnerability implies produce information to evaluate theconsequences of development strategies, policies and actions on the developmentcomponents and processes. It is necessary to define a pragmatic framework, based on

M. Winograd94

what is known from theories and what is learned in practice, that can be used as amodel to guide, define and use appropriate indicators for the system (i.e. structure/functions, scales/levels, viability/integrity, goods/services) and the steps for decisionand policy making (i.e. conditions, diagnosis, forecasts, responses and evaluation)(Figure 1). The system model defines a way to map and evaluate the systemcharacteristics onto the identified components of sustainability and vulnerability.The proposed model defines a hierarchical decomposition, by scales and levels, whereparticular system variables can be fitted in order to allow consistent identificationand selection of indicators (Kasperson and Kasperson, 2001; Metzger et al., 2004;Villa and McLeod, 2002; Winograd et al., 2000a,b).

Figure 1 Framework to assess sustainability and vulnerability

2.1 The indicators set

Given the broad nature of vulnerability and sustainability issues, and to ensure somepractical application, the definition of a first set of information was done for climaticissues (Table 1). Natural events (as a single issue) and climate change (as a complexproblem) were used as good examples for defining the indicators. The set of indicesare to be used mostly at regional and national scales and to define strategies andpolicies, while the indicators set is to be used mostly at national and local scales andto apply and to monitor policies and actions (Table 1). At the same time it isimportant to note that the majority of indices come from the aggregation andweighting of the proposed indicators, to allow users to use detailed or aggregatedinformation (Dilley et al., 2005).

Sustainability and vulnerability indicators for decision making 95

Table 1 Examples of indices and indicators for climateSteptype

Condition

Diagnosis

Forecast

Responses

Environment

Society

Environment

Society

Environment

Society

Environment

Society

Index

Climaticrisk

index.

Forest

risk

index.

Coastalrisk

index.

Human

development

index.

Accessibility

index.

Environmental

vulnerability

index.

Landuse

vulnerability

index.

Population

vulnerability

index.

Infrastructure

vulnerability

index.

Social

vulnerability

index.

Landuse

index.

Prioritised

vulnerability

index.

Vulnerable

environmental

hotspots.

Vulnerable

social

hotspot.

Indicator

Frequency

of

naturalevents.

Probabilityof

naturalevents.

Riskofflood.

Riskof

drought.

Riskof

landslides.

Locationof

fires.

Forestssurfaces.

Forest

fragmentation.

Actuallanduse.

Populationin

poverty.

Welfare

access.

Economic

loss

bynatural

disasters.

Population

affectedby

naturaldisasters.

Humanloss

by

naturaldisasters.

Changes

in

temperature.

Changes

in

precipitation.

Landuse

changes.

Ecosystem

s

recovery.

Healthand

emergency

servicecoverage.

Household

structure.

Housingquality/

condition.

Housing

location.

Landuse

options.

Production

system

s

options.

Urbanlanduse

options.

Institutional

options.

Insurance

coverage.

Potentialland

use.

Future

predominant

habitats.

Prevention

standards/

policies.

Preventionplans.

Emergency

programmes.

Locationof

reconstruction

andmitigation

projects.

Earlywarning

system

s.

Resettlem

ent

programmes.

M. Winograd96

The main goal of the project was to determine the areas with a high degree ofvulnerability with regard to climate-related natural hazards, in order to prioritise thedifferent Honduras municipalities and identify reconstruction and mitigation optionsafter Mitch and adaptation and coping strategies in anticipation of possible furthernatural events. To reach this goal a methodology to analyse natural disastervulnerability was defined and validated and harmonised with partners and nationalinstitutions. The methodology is based on the development of different vulnerabilityindices (environmental vulnerability index, population vulnerability index, socialvulnerability index and infrastructure vulnerability index) (Figure 2) in order toprioritise vulnerable groups, areas and sectors.

