Understanding livelihood vulnerability to climate change

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236576858

Understanding livelihood vulnerability to climate change: Applying the livelihood vulnerability index in Trinidad and Tobago ARTICLE in GEOFORUM · APRIL 2013 Impact Factor: 1.93 · DOI: 10.1016/j.geoforum.2013.04.004

CITATIONS

READS

12

783

4 AUTHORS: Kalim U. Shah

Hari Bansha Dulal

Indiana University

World Bank

21 PUBLICATIONS 132 CITATIONS


SEE PROFILE

SEE PROFILE

Craig Anthony Johnson

April Karen Baptiste

University of Guelph

Colgate University



SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.

SEE PROFILE

Available from: Kalim U. Shah Retrieved on: 14 January 2016

Geoforum 47 (2013) 125–137

Contents lists available at SciVerse ScienceDirect

Geoforum journal homepage: www.elsevier.com/locate/geoforum

Understanding livelihood vulnerability to climate change: Applying the livelihood vulnerability index in Trinidad and Tobago Kalim U. Shah a, Hari Bansha Dulal b,⇑, Craig Johnson c, April Baptiste d a

York University, Toronto, Canada World Bank, 1818 H Street, NW, Washington, DC 20433, USA c University of Guelph, Ontario, Canada d Colgate University, New York, USA b

a r t i c l e

i n f o

Article history: Received 25 June 2012 Received in revised form 31 March 2013

Keywords: Livelihood vulnerability index Climate change Wetlands Gendered vulnerability Trinidad and Tobago Rural households Small island developing states

a b s t r a c t This paper develops and tests the application of a Livelihood Vulnerability Index (LVI) for agricultural and natural resource-dependent communities in developing countries. The index is applied in a comparative study of two wetland communities in Trinidad and Tobago, a country that is expected to bear some of the most severe impacts of climate change. Our application of the LVI entailed a series of critical focus group discussions involving local community representatives, government officials and researchers. Researchers collected household data for eight types of assets, which were aggregated into composite LVIs and differential vulnerabilities of the two communities being compared. The results of the analysis suggest that one of the communities, ‘‘Nariva’’, was more vulnerable than the other, ‘‘Caroni’’, particularly in relation to socio-demographics, health and water security, natural disaster and climate variability. Caroni on the other hand was more vulnerable in relation to other LVI indicators with the exception of food security. On questions of gender, the study found that female-headed households were marginally more vulnerable than male-headed households. Overall, the study suggests that the livelihood vulnerability index can be broadly applied in comparable settings in small-island developing states and other developing countries. In so doing, it provides a reliable methodology that can be used to assess community vulnerability and design management plans in areas with limited resources and access to reliable data. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Vulnerability indicators provide a potentially useful means of monitoring vulnerability over time and space, identifying the processes that contribute to vulnerability, prioritizing strategies for reducing vulnerability, and evaluating the effectiveness of these strategies in different social and ecological settings (Adger et al., 2009; Dow, 1992). To date however, definitions and assessments of climate change vulnerability are often applied inconsistently. The Intergovernmental Panel on Climate Change provides a useful typology, suggesting that vulnerability may be characterized as a function of three components: adaptive capacity, sensitivity, and exposure (Schneider et al., 2007). Adaptive capacity describes the ability of a system to adjust to actual or expected climate stresses, or to cope with the consequences. It is considered ‘‘a function of wealth, technology, education, information, skills, infrastructure, access to resources, and stability and management capabilities’’ (McCarthy et al., 2001). Recent research also indicates that perceptions of social identity by communities play a strong role in climate ⇑ Corresponding author. E-mail addresses: [email protected], [email protected] (H.B. Dulal). 0016-7185/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.geoforum.2013.04.004

risk perception and adaptive ability (Frank et al., 2011). Sensitivity refers to the degree to which a system will respond to a change in climate, either positively or negatively. Exposure relates to the degree of climate stress upon a particular unit of analysis; it may be represented as either long-term changes in climate conditions, or by changes in climate variability, including the magnitude and frequency of extreme events. Landmark studies of disasters, risk and climate change highlight three broad characterizations about the dynamic and integrated nature of social and environmental vulnerability (Watts and Bohle, 1993; Blaikie et al., 1994; Kelly and Adger, 2000). One concerns the impact of exposure to hazardous events on human populations and social structures. A second explores the social and historical conditions under which people are put at risk to a diverse range of climate-related, political, or economic stresses. A third integrates physical event and the underlying causal characteristics of populations that lead to risk exposure and limited capacity of communities to respond (Adger, 1999; Cutter et al., 2000). Correspondingly, livelihood vulnerability to climate change can be usefully understood as an outcome of biophysical and social factors (Cutter et al., 2000). Biophysical climate change vulnerability refers to the level of exposure communities face from the physical

126

K.U. Shah et al. / Geoforum 47 (2013) 125–137

impacts of sea level rise, increase in sea surface and/or atmospheric temperatures. Climate-induced variability increases the vulnerability of rural livelihoods and reduces the ability of households to deal with risks, shocks and stresses (Prowse and Scott, 2008). Since these households typically have limited assets, they are at increased risk (exposure) and their ability to cope is restricted (Dulal et al., 2010b). Social vulnerability is partially the product of those factors that shape the susceptibility of communities to harm and those that govern their ability to respond. It also includes ‘‘place inequalities’’ - those characteristics of communities and the built environment, such as the level of urbanization, growth rates, and economic vitality - that contribute to the social vulnerability of particular places (Cutter et al., 2000). Kelly and Adger (2000) differentiate between ‘end-point’ and ‘start-point’ features of climate change vulnerability. End-point studies define vulnerability in terms of net impacts and inevitably frame adaptive options in terms of ‘‘fixes’’, often technological in nature, which will minimize particular impacts that have been projected. The ‘starting-point’ approach, which is employed in this study, defines vulnerability as a pre-existing state generated by multiple factors and processes, such as political or economic marginalization, that conditions the ability to respond to stress. Methods of vulnerability assessment take diverse approaches to systematically examining and integrating interactions between humans and their physical and social surroundings. Many approaches use indicators to characterize and quantify multidimensional issues, often combining diverse indicators into a single composite index of vulnerability. Vulnerability indices are constructed for three primary purposes. First, they offer a reference point for evaluating frameworks for development policy (Kelly and Adger, 2000; Eriksen and Kelly, 2007). Second, they can provide information for developing adaptation and mitigation plans (Gbetibouo et al., 2010). Third, they can provide a means of standardizing vulnerability measurement, thereby allowing comparison of different contexts. This in turn provides a means of setting priorities in resource allocations for adaptation and mitigation (Preston et al., 2011; Heltberg and Siegel, 2009). While indexes provide a useful means of comparing and evaluating different units of analysis (e.g. households, geographic regions), they must also be able to incorporate local, context-specific variables (Eakin and Bojorquez-Tapia, 2008). Without such flexibility, assessments can suffer from a lack of specific, local indicators that may be used to differentiate between vulnerability assessments based on the best quality information obtainable and the limited resources and expertise available (Shah and Rivera, 2007). At the household level, an index assessing livelihood vulnerability should provide an explicit indication of the capabilities, assets, and activities required for a sustainable means of living for the respective household (Chambers and Conway, 1992). A livelihood is considered sustainable when it can cope with and recover from shocks, and maintain or enhance its capabilities and assets, while not undermining the natural resource base. Livelihood vulnerability assessments can provide decision-making information at two adaptation and planning levels. First, among multilateral institutions they are increasingly being adopted and developed as an effective policy framework to address poverty and vulnerability, consistent with maximizing growth and development objectives. Second, among national social development and environmental protection agencies, they assist in developing community-specific plans that balance environmental, socio-economic and socio-cultural needs and rights of rural communities whose livelihoods are dependent upon natural resources (Arvai et al., 2006). The Sustainable Livelihoods Approach (SLA) is a conceptual tool used to improve understanding of the livelihoods of the poor. It looks at five types of household assets – natural, social, financial, physical and human capital, using multiple indicators to assess

exposure to natural disasters and climate variability, social and economic characteristics of households that affect their adaptive capacity, and current health, food and water resource characteristics that determine their sensitivity to climate change impacts (Chambers and Conway, 1992). Drawing upon the SLA, Hahn et al. (2009) developed a Livelihood Vulnerability Index (LVI) aimed at using household-level data to inform strategic community level planning. Having incorporated climate exposures and household adaptive practices into their approach, they tested the LVI in two communities in Mozambique, where it proved insightful in capturing differentials in community-level climate vulnerability. The ability of the LVI to draw out subtle yet critical differences in specific vulnerabilities (e.g. related to water, food etc.) is valuable in tailoring policies that can meet the needs of resource-dependent communities in the developing world. Although used in the southern African context of Mozambique, its structured approach provides a realistic framework for the developing country context in general. Drawing upon Hahn et al. (2009), this study explores the analytical utility of using the LVI to understand livelihood and climatic vulnerability in small-island developing-states (SIDS). It does so by applying the model in selected coastal wetland communities in Trinidad and Tobago. The communities were selected because they are directly and indirectly dependent upon the ecological services of wetlands, which provide important livelihood options in farming, hunting and fishing (Dugan, 1990; Dolan and Walker, 2004). They were also selected because rural communities in coastal, wetland areas of small island developing states are likely to be among the first to feel the impacts of climate change and therefore merit immediate attention. This study builds upon the approach developed by Hahn et al. (2009) in three significant ways. First, it incorporates local and indigenous knowledge into the selection of indicators. At the community level, local perceptions and experiences of climate extremes can help in identifying the factors that enable or constrain the ability of communities to respond, recover and adapt to climate change. As such, the approach incorporates local and traditional knowledge in ways that can inform more effective decision-making, planning and management in remote areas susceptible to climate change hazards. Second, the study starts from the premise that resilience and vulnerability are gendered by important norms in society. The empirical literature has shown that adaptation strategies are gendered by sector-specific employment, lower wages, and family care responsibilities (Enarson and Scanlon, 1999; Morrow, 1999). Compared to men, women and children are fourteen times more likely to die during disasters (Brody et al., 2008). Jankowska et al. (2012) found that climate change had varying levels of human health effects (e.g. malnutrition) in men, women, infants and young children in Mali. However, all women are not equally vulnerable because of capital asset differentials (Neumayer and Plumper, 2007). In order to reduce gender disparity in climate vulnerability, marginality needs be viewed through the power relations that produce the vulnerability in the first place (Arora-Jonsson, 2011). How men and women are impacted by, and respond to climate change is directly related to gender roles, relative socio-economic status and political power differentials (Kumar-Range, 2001). The social experience of climate change vulnerability and adaptation affirms, reflects, disrupts and otherwise engages gendered social relationships, practices and institutions (Enarson and Morrow, 1998). Finally, the study presents original empirical data that can be used to inform its assessment of the LVI. According to a recent evaluation by Preston et al. (2011), only 9% of the 45 climate change vulnerability mapping studies they addressed in their


study collected some form of primary empirical data. On the contrary, most were dependent on secondary data from various sources to populate their indicator measures. Invariably, data quality is likely to suffer in the latter approach. This study therefore evaluates the LVI on the basis of empirical observations and local needs; it tests the LVI in the field in a setting which is in critical need of such tools and validates the ability of the tool to provide insight for informing local climate change adaptation and mitigation planning.

2. Caribbean climate change vulnerability According to the Intergovernmental Panel on Climate Change’s (IPCC) Fourth Assessment Report, ‘‘small island states are likely to experience large impacts due to the combination of higher exposure, for example to sea-level rise and storm surge, and limited ability to adapt’’, (Schneider et al., 2007). A number of studies suggest that Caribbean islands are already experiencing climate change impacts, highlighting the observation of increased temperatures (McWilliams et al., 2005); rises in sea level (Kelman and West, 2009); severe weather conditions that threaten lives, property and livelihoods throughout the region (Pulwarty et al., 2010); and agricultural losses and ecosystem degradation (Fischer et al., 2005). Climate change impacts are expected to have a disproportionate effect on poor and vulnerable groups, such as children, the elderly, the sick and people living with disabilities (Chambers, 1989; Alexander, 1997; Dulal et. al., 2010). Others have also raised the concern that entire island communities will be permanently displaced by climate change (McNamara and Gibson, 2009). Rural communities along the coastal wetland areas of many Caribbean islands are likely to be among those highly vulnerable groups since they depend on natural resources for their livelihoods, be it agriculture, fishing and hunting and/or nature based tourism (Shah, 2011a). That said, considerable uncertainty exists about the long-term patterns of environmental variability and their likely impacts on the livelihoods of the poor (Brown and Crawford, 2009). Amongst the diverse consequences of Caribbean climate change are substantial impact on agricultural production and community-based tourism, consequently reducing the scope of poverty-reducing policies when there is majority smallholder agriculture and tourism activities (Slater et al., 2007). Climate change could therefore have direct implications for creating unsustainable livelihoods and or reducing the livelihood options of the rural poor (Brown and Crawford, 2009). As the reality of impending climate change ensues, the very social, cultural and economic well being of rural communities come into question. Such communities become the most vulnerable to the effects and impacts of climate change, particularly sea level rise, coastal inundation, loss of agricultural lands and fisheries. Coupled to this is the increasing onset of health impacts from vector borne diseases such as yellow fever and dengue fever caused by changes in temperature and seasonal durations (Dulal et al., 2010a). Climate-induced extreme weather events increase flooding in low-lying regions; cause loss of life and; loss of physical assets such as housing, water, crops and livestock. In these wetland communities, vulnerability is even more exacerbated because of political marginalization (Assan, 2009) and weak institutions (Osbahr et al., 2008). Political marginalization can have a negative effect when communities lack political power in numbers, financial influence, resources and votes (Shah, 2011b). Institutional conditions impact on local responses to climate shock, food security and poverty reduction as well as ability to diversify livelihoods (Osbahr et al., 2008). The 1994 Barbados Conference on Sustainable Development of Small Island Developing States called for ‘‘the development of vul-