Figure 2 Methodology for vulnerability assessment

Source: Winograd et al. (2000b)

3 An example of application: vulnerability in Honduras

The use of information for decision making and planning needs to go beyond simplymonitoring the catastrophic effects of natural events on the economy, society and theenvironment and the planning of reconstruction after the catastrophe has alreadyoccurred. What is needed is the possibility to go from short-term reaction to theimmediate impact, to the prevention of the direct and indirect causes in order tomitigate and adapt against future possible effects. A first step toward this is toanalyse the underlying causes of the situation caused by Mitch, which does not onlyconsist of abnormal amounts of rain during October 1998. In fact, a combination ofpast problems had an influence on the severity of the impact of Hurricane Mitch.These problems include inappropriate land use practices, high deforestation rates inthe watersheds, and a lack of urban planning, early warning systems and investmentin works to prevent and mitigate natural events (Van Aalst and Burton, 2002).

Sustainability and vulnerability indicators for decision making 97

3.1 What did the vulnerability mean for the environmental and socio-economicconditions of a country like Honduras?

The environmental vulnerability for Honduras (based on areas vulnerable todrought, landslides and flooding) shows that the situation in the nation is clearlyalarming. In fact, more than 50% of the country is at risk of landslides or droughts(60% of the used land). More than 25% of the country is at risk of flooding (50% ofthe used land), particularly in the more populated agricultural areas in the San PedroSula-Tegucigalpa-Choluteca axis.

Thus for example, Figure 3 shows the result of overlap of the environmentalvulnerability index with the population vulnerability index, the social vulnerabilityindex and the infrastructure vulnerability index to obtain a combined vulnerabilityindex that helps to identify the priority or `hotspot' areas. The populationvulnerability index is the total population by `village' at risk of flooding plus thetotal population by `village' at risk of landslide, given the importance of floods andlandslides resulting from Mitch. The social vulnerability index is the total populationby `village' by poverty levels at risk of flood plus the total population by `village' bypoverty level at risk of landslides. The infrastructure vulnerability index is based onroads at risk of flood plus roads at risk of landslide plus electricity lines at risk offlood plus electricity lines at risk of landslides. In the case of Figure 3, priority areasare the counties in Honduras that need to be prioritised as a function of thecombined vulnerability index to identify mitigation and adaptation options.

Figure 3 Vulnerability index at national level (`Municipios' in Honduras)

Source: Winograd et al. (2000b)

M. Winograd98

3.2 What did the vulnerability mean in terms of diagnosis of the direct andindirect impacts of Hurricane Mitch and other natural events at anational level?

Although the poor in Honduras would appear to be the most vulnerable section ofthe population faced with natural disasters, they do not necessarily seem to have beenharder hit than other groups in society, when considering only the direct effects ofMitch. However, if the indirect effects (i.e. in the form of loss of accessibility) aretaken into account the picture looks a bit different. Figure 4(a) shows, in dark line/circle the areas that lost more than 10% of the accessibility (defined as travel timetaken to reach a defined location) to markets and in bright grey and dark grey circlesthe location of villages with severe and critical levels of poverty. This illustrates howsome areas only suffered indirectly as a result of damage to infrastructure. Thedifficulty of access between population centres and rural areas caused serious delaysin the arrival of medical aid, food and seeds. It is important to note that Figure 4(a)shows those areas that have lost accessibility (as opposed to all areas that wereaffected by Mitch). This was calculated by comparing the accessibility before andafter Mitch. The loss is calculated as a percentage rather than a value in hours andtherefore highlights the perceived loss of accessibility.

Figure 4 Socioeconomic `hotspots' and reconstruction options at national level

3.3 What did the vulnerability mean in terms of prioritisation for immediatereconstruction actions and long-term mitigation responses at national andlocal level?

Figure 4(a) also shows two potential reconstruction options that would help alleviatethe impacts of Mitch on Honduras' poorer population. The first option is orientatedtoward reconstructing strategic roads in Honduras affected by Mitch and recoveringthe accessibility between the most important cities, ports and productive resources(with an indirect goal of recovering the accessibility of poor villages, when poverty is

Sustainability and vulnerability indicators for decision making 99

calculated based on unsatisfied basic needs). The second option is orientated towardsreconstructing the road network that allows access to isolated poor villages. Figure4(b) shows the costs and benefits of these two options. It is worth observing that inthe first case the cost is lower and the number of poor people who benefit is higherthan in the second option. Obviously, these two options are exploratory and notexclusive and the analysis was made with the goal of showing the importance of usinginformation for decision making, to explore planning strategies and to prioritisereconstruction options for the sectors most affected.