127

nerability indices and other indicators that reflect the status of small island developing countries and integrate ecological fragility and economic vulnerability’’ (Hurley, 2010). However, as noted earlier, indicators aimed at measuring vulnerability to climate change vary widely, highlighting the need to develop and test analytical models that may be replicated in multiple field settings. 3. Research methodology The central focus of this study is on wetland communities in two geographic regions of Trinidad and Tobago. One is on the sheltered west coast of the island of Trinidad, the Caroni Swamp, while the other is on the eastern Atlantic coast of the island, the Nariva Swamp. The different socio-economic and developmental context of each area allows for spatial comparative analysis. The Caroni Swamp is located on the Gulf of Paria, 3.5 km from the Port of Spain. It is the largest mangrove swamp in Trinidad, accounting for 60% of the country’s total mangrove area and draining approximately 675 km2 in catchment area. its ecology is diverse, consisting of marshes, mangrove swamp, brackish and saline lagoons, and tidal mudflats. The swamp has been modified by attempted reclamation, and there is seasonal cultivation on the landward fringe. In terms of livelihoods, the Caroni Swamp provides an important source of income and livelihood from oyster and fish harvesting, hunting and ecotourism. As with any wetland, water is the driving force of wetland ecosystems and changes in flow can have severe consequences for the benefits that wetlands provide (Dugan, 1990). Surrounding villages still heavily depend on the swamp both within and around the park boundaries for fishing, crab catching, oyster harvesting and ecotourism. The swamp is surrounded by agricultural land and commercial rice farming continues to create negative environmental impacts through chemical/ fertilizer run-off and land clearing activities (Sookhoo, 1987). The second research site was located adjacent to the Nariva Swamp, which at seven thousand hectares, is the largest wetland in Trinidad and Tobago (Nathai-Gyan, 1996). Its ecosystem supports a diverse population of flora and fauna. Nariva swamp also offers recreation in the form of hunting, fishing, and eco-tourism. Furthermore, the swamp supports subsistence agricultural production, including rice and vegetable farming and natural sources of freshwater fish and conchs (Kacal, 1999; Carbonell et al., 2007). However, agricultural production by local residents, who do not have legal ownership of land, causes serious environmental damage to the swamp (Carbonell et al., 2007). Most of its water derives from a series of rivers in the central mountain range. The swamp is able to provide for a diverse faunal biodiversity because of the various distinct types of vegetation ranging from tropical forest, swamp forest, palm swamp forest, mangrove areas, marshland, and open waters (Bacon, 1996). It is listed as a Wetland of International Importance under the Ramsar Convention as of 1992 and in 2006 it became an Environmental Sensitive Area under the Environmental Management Act of Trinidad and Tobago. 3.1. Survey design and fielding Drawing upon the SLA and the Hahn et al. (2009) LVI, the research methodology was modified and refined on the basis of a series of consultations conducted with a panel of local stakeholders, climate vulnerability experts and researchers. The consultations and ensuing recommendations produced two important modifications to the LVI and the research methodology. First one more component – Housing and Land Tenure – was added to the LVI as a means of capturing the sensitivity of households to climate change. Second, the consultation process helped to refine the range of

128


questions being asked in the surveys and the potential range of interpretations of local responses. Community leaders recommended that several questions in the base model were not relevant in the Trinidad context. For example, there were no significant ‘‘water conflicts’’ among rural communities in Trinidad and it was likely that more dependable responses would be forthcoming if the question asked about ‘‘average number of floods/droughts in the past three years’’ rather than over the ‘‘last six years’’ as used by Hahn et al. (2009). Other questions such as ‘‘households reporting after conflicts’’ were not considered relevant at all in the local context, and were consequently removed. Table 1 summarizes the finalized major LVI, discusses the modifications made and the questions used to collect the data. 3.2. Sampling and data collection Gibbes et al. (2009) suggest that community dependence on wetlands tends to increase with physical proximity. Therefore rural village communities in closest proximity to each respective wetland area were selected for the study. These were Cacandee village on the southern edge of the Caroni wetlands and Cascadoux/Kernahan villages in the southern part of the Nariva wetlands. Prior to the survey, a reconnaissance tour of each village was conducted in order to design an appropriate sampling approach. Cacandee village consists of approximately 400 households (PADF, 2006). At 95% confidence level and ±10% interval, a minimum sample size of 78 households was required. The research entailed a systematic sampling approach, in which interviewers started at either end of the main road and at a random end of the four main branch roads, using a random number table to select the households. This continued until the minimum sample size was achieved. Kernahan and Cascadoux settlements are considered as a single community because of the geography of their historically converging development (village centers are about 4 km apart with increasing ribbon development between). They are also not marked by any distinct socio-cultural or socio-economic differences (Baptiste and Nordenstam, 2009). Kernahan and Cascadoux have 91 and 42 households respectively. At 95% confidence level and +10% interval, a minimum sample size of 56 households was required. For proportionality, the sample consisted of at least 38 households from the Kernahan and 18 from Cascadoux. The surveys were administered on weekends between May– July 2009. As recommended by community leaders, weekend surveying increased the chances of meeting household heads at home rather than during weekdays, when they were likely to be in the fields, fishing, at market or working outside of the community. Only the responses of the household heads were recorded. Cultural expectations dictated that the male be interviewed as the head of the household unless otherwise stated (Baptiste and Nordenstam, 2009). Each interview took between 20 and 40 min to complete.

components is viewed as having an equal contribution (i.e. balanced weight) to a community’s overall vulnerability (Sullivan et al., 2002). There are four steps in calculating each LVI. Step 1 involves transforming the raw data into appropriate measurement units, such as percentages, ratios and indices (see ‘‘Units’’ in Table 2). Step 2 is the standardization of each sub-component since they are measured on different scales. This is necessary in order to combine all measures in a single LVI index. To standardize a main component, the quotient of the difference between the actual score and the minimum value obtained from the total sample (i.e. both communities) and the difference between the maximum and minimum values from the total sample was calculated. In Step 3, the average of the standardized scores of each main component is calculated, giving a final score for each main component. For example, the average of the scores for sub-components – ‘‘average receive: give ratio’’, ‘‘average borrow: lend ratio’’ and ‘‘% households that have not gone to government for assistance in 12 months’’ gives the score for the major component, ‘‘Social Networks’’. Lastly, Step 4 combines the weighted averages of all the major components to generate the LVI score. The weights of each main component are determined by the number of indicators of which it is comprised. This ensures that all main components contribute equally to the overall LVI (Sullivan et al., 2002). The LVI is scaled from 0 (least vulnerable) to 0.5 (most vulnerable). 3.3.2. Model 2 The IPCC definition characterizes vulnerability (to climate change) as a function of a system’s exposure and sensitivity to climatic stimuli and its capacity to adapt to their (adverse) effects, which corresponds to outcome (or end-point) vulnerability, but it does not provide a clear definition of these attributes or the relationship between them. Using the same data, the LVI is based on the IPCC vulnerability definition of adaptive capacity, exposure and sensitivity (and as described by Hahn et al. (2009)). This entails grouping the eight major components into each of these three categories (see Fig. 1). The results of this analysis are reported in Table 3. Each of the three IPCC factors is calculated based on the equation:

CF d ¼

Pn i¼1 wMi M di P n i¼1 wMi

where CFd is an IPCC-defined contributing factor (exposure, sensitivity, or adaptive capacity) for community d, Mdi are the major components for community d indexed by i, WMi is the weight of each major component, and n is the number of major components in each contributing factor. Once exposure, sensitivity, and adaptive capacity were calculated, the three contributing factors were combined using the following equation:

LVI IPCCd ¼ ðed ad Þ Sd

Three sets of analysis were undertaken: (1) calculation of a balanced weighted average LVI (referred to as model 1); (2) calculation of a LVI based on the IPCC framework (IPCC, 2007; (Schneider et al., 2007) (referred to as model 2); (3) gendered analysis utilizing both models 1 and 2. Additionally, t-tests were used to compare the mean scores of Nariva and Caroni’s main components and sub-components.

where LVI–IPCCd is the LVI for community d expressed using the IPCC vulnerability framework, e is the calculated exposure score for community d (equivalent to the Natural Disasters and Climate Variability major component), a is the calculated adaptive capacity score for district d (weighted average of the Socio-Demographic, Livelihood Strategies, and Social Networks major components), and s is the calculated sensitivity score for district d (weighted average of the Heath, Food, and Water major components). The LVI–IPCC was scaled from 1 (least vulnerable) to +1 (most vulnerable).

3.3.1. Model 1 In this approach, several subcomponent indicators combine into each of the eight main components (Table 1). No a priori assumption is made about the importance of each indicator or main components in the overall sum. In this way, each of the eight main

3.3.3. Gender based analysis Three sets of comparisons were conducted: Female headed versus male headed households for the total sample, female headed versus male headed households for each wetland community (Caroni and Nariva) and female headed households in Caroni

3.3. Data analysis

129

K.U. Shah et al. / Geoforum 47 (2013) 125–137 Table 1 Design of the Livelihood Vulnerability Index (LVI). Main component/sub-components

Status in LVI

Explanation of sub-component

Survey question

Explanatory notes

Socio-demographic Profile (1) Dependency ratio

Ratio of the population 65 years of age to the population between 19 and 65 years of age

Please list the ages and sex of every person who eats and sleeps in this house?

This data was collected to facilitate gender based comparative analysis. Literacy rate has been over 80% since the 1980s, so a more refined measure is needed. Completing primary school indicates basic competency in reading, writing and arithmetic

(2) % of female headed households

Percentage of households where the primary adult is female. If male head is away from home >6 months per year, the female is counted as the head Average of ages of all female head of households Percentage of households where the head reports not finishing primary school % of households with at least one member requiring daily care because of age, physical or mental condition, illness or disability

Are you the head of the household?

Percentage of households reporting at least one member working outside the community for their primary livelihood

How many people in your household go to a different community to work?

Percentage of households reporting these activities as the main source of income

Do you or anyone in your household grow crops, raise animals? Do you or anyone in your household collect something from the forest, swamp, rivers of surroundings? Do any of these activities account for your main income? Same as for indicator (7) above

(3) Avg. age of female headed household (4) % Household heads did not attend school

New

(5) % Households with members needing dependent care

New

Modified

Livelihood strategies (6) % Household without members working outside the community

(7) % Households mainly income dependent on agriculture/ fishing/hunting

Modified

(8) Avg. agricultural livelihood diversity index

(9) % Households without non agricultural livelihood income contribution

New

Social networks (10) Avg. receive: give ratio

Modified

(11) Avg. borrow: lend ratio

(12) % Households that have gone to government for assistance in last 12 months Health (13) Avg. time to health facility

The inverse of (the number of agricultural livelihood activities + 1) reported by a household Percentage of households reporting livelihoods other than agriculture/ fishing/hunting as the main source of income

Did you ever go to school? If yes, did you finish primary school? Do any members of your household require daily care because of age, physical or mental condition, illness or disability? Extended to include hunting and fishing. ‘‘Main source of income’’ is defined as maintaining the household for at least 6 months of the past year or > 50% of total yearly income. Added to address the potential limitation cited by Hahn (2009) in the Mozambique LVI

Do you or anyone in your household work in an activity not related to growing crops, rearing animals or collecting something from the forest, swamp or rivers? Do any such activities account for your main income? ‘‘Help given to another household’’ rather than ‘‘anyone’’ reduced ambiguity and improves household-to-household data accuracy. Government assistance is received from the local government office and officials rather than community leaders. Indicator positively worded – having asked for assistance indicates heightened vulnerability

Ratio of (the number of types of help received by a household in the pat month +1) to (number of types of help given by household to another household in the past month +1)

Percentage of households that reported that they have asked their local government for any assistance in the past 12 months

Did you borrow any money from relatives or friends in the past month? Did you lend any money to relatives or friends in the past month? In the past 12 months, have you or another in your household gone to your local government office/ official for help?

Average time to travel to the nearest health facility

How long does it take you to get to a health facility?

Ratio of a household borrowing money in the past month to a household lending money in the past month Modified

What is your age?

Question (16) modified to highlight the major insect vector health threat in Trinidad, Dengue (continued on next page)

130


Table 1 (continued) Main component/sub-components

Status in LVI


Survey question

Explanatory notes Hemorrhagic Fever, instead of malaria

(14) % Households with members suffering chronic illness (15) % Households where member missed work/school in last 2 weeks due to illness (16) Average dengue exposure prevention index

Modified

Food (17) % Households primarily dependent on self farmed food.

New

(18) Average crop diversity index

Percentage of households reporting at least 1 member with chronic disease Percentage of households reporting at least one member who missed school or work due to illness in the last two weeks Months reported exposure to dengue⁄ Owning at least one bednet indicator (yes = 0.5; no = 1)

Is anybody in your family chronically ill (get sick very often?)

Percentage of households that get their food primarily from their personal farms

Questions on ‘‘struggle to find food’’, ‘‘saving seeds’’ and ‘‘saving crops’’ were omitted as less relevant to the Trinidad context. Where does your family get most of its food? What kind of crops does your household grow? Added to address access of Do you trade the food you grow households to the variety of foods with others for different food? Do needed for a balanced, healthy diet. you sell the crops your grow for money to buy other food goods? Is most of your food obtained from Added to emphasize importance of hunting and/or fishing? fishing, oyster catching in the coastal wetland context.

The inverse of (the number of crops grown by a household + 1) Percentage of households unable to trade self grown crops

Has anyone in your household been too sick in the past two weeks to miss work or school?

(19) % Households that do not sell/ barter crops for other food supplies

New

(20) % Households depending significantly on fishing/ hunting for food

New

Percentage of households that get their food primarily from what is hunted or fished

Water (21) % Households without pipe borne water

New

Percentage of households not receiving water through the public water system

Do you receive pipe bourne water to your home?

Percentage of households obtaining water from wells, rainwater, springs and other means apart from the public system Percentage of households reporting that water is not available at their primary water supply Average water supply security per household

Where do you collect your water from?

(22) % Households utilizing natural water system

How many days for the month is water unavailable from your main supply source? Do you store water? How many days of water supply do you have stored to meet the needs of your household?

(23) Avg. days without regular water supply per month

New

(24) Average days supply stored per household

Modified

Housing and land tenure (25) % Houses with weak storm resistant construction (wood, mud)

New

Percentage of houses that will be unable to withstand a severe climatic event (e.g. hurricane winds)

Can you describe the materials your house is built of including ground, walls and roof?

(26) % Houses not elevated by posts/high ground to avoid floods (27) % Households without ownership of the lands they live on

New

Percentage of houses that will be unable to withstand storm surges and floods Percentage of households that can be removed from lands on which they presently reside

Not in questionnaire. Directly observable

New

Question on ‘‘water conflicts’’ removed since not context relevant. Adjusted from daily basis to monthly basis to take into account, the public water system arrangements made for the community. Modified to measure in ‘‘days’’ instead of ‘‘liters of water’’ to remove ambiguity about ‘‘size of water containers’’

New Main Component. Question (25) supported by Cutter et al. (2003); Question (26) supported by Rygel et al. (2006); Question (27) supported by Dulal et al. (2010a)

Do you own a deed for the land you reside on? Do you rent the land you reside on? Are you a regularized/ unregularized squatter?

Natural disasters and climate variability (28) Average number of floods/ droughts in past 3 years

Modified

Total number of floods, droughts, reported by households in the past 3 years

How many times has this area been affected by flooding/drought between 2007–2009

(29) % Households with losses to physical assets (house/ machinery) due to flooding

New

(30) % Households with injury or

Modified

Percentage of households that have suffered losses of physical assets that causes loss/damage of livelihood Percentage of households that

In the last 3 years, what physical assets have you lost or been severely damaged due to flooding/ storm events? Was anyone in your household

Asks for recall on past 3 years instead of 6 years to improve response accuracy. Added to emphasize importance of physical assets to rural households. Focused on ‘‘households’’ rather than family members’’ and narrows to a specific time period (the last three years) to reduce recall bias

131

K.U. Shah et al. / Geoforum 47 (2013) 125–137 Table 1 (continued) Main component/sub-components death from natural disasters in last 3 years (31) Mean standard deviation of monthly avg. of avg. max. daily temperature (1999–2005) (32) Mean std. deviation of monthly avg. of avg. minimum daily temperature (1999–2005) (33) Mean std. dev. of monthly avg. precipitation

Status in LVI


Survey question

reported death or injury as a result of severe/prolonged weather conditions/events Standard deviation of the average daily maximum temperature by month between 2000–2005 was averaged for each area Standard deviation of the average daily minimum temperature by month between 2000–2005 was averaged for each area Standard deviation of the average monthly precipitation between 1999–2005 was averaged or each area

injured in any severe weather events in the last 3 years?