3.4 What did the vulnerability mean in terms of exploration of adaptationoptions and evaluation of responses at national and local levels?

The flat areas, where the majority of permanent crops are located, were the areasmost affected by flooding after Mitch. In addition, flooding in the north, centre andsouth of the country could be due in part to the low forest coverage and the land usechanges in the upper watersheds and, as a consequence, potential future vulnerabilitycould be increased. This land use could exacerbate highly erosive processes in thehillside areas resulting in a high risk of landslides and dammed rivers in theheadwaters with inundation in the lower part of the river. This type of informationon the underlying causes to the observed effects could be potentially crucial inhelping to prioritise interventions, identify the financial aid and investment neededand the people and sectors that would benefit, and finally to plan mitigation andprevention actions.

3.5 What did vulnerability mean in terms of typical situations or interactionsrelevant for the persistence of patterns that affect prevention, mitigationand adaptation to natural events?

Figure 5(a) shows the environmental vulnerability index at local levels in four villagesof the Department of Olancho, Honduras. Figure 5(b) shows the actual predominantproduction systems and associated land uses, where the actual and futurevulnerability is high given that current land-use practices increase the risk oflandslides and flooding. Depicted in Figure 5(c) are alternative production systemsand land-use options that help to diminish vulnerability and increase resilience giventhat land use decreases the risks of landslides, erosion and flooding.

As depicted in Figure 6, mangroves provide a key ecosystem for goods (i.e.fisheries, aquaculture, and wood) and services (i.e. protecting coasts from floods anderosion and serving as nurseries for wild fish and birds). In Roatan Island themangroves and forests are protected and/or managed (Figure 6(a)). In contrast inGuanaja Island (Figure 6(b)) we witnessed poor management and high deforestationin mangroves and upper forests. After Mitch, the Guanaja system in Figure 6(c)changed to that in Figure 6(e) (dead mangrove and salt marsh, a new system withnew functional and ecological patterns, goods and services). On the other hand, inthe case of Roatan Island (Figure 6(a)), good management and low deforestation inmangroves and upper forests allowed the system to recover to a steady-state system(Figure 6(d)).

M. Winograd100

Figure 5 Exploration of adaptation options of productions systems at local level (hillsides inHonduras)

Source: Winograd et al. (2000a)

Figure 6 Resilience assessment at local level: the case of mangroves in Honduras

Source: Winograd et al. (2000a); photo sources: USGS (2001)

Sustainability and vulnerability indicators for decision making 101

In the case of land use and production systems, the analysis of resilience (defined asthe capacity of a system to absorb perturbation or stress and maintain structure andfunction (Carpenter et al., 2002)) also shows interesting results (Figure 7). For example,in hillside areas when a hurricane strikes, like Mitch, land-use patterns will rapidly bechanged, during the disturbance phase (circle 2 in Figure 7(a) and (b)). In the case oflow capital and mono-cropping systems (Figure 8(a)), the low diversity of land-usepatterns decreases resilience and complexity, which explains the slow reorganisationphase initialised afterwards. In many cases the system is unable to return to itsformer equilibrium (circle 3 in Figure 7(a)) and key variables (agriculturalproductivity, production diversity, and food surplus production) will certainlyhave changed. A new (re)colonisation phase will start, which could result in newland-use patterns and production systems. Conversely, in the case of low capital andmulti-cropping production systems (Figure 7(b)), the diversity of land use increasesresilience and complexity, which explains the fast reorganisation phase initialisedafterwards. Thus the multi-cropping production system returns to its formerequilibrium with key variables unchanged (agricultural productivity, productiondiversity, and food surplus production). In general the land-use system in Hondurasappears to be resilient, but not necessarily stable (Kok and Winograd, 2002).

Figure 7 Resilience of land use and production systems in Honduras

Source: Kok and Winograd (2002), modified

4 Lesson learned: from assessment to decision making

In post-Mitch Honduras, prevention and adaptation were perceived as costs ratherthan investments and strategies. Policies and actions were oriented to solve theconsequences instead of preventing the causes. After hurricane Mitch, less than 40%of the reconstruction and mitigation projects were located in highly affectedMunicipios (Figure 8(a)). At the same time, given the urgency and the lack ofappropriate information, less than 10% of the reconstruction and mitigation projectswere in highly vulnerable Municipios (Figure 8(b)).