Explanatory notes

Data obtained from the Trinidad and Tobago Meteorological Services



Notes: All comparisons are to the LVI methodology described in Hahn et al. (2009); unless otherwise stated as ‘‘new’’ or ‘‘modified’’ indictors are adapted/adopted from Hahn et al. (2009); further notes on limitations of various indictors and original references for some indicators can be found in Hahn et al. (2009). All indicators are structured to accommodate data interpretation as, the higher the value being the higher the vulnerability.

versus those in Nariva. To compare indicator scores between two groups (e.g. male headed versus female headed households) the raw data was transformed into the units used in producing the LVI (mathematical transformations are noted in Table 1 and as described by Hahn et al. (2009)). For example, if a household farmed 2 crops the raw score is 2, which is then transformed into the average agricultural livelihood index by the function 1/(2 + 1) = 0.33). These (transformed) mean scores were compared to identify significant differences. The results are reported in Tables 4–6.

4. Results and analysis The aggregate scores presented in Table 2 (Model 1) suggest that Nariva was more vulnerable to climate change than Caroni (LVIs of 0.41 compared to 0.36). The results also suggest that Nariva was more vulnerable in terms of socio-demographics (0.29), health (0.46), water (0.56) and natural disaster and climate variability (0.37). Caroni on the other hand was more vulnerable in terms of livelihood strategies (0.47), social networks (0.32) and housing and land tenure (0.51). Both communities exhibited the same level of food vulnerability (0.36). The highest levels of vulnerability in Nariva were documented in relation to water resources/access (0.56) and housing and land tenure (0.51). On the contrary, the lowest levels of vulnerability were measured in relation to socio-demographic profile (0.29) and social networks (0.32). When the main components were reviewed by sub-components (i.e. indicators), Nariva was most vulnerable in terms of exposure to dengue (0.89), lack of access to employment outside of the community (0.75) and resource dependence on agriculture/hunting/fishing (0.71). By main components, Caroni communities were found to be highly vulnerable in terms of housing and land tenure (0.53) and livelihood strategies (0.50). By subcomponents (i.e. indicators), they were least vulnerable in terms of socio-demographic profile (0.27) and water resources/access (0.30). By comparison of main component scores (Table 2), the widest differentials between the communities were water resources/access (Nariva is +0.27 more vulnerable than Caroni) and; health (Nariva is +10 more vulnerable than Caroni). Comparison of means (p < 0.05 level) for Nariva and Caroni found statistically significant differences for 11 of the 33 sub-components in the LVI. Vulnerability was statistically different in Nariva versus Caroni with respect to: the percentage of households with no members working outside the community (75% vs. 68%); average agricultural livelihood diversity index (0.20 vs. 0.47); the

percentage of households without non-agricultural income (21% vs. 12%); percentage of households going for government assistance (57% vs. 77%); average time to a health facility (64.9 vs. 34.7); all four indicators comprising main component – water (0.56 vs. 0.30 overall); the average number of floods/droughts in the past 3 years (8.62 vs. 9.08) and households that sell/trade their crops or catch (73% vs. 58%). Turning now to Model 2 (the IPCC-LVI), documented in Table 3, the results suggest that both Nariva and Caroni communities shared a very similar degree of vulnerability (0.02 and -0.03 respectively) that could be described as ‘‘mid range’’ (on the 1 to +1 scale). The model does however suggest that Nariva was more acutely sensitive than Caroni (0.47 versus 0.38) while Caroni was less exposed than Nariva (0.29 versus 0.37 respectively). The model also suggests that sensitivity in Caroni was most affected by housing and land tenure while in Nariva, it was most affected by water resources/access. Adaptive capacity of both Caroni and Nariva communities was most affected by their livelihood strategies (0.50 and 0.47 respectively).

4.1. Gender-based analysis Table 4 documents the results of the gender-based comparisons of livelihood vulnerability between female (n = 44) and maleheaded (n = 90) households (in actual measurement units). Table 5 presents male to female comparisons of standardized scores. Overall, the results suggest little difference in vulnerability between male and female-headed households (female-headed households +0.02). Female-headed households had more vulnerable sociodemographic profiles than male-headed households (+0.06) but were less vulnerable than male-headed ones in terms of housing and land tenure (0.04). The results (Table 5) also suggest a 6±0.02 difference in male-headed and female-headed scores for health and food, livelihood strategies, social networks and natural disaster/climate change variability. That said, the vulnerability of female-headed households was significantly greater (p < 0.05) than male-headed households in terms of members requiring dependent care; having lower than average agricultural livelihood diversity index (i.e. practices less types of livelihoods) and being more income dependent on hunting/fishing. In Caroni, the findings presented in Table 5 suggest that femaleheaded households were more vulnerable in terms of socio-demographic profile, social networks and health (+0.06), food (+0.03) and natural disasters/climate variability (+0.02); less vulnerable than male-headed households on shelter and land tenure (0.08)

132


Table 2 LVI results for Caroni and Nariva communities. Major component

Sub-component

Units

Caroni wetland communities

Nariva wetland communities

Actual value

Actual value

Socio-demographic profile

Standardized

Standardized

Max. value in both communities

Min. value in both communities

Dependency ratio % of female headed households Avg. age of female headed household % Household heads did not complete school % Households with members needing dependent care

Ratio %

1.49 35

0.27 0.22 0.35

1.69 30

0.29 0.26 0.30

6.0 100

0.2 0

1/Years

0.02

0.24

0.02

0.28

0.04

0.01

%

29

0.29

34

0.34

100

0

%

26

0.26

27

0.27

100

0

% Household without member(s) working outside community % Households mainly income dependent on agriculture/fishing/hunting Avg. agricultural livelihood diversity index % Households without non agricultural livelihood income contribution

%

68

0.5 0.68

75

0.47 0.75

100

0

%

73

0.73

71

0.71

100

0

1/No. of livelihoods

0.26

0.47

0.23

0.20

100

0

%

12

0.12

21

0.21

100

0

Avg. receive: give ratio Avg. borrow: lend ratio % Households that have gone to government for assistance in last 12 months

Ratio Ratio %

1.41 1.06 77

0.36 0.25 0.06 0.77

1.3 1.16 57

0.32 0.23 0.16 0.57

5 2 100

0.2 1 0

Avg. time to health facility % Of households with members suffering chronic illness % Households where member missed work/ school in last 2 weeks due to illness Average dengue exposure prevention index

Min %

34.7 18

0.36 0.21 0.18

64.9 20

0.46 0.64 0.20

90 100

20 0

%

10

0.10

11

0.11

100

0

Months Bednet indicator

6.73

0.93

6.56

0.89

7

3

% Households primarily dependent on self farmed food. Average crop diversity index % Households that do not sell/barter crops for other food supplies % Households primarily dependent on fishing/ hunting

%

32

0.36 0.32

43

0.36 0.43

100

0

1/number of crops

0.24

0.28

0.24

0.28

0.5

0.14

%

42

0.42

27

0.27

100

0

%

40

0.40

45

0.45

100

0

% households without pipe borne water % households utilizing natural water system Avg. days without water supply per month Average days supply stored per household

%

46

0.3 0.46

100

0.56 1.00

100

0

%

29

0.29

68

0.68

100

0

Days

5.69

0.28

9.71

0.5

10

4

1/Days

0.63

0.18

0.39

0.04

2

0.33

% Houses of non storm resistant construction e.g. wood, mud) % Houses not flood resistant (e.g. built on stilts, high foundation).

%

38

0.53 0.38

55

0.51 0.55

100

0

%

60

0.60

52

0.52

100

0

Livelihood strategies

Social networks

Health

Food

Water

Housing and Land Tenure

133

K.U. Shah et al. / Geoforum 47 (2013) 125–137 % Households without legal ownership of the land they live on.

%

62

Natural disasters and climate variability

0.62

46

0.29 Average number of floods/ droughts in past 3 years % Households with losses to physical assets (house/ machinery) due to flooding % Households with injury or death from natural disasters in last 6 Years Mean standard deviation of monthly avg. of avg. max. daily temperature (1999– 2004) Mean std. deviation of monthly avg. of avg. minimum daily temperature (1999–2004) Mean std. dev. of monthly avg. precipitation

0.46

100

0

0.37

Count

9.08

0.69

8.62

0.54

10

7

%

23

0.23

23

0.23

100

0

%

4

0.04

11

0.11

100

0

°C

0.6

0.4

0.7

0.6

0.9

0.4

°C

0.5

0.2

0.8

0.5

1.4

0.3

mm

55.4

0.17

72.3

0.25

224.4

21.2

Livelihood vulnerability index

and livelihood strategies (0.02) and; equally water resource/access vulnerable. Vulnerability of female-headed households was significantly greater (p < 0.05) than male-headed households in terms of lower average agricultural livelihood diversity index and dependence on income from hunting/fishing. In Nariva, femaleheaded households were more vulnerable on the sub-components socio-demographic profile (+0.06) and less vulnerable than maleheaded households in terms of housing and land tenure (0.04). There was a 6±0.02 difference between male and female-headed households on all other main components. Vulnerability of female-headed households was significantly greater (p < 0.05) than male-headed households in terms of households with members requiring dependent care; households primarily dependent on self farmed food; and households not selling/bartering their crops to obtain other food supplies. Table 6 summarizes the results of gendered analysis using the IPCC-LVI model. Again, using this methodology, the results do not suggest a significant difference in overall vulnerability between male and female-headed households. However, in Nariva, of the three IPCC contributing factors (adaptive capacity, sensitivity and exposure), female-headed households exhibited more sensitivity (+0.07) than male-headed households. Differences between male and female households in Caroni for each IPCC contributing factor were negligible.

5. Discussion At first glance, the preceding suggests that vulnerability differences between the two communities were minimal. However, upon closer inspection, the analysis reveals a number of subtle yet important differences between Nariva and Caroni. These are now discussed in greater detail. According to the findings, Caroni had a smaller range of agricultural livelihoods, more households with members working outside the community and more households with some dependence on non-agricultural livelihood income. Over the last generation, households have continually moved in a direction of specializing in singular livelihoods (e.g. oyster gathering or vegetable gardening), usually the one with the most income potential and/or relying on the available household manpower available. This is tied to the

0.36

0.41

increases in more household members taking up artisan, merchandizing, skilled livelihoods (e.g. mechanics, construction, taxis) than before and taking employment in nearby towns. In contrast to Caroni, Nariva has experienced less economic change, and remains geographically isolated. Although some household members travel to towns for daily employment, returning to their villages, the most prominent external employment is from field labor in the numerous oil and gas fields of South-Eastern Trinidad. Populations engaged in these activities can stay outside the villages for months on end, highlighting important implications for climate adaptation planning (see below). The findings also suggest that more Caroni households went to local government offices for assistance in the past year than Nariva households. Documenting political participation on the basis of local government visits provides an important way of measuring the extent to which vulnerable populations have developed a relationship with agencies responsible for providing social assistance. It also helps to illustrate the ways in which local populations perceive the potential costs and benefits of engaging with local government. In Caroni, for instance, many households reported that they approached local government officials because they perceived rewards for their support in recent elections. Similarly, some Nariva households suggested that they did not go to local government offices because they did not support the ruling political party in the last round of local government elections. The findings from Nariva suggest higher vulnerability to health shocks than Caroni, indicating that the average time of travel to health facilities is an important determinant of vulnerability. According to the data, it took nearly twice as long to travel to the nearest health facility in Nariva as it did in Caroni, reflecting the lack of serviceable roadways connecting the Nariva communities to the nearby towns. At night, the lack of street lighting makes the situation even more dire. Health vulnerability therefore reflects a lack of viable transportation options as opposed to a lack of healthcare facilities. Roads, however, were not the only infrastructural factors affecting vulnerability to climate-induced health shocks. Another was access to clean drinking water. According to the study, access to pipe-borne water was limited in Caroni (50% of households had access) and non-existent in Nariva. The Water and Sewerage Authority provides truck borne water to Kernahan/Cascadoux twice a

134


week, at which time households must fill their rain barrels and containers. Ponds and swamp irrigation provide water for crops and livestock. Rainwater and wells in the dry season supplement this supply. Water vulnerability was therefore highly dependent on the national water resource authorities and infrastructure development plans for the area. Inadequate transportation also affected the ability of Nariva households to earn a living from fishing and agriculture. According to the study, nearly three out of four Nariva households derived some form of income from the sale of produce/catch. However, their income was highly dependent on selling at the roadside, where the price is typically poor. In Caroni by contrast, better road access meant that households were able to transport their goods in local urban centers, where they can command premium prices.

Fig. 1. Grouping of the eight major components into the IPCC LVI.

Table 3 Model 2: IPCC-LVI results. IPCC Contributing Factors Caroni

Major LVI Components

Major Component scores

Nariva

Number of sub-components per major component Caroni

Adaptive Capacity Livelihood strategies Social networks Sensitivity Food Water Housing and land tenure

Socio-demographic profile 4 3

5 0.50 0.36

0.27 0.47 0.32

0.29

0.37

0.36

Health 4 4 3

4 0.36 0.30 0.53

0.36 0.36 0.56 0.51

0.46

0.38

0.47

0.29

0.37

0.29

0.37

Exposure


6

Final IPCC weighted LVI scores

0.03

0.02

IPCC Contributing Factors scores

Nariva

Table 4 Gender based Comparison of LVI results (actual measurement units). Household heads

Dependency ratio % Household heads did not complete school % Households with members needing dependent care % Household without member(s) working outside community % Households mainly income dependent on agriculture/fishing/hunting Avg. agricultural livelihood diversity index % Households without non agricultural livelihood income contribution Avg. receive: give ratio Avg. borrow: lend ratio % Households that have gone to government for assistance in last 12 months Avg. time to health facility % Of households with members suffering chronic illness % Households where member missed work/school in last 2 weeks due to illness Average dengue exposure prevention index % Households primarily dependent on self farmed food. Average crop diversity index % Households that do not sell/barter crops for other food supplies % Households primarily dependent on fishing/hunting % Households without pipe borne water % households utilizing natural water system Avg. days without water supply per month Average days supply stored per household % Houses of non storm resistant construction e.g. wood, mud) % Houses not flood resistant (e.g. built on stilts, high foundation). % Households without legal ownership of the land they live on. Average number of floods/droughts in past 3 years % Households with losses to physical assets (house/machinery) due to flooding % Households with injury or death from natural disasters in last 6 years Mean standard deviation of monthly avg. of avg. max. daily temperature (1999–2004) Mean std. deviation of monthly avg. of avg. minimum daily temperature (1999–2004) Mean std. dev. of monthly avg. precipitation *

Comparisons performed with actual values.