M. Winograd102

Figure 8 Evaluation of responses and actions at national level

Source: Winograd et al. (2000b)

Thus, for example, estimates show that the losses from hurricane Mitch throughoutCentral America reached US$8.8 billion. Multilateral institutions, internationalagencies and the governments of developed countries promised to donate US$8.7billion to the countries of the region (although less than US$3 billion had reachedthose nations by the end of 2004, 6 years after the disaster). However, investments ofonly US$0.35 to US$0.5 billion for mitigation, attenuation and adaptation measures(land use and early warning systems) would have resulted in US$3.5 billion in lossesbeing prevented and/or avoided (The Economist, 2005).

Although analysts understood the dynamics of vulnerability, the spatial mappingof static indicators failed to represent the importance of some drivers and key issues(i.e. infrastructure and economic access to isolated villages, effect of uplandsdeforestation in lowlands croplands and mangroves). In fact, many assessments arecurrently adapted for short-term evaluation when there is a need for analysis of thelong-term risks (i.e. cumulative effects of climate natural events on vulnerability,possible synergies and thresholds of climate changes).

The use of information for decision making and planning needs to go beyondsimply monitoring the catastrophic effects of natural disasters on the economy, thesociety and the environment and the planning of reconstruction after the catastrophehas already occurred. What is needed is the prevention of the direct and indirectcauses in order to mitigate against future possible effects (Downing et al., 2003;Ziervogel et al., in press). It is important to note that, in the case of vulnerability

Sustainability and vulnerability indicators for decision making 103

research, there is a need to assess vulnerability, not only as the exposure to risk, butalso as the ability to cope with risks and their impacts: in other words, putdevelopment first. To prevent natural events from becoming disasters, policies andactions need to be proactive instead of reactive (Tompkins and Adger, 2003).

A crucial point, related with GIS tools and models, is that we need `socialise' thepixel to link spatial entities (countries, landscapes, watersheds, ecosystems) with thesocial actors and users (peoples, households, communities, decision makers) and thecauses of vulnerability (livelihoods, landslides, drought, poverty) (Verburg et al., 2004).

Finally, one of the key issues to success with any assessment initiative is to createand strengthen capacities to produce and use information and identify best practices.To create real capacity, it is not enough to give already developed frameworks,methods, tools, recommendations and information to the institutions, people anddecision-makers. More importantly, training is needed on how to developvulnerability assessments, how to use information and how to improve policy anddecision making and planning. This allows actors and institutions to gain `ownership'of the initiative and convert information into action. Without good information andbest practices to use methods and tools on vulnerability indicators, it is impossible togenerate participation, manage risk and cope with the consequences of changes. Thefirst essential step is to correct existing practices that misuse data, information andtools to ensure better decisions, facilitate actors' participation at different levels andenable non-specialists to use and access information effectively.

5 Conclusions

To conclude, we should strongly emphasise the need to ensure that the situation atthe grassroots level (development and institutional contexts, technical capacities andactors needs) is linked to the methods and tools used as well to the processes inprogress. Information needs vary according to the realities, the people involved,interest groups and decision-making objectives. The information set should bepolitically relevant, socially acceptable and technically valid to ensure that assessmentsare not based solely on partial, anecdotal or temporary issues. To achieve this, theinformation on vulnerability, impacts and adaptation options, like those presented inthis section, should be translated from scientific language and time scales to thelanguage and time scales of the decision-makers and actors at all levels. In allcircumstances, it is important to link spatial entities (villages, landscapes, watersheds,ecosystems) to social entities (individuals, families, communities, towns) in such away as to acknowledge a reality that implies both spatial and temporal dimensions.

This approach will lead to credibility and the political and social appropriation ofthe problems of vulnerability, prevention and adaptation, so that they end up beingperceived as investments and not just as costs. Evaluation processes should be usedto define regional strategies, draw up national policies, identify options andimplement actions at a local level. It then becomes possible to change the cycle thatmoves from ad hoc responses to natural disasters, climate variability and climatechange towards participation and empowerment of communities aimed at preventingthe causes, based on the planning of possible impacts, adapting to adverseconsequences, preventing negative effects and reducing the direct and indirect causes.