Units*

Ratio % % % % 1/No. of livelihoods % Ratio Ratio % Min % % Months Bednet % 1/no. of crops % % % % Days 1/Days % % % Count % % °C °C mm

Total sample

Caroni

Male

Female

Male

Female

Nariva Male

Female

1.58 31 21 72 76 0.24 18 1.32 1.12 66 45.78 17 12 6.65 36 0.24 41 36 69 44 7.53 0.58 47 57 58 8.93 22 4 0.65 0.48 63.9

1.57 32 36 68 66 0.27 11 1.46 1.08 75 50.57 23 7 6.68 39 0.24 25 55 68 45 7.05 0.61 43 57 50 8.80 25 11 0.65 0.48 63.9

1.48 27 22 71 78 0.25 14 1.37 1.09 75 31.3 16 12 6.73 35 0.25 43 29 45 29 5.8 0.63 43 63 63 .14 22 2 0.50 0.50 55.4

1.53 33 33 63 63 0.28 7 1.48 1.02 81 41.3 22 7 6.74 26 0.23 41 59 48 26 5.48 0.73 30 56 5 8.96 26 7 0.50 0.50 55.4

1.71 36 21 74 72 .22 23 1.24 1.15 54 64.7 18 13 6.55 36 0.24 38 44 100 64 .7 0.42 51 4 51 8.67 23 8 0.70 0.8 72.3

1.64 2 41 76 71 0.24 18 1.43 1.18 65 65.3 24 6 6.59 59 0.25 12 47 100 76 9.53 0.43 65 5 35 8.53 24 18 0.70 0.80 72.3

135

K.U. Shah et al. / Geoforum 47 (2013) 125–137 Table 5 Gendered model 1 LVI comparisons (standardized scores). Main component

Household heads Males (n = 90)

Females (n = 44)

Caroni: Males (n = 51)

Caroni: Females (n = 27)

Nariva: Males (n = 39)

Nariva: Females (n = 17)

Socio-demographic profile Livelihood strategies Social networks Health Food Water Shelter and land tenure Natural disasters and climate variability

0.25 0.49 0.36 0.39 0.33 0.54 0.54 0.32

0.31 0.50 0.36 0.41 0.35 0.53 0.50 0.33

0.24 0.51 0.36 0.37 0.35 0.46 0.56 0.32

0.30 0.49 0.42 0.43 0.38 0.46 0.48 0.34

0.25 0.49 0.36 0.39 0.33 0.54 0.54 0.32

0.31 0.50 0.36 0.41 0.35 0.53 0.50 0.33


0.42

0.44

0.39

0.41

0.40

0.41

Notes: ‘‘% of female headed households’’ and ‘‘Average age of female household head’’ are questions not used in assessing ‘‘All males’’ Caroni males’’ and ‘‘Nariva males’’ and do not form part of sub-component assessments when a comparative analysis is undertaken between male and female.

Table 6 Gendered model 2 LVI comparisons (standardized scores). Main component

IPCC contributing factor

Household heads Males (n = 90)

Females (n = 44)

Caroni: Males (n = 51)

Caroni: Females (n = 27)

Nariva: Males (n = 39)

Nariva: Females (n = 17)

Socio-demographic profile Livelihood strategies Social networks

Adaptive capacity

0.36

0.39

0.36

0.35

0.35

0.39

Health Food Water Shelter and land tenure

Sensitivity

0.45

0.45

0.43

0.44

0.45

0.38


Exposure

0.32

0.33

0.32

0.34

0.32

0.33

0.02

0.03

0.02

0.01

0.01

0.02


Turning now to the impact of housing and land tenure on household vulnerability, the results suggest that levels of vulnerability in Caroni and Nariva were similar but based on very different economic and structural determinants. For one, the material used for building houses in the two communities varied in relation to underlying ecological dynamics. On the alluvial plains of the Caroni River, for instance, close proximity to clay brick factories facilitated the construction of stronger homes that were less vulnerable to damage caused by windstorms and other climatic events. By virtue of their relative isolation, Nariva households were more dependent on timber from local forests. In terms of land ownership, the empirical findings suggest that many households in both communities lacked basic forms of land title, which is also an important indicator of vulnerability (cf. Hahn et al., 2009). In Caroni, the surveys reveal that three out of four landless households were renting land, while in Nariva, 3=4 of landless households were classified as illegal squatters seeking ‘‘regularization’’ by land settlement authorities. Again, the differences are subtle, but they reveal important structural factors that contribute to our understanding of vulnerability. In Nariva, for instance, the illegal status of landless households suggests a more pronounced form of vulnerability, in which access to formal safety nets and entitlements was limited. Lastly, while a growing body of literature points to the impact of gender on vulnerability, there was little evidence to suggest that female-headed households were significantly more vulnerable than male-headed households. The most significant difference on average and within each study site was documented in relation to socio-demographic profile. Here one of the most important findings was that female-headed households were 15% more likely to

have a household member requiring dependent care than maleheaded households. Moreover, the findings suggest, female-headed households had about 6% more occurrences of members suffering chronic illness than male-headed households. Whether higher dependency ratios were due to the fact that the male breadwinner was no longer present or whether they caused the male family member to abandon the family, the findings help to illustrate the crucial relationship between gender, household responsibility and vulnerability. 6. Conclusions This paper has explored the analytical utility of using the livelihood and vulnerability index (LVI) developed by Hahn et al. (2009) to understand local vulnerability to climate and environmental change. In so doing, it makes two significant contributions to the development of our understanding of vulnerability indicators. First, it provides an assessment tool that can be tailored to the local context of environmental vulnerability and policy analysis. Second, it provides a systematic means by which researchers can compare and contrast livelihood vulnerability in different socio-ecological settings. As noted in the introduction, there are three essential qualities that any vulnerability assessment tool should aim to achieve (Hooke and Pielke, 2000). First, it should provide a range of indicators that can be used and replicated in different empirical field settings. Building upon the model developed by Hahn et al. (2009), the comparison of Caroni and Nariva highlights the merits and challenges of comparing livelihood vulnerability. On one hand, the methodology incorporates and compares variables that have