M. Winograd104

References

Bossel, H. (1999) `Indicators for sustainable development: theory, method, application',Report to the Balaton Group, IISD, Winnipeg, Canada.

Carpenter, C.S., Elmqvist, T., Gunderson, L., Holling, C.S. and Walker, B. (2002) `Resilienceand sustainable development: building adaptive capacity in a world of transformations',Ambio, Vol. 31 pp.437±440.

Dilley, M., Chen, R.S., Deichmann, U. et al. (2005) `Natural disaster hotspot: a global riskanalysis', Disaster Risk Management Series No. 5, Hazard Management Unit, The WorldBank, Washington, DC.

Downing, T., Patwardhan, A., Klien, R., Mukhala, E., Stephen, L., Winograd, M. andZiervogel, G. (2003) `Vulnerability assessment for climate adaptation', AdaptationPlanning Framework, Technical Paper 3, UNDP, New York.

Holling, C.S. (2001) `Understanding the complexity of economic, ecological, and socialsystems', Ecosystems, Vol. 4, pp.390±405.

IPCC (2001) Climatic Change 2001: Impacts, Adaptations and Vulnerability, A Report ofWorking Group II, Technical Summary, IPCC.

Kasperson, J. and Kasperson, R. (2001) International Workshop on Vulnerability and GlobalEnvironmental Change: A Workshop Summary, Stockholm Environment Institute (SEI),SEI Risk and Vulnerability Programme, Report 2001±01, Stockholm, Sweden.

Kok, K. and Winograd, M. (2002) `Modelling land-use change for Central America, withspecial reference to the impact of hurricane Mitch', Ecological Modeling, Vol. 149, pp.53±69.

Metzger. M., Leemans, R. and Schroter, D. (2004) `A multidisciplinary multi-scale frameworkfor assessing vulnerability to global change,, Millennium Ecosystem AssessmentConference: Bridging Scales and Epistemologies, 17±20 March 2004, Alexandria, Egypt.

The Economist (2005) Vol. 374, Number 8408, p.27.

Tompkins, E. and Adger, N. (2003) `Building resilience to climate change through adaptivemanagement of natural resources', Tyndall Centre for Climate Change, Research WorkingPaper 27.

Van Aalst, M. and Burton, I. (2002) `The last straw integrating natural disaster mitigation withenvironmental management', Disaster Management Facility Unit, Working Paper No. 5,The World Bank, Washington, DC.

Verburg, P., Veldkamp, A., Willement, L., Overmars, K. and Castella, J.C. (2004) `Landscapelevel analysis of the spatial and temporal complexity of land-use change', in R. DeFries,G. Asner, R. Houghton (Eds) Ecosystems and Land Use Change, Geophysical MonographSeries, Vol. 153, pp.143±167.

Villa, F. and MacLeod, H. (2002) `Environmental vulnerability indicators for environmentalplanning and decision-making: guidelines and applications', Environmental Management,Vol. 29, No. 3, pp.335±348.

Winograd, M., Farrow, A., Aguilar, M. and Kok, K. (2000a) Indicadores de SosteniblidadRural: Una vision para America Central, CD-ROM ArcView Data Publisher 3.1, CIAT,The World Bank, UNEP and ESRI, Cali, Colombia.

Winograd, M., Schillinger, S., Farrow, A. and Hernandez, I. (2000b) `Vulnerabilidad frente adesastres naturales en Honduras', CIAT, Cali, Colombia, available from: http://gisweb.ciat.cgiar.org.

Ziervogel, G., Cabot, C., Winograd, M., Segnestam, L., Wilson, K. and Downing, T. (In press)`Vulnerability assessments and risk maps: do they help to protect the vulnerable? A casestudy of Honduras', in L. Stephen, T. Downing and A. Rahman (Eds) Approaches toVulnerability: Food Systems and Environments in Crisis, London, UK, EarthscanPublisher, Chapter 8, in press.

Sustainability and vulnerability indicators for decision making 105