136


important bearing on people’s vulnerability, including assets, infrastructure and access to local government. On the other, it highlights the challenge of incorporating new factors, such as the building materials used in the construction of local housing and the impact of voting behavior on people’s access to local government. Second, it should provide a means of incorporating local, contextual factors affecting vulnerability to environmental change. As noted earlier, local stakeholders played a significant role in framing and designing the research. This approach adds a degree of tailoring to the particular context of rural agro/natural resource-dependent communities that proves useful in simultaneously providing an assessment framework that allows comparisons of different locales and a precision of insight that highlights critical differentials in vulnerability that have to be considered in decision-making. Through this consultative process several indicators previously used by Hahn et al. (2009) were revised for the Trinidad and Tobago context, some dropped as less relevant and new indicators added (see Table 1). Third the assessment should inform policy. The LVI fulfils the objective of providing a benchmarking baseline that can be used to inform macro-policy development and resource allocation decisions and micro-action planning for mitigation and adaptation within communities. The LVI provides a useful evaluation tool that gathers data on critical indicator variables. It also provides a useful means of undertaking spatial comparisons of different communities and gender-based comparisons at the household level. Finally, it identifies new variables (e.g. voting patterns and access to local government) that may be addressed in future policy discussions about improving people’s access to services aimed at reducing vulnerability to environmental change. However, assessment methodologies are of little use if they cannot be easily communicated and understood. Here the experience in this study suggests that the revised LVI required minimal training on the part of agricultural and rural extension officers whose offices can conceivably be used in future data collection. Further, since the intention was to develop an assessment tool that is accessible to a diverse set of users in resource-poor settings, the LVI formula provided a simple approach of applying equal weights to all major components, the advantage being that the weighting scheme can be adjusted by future users. There are several methods to combine indicators including benchmarking against a particular standard, weighting methods based on statistical testing, expert opinions and stakeholder consultations. Other practical benefits of this approach are the minimal time and effort burden on respondents and; the minimization of potential sources of bias (by how the indicators are constructed and by primary data collection). As with any assessment index methodology, users must be cautious about applying the approach and interpreting the results. Indices by nature provide an ‘‘average’’ measure of phenomena that can mask subtle but important data. Also, while this LVI is fundamentally driven by prevailing climate change vulnerability theory and efforts have been made to include indicators for which data can be collected in the future, the index cannot provide medium or long term predictions as would more complex simulation models. Lastly, the index method brings together a strong, coherent set of indicators that represents complex phenomena and multidimensional realities of vulnerability. However, it is possible that other types of indicators could also add confidence to such an LVI, including for example ‘‘attitudes and values’’ to climate change vulnerability (Mustafa, 1998). The study also provides a practical means of measuring and comparing gendered forms of livelihood vulnerability (AlvarezCastillo, 2001; Morrow and Enarson, 1996). In so doing, it identifies a number of indicators suggesting that female-headed households are marginally (although not significantly) more vulnerable than

male-headed households. Looking forward, future studies could refine the LVI in ways that identify and incorporate more specific indicators of gender-based differences in livelihood vulnerability. The results of the study can provide a useful guide for future research. That said, one of the major challenges of developing vulnerability indicators is the unit of analysis (Hinkel, 2011). Here it is crucial that local interpretations and assessments take into account the wider political, institutional, economic and social factors that determine vulnerability, as expressed by the concept of ‘‘contextual vulnerability’’ (O’Brien et al., 2007; (Preston et al., 2011)). They not only determine vulnerability but also influence adaptation outcomes. The institutional rules that govern the ways in which individuals and households respond to environmental shocks and changing trends, for example, also play a large role in adaptive capacity (Dulal et al., 2010b). A final implication that stems from the study is the possibility of using the assessment methodology as a means of ranking or weighting the indicators (sub-components) and the eight main components in ways that can be used in designing future programs and policy. The LVI provides a combination of objective (i.e. statistical indicators) and subjective (through stakeholder inputs) indicators that can be used to frame and evaluate policies aimed at building livelihoods and resilience to climate change. That said, managerial and institutional capacity cannot be taken for granted, highlighting the need to support the ability of government and civil society actors to invest in adjustment (coping), resistance and recovery (resilience).

References Adger, W.N., 1999. Social vulnerability to climate change and extremes in coastal Vietnam. World Development 27, 249–69 (Lars Otto Naess). Adger, W.N., Dessai, S., Goulden, M., Hulme, M., Lorenzoni, I., Nelson, D.R., Otto Naess, L., Wolf, J., Wreford, A., 2009. Are there social limits to adaptation to climate change? Climatic Change 93 (3–4), 335–354. Alexander, D., 1997. The study of disasters, 1977–1997: some reflections on a changing field of knowledge. Disasters 21 (4), 284–304. Alvarez-Castillo, F., 2001. Gender and Health: An Overview. Health Science Association. Manila, Philippines. Arora-Jonsson, S., 2011. Virtue and vulnerability: discourses on women, gender and climate change. Global Environmental Change 21 (2), 744–751. Arvai, J., Bridge, G., Dolsak, N., Franzese, R., Koontz, T., Luginbuhl, A., Thompson, A., 2006. Adaptive management of the global climate problem: bridging the gap between climate research and climate policy. Climatic Change 78 (1), 217–225. Assan, J.K., Caminade, C., Obeng, F., 2009. Environmental variability and vulnerable livelihoods: minimising risks and optimising opportunities for poverty alleviation. Journal of International Development 21 (3), 403–418. Bacon, P.R., 1996. Wetlands and biodiversity. In: Hails, A.J. (Ed.), Wetlands, Biodiversity and the Ramsar Convention: The role of the Convention on Wetlands in the Conservation and Wise Use of Biodiversity. Ramsar Convention Bureau, Gland, p. 17. Baptiste, A.K., Nordenstam, B.J., 2009. Impact of oil and gas drilling in Trinidad: factors influencing environmental attitudes and behaviors within three rural wetland communities. Environmental Conservation 36 (1), 14–21. Blaikie, P.C., Cannon, T.T., Davis, I., Wisner, B., 1994: At Risk: Natural Hazards, People’s Vulnerability, and Disasters. Disaster and Society – From Hazard Assessment to Risk Reduction, Berlin, pp. 75–82. Brody, A., Demetriades, J. Esplen, E., 2008. Gender and Climate Change: Mapping the Linkages. A Scoping Study on Knowledge and Gaps. A Scoping Study on Knowledge and Gaps. BRIDGE, Institute of Development Studies (IDS). Brown, O., Crawford, A., 2009. Climate Change and Security in Africa. IISD, Manitoba, p. 8. Carbonell, M., Alleng, G., Browne, D., Massey, B., 2007. Nariva Swamp Restoration Initiative. Ducks Unlimited Inc., Trinidad and Tobago, Memphis, TN, USA. Chambers, R., 1989. Vulnerability, coping and policy. IDS Bulletin 20, 1–7. Chambers, R., Conway, G., 1992. Sustainable Rural Livelihoods: Practical Concepts for the 21st Century. Cutter, S.L., Mitchell, J.T., Scott, M.S., 2000. Revealing the vulnerability of people and places: a case study of Georgetown County, South Carolina. Annals of the Association of American Geographers 90 (4), 713–737. Cutter, S.L., Boruff, B.J., Shirley, W.L., 2003. Social vulnerability to environmental hazards. Social Science Quarterly 84 (1), 242–261. Dolan, A., Walker, I., 2004. Understanding vulnerability of coastal communities to climate change related risks. Journal of Coastal Research, Special Issue 39, Proceedings of the 8th International Coastal Symposium, Brazil.

K.U. Shah et al. / Geoforum 47 (2013) 125–137 Dow, K., 1992. Exploring differences in our common future(s): the meaning of vulnerability to global environmental change. GeoForum 23 (3), 417–436. Dugan, P., 1990. Wetland Conservation: A Review of Current Issues and Required Action. World Conservation Union. Dulal, H., Brodnig, G., Shah, K., 2010a. Capital assets and institutional constraints to implementation of greenhouse gas mitigation options in agriculture. Mitigation and Adaptation Strategies for Global Change 16 (1), 1–23. Dulal, H.B., Brodnig, G., Thakur, H.K., Green-Onoriose, C., 2010b. Do the poor have what they need to adapt to climate change? A case study of Nepal. Local Environment 15 (7), 621–635. Eakin, H., Bojorquez-Tapia, L., 2008. Insights into the composition of household vulnerability from multicriteria decision analysis. Global Environmental Change 18 (1), 112–127. Enarson, E., Morrow, B.H., 1998. Why Gender? Why Women? An Introduction to Women and Disaster. The Gendered Terrain of Disaster: Through Women’s Eyes, pp. 1–8. Enarson, E., Scanlon, J., 1999. Gender patterns in flood evacuation: a case study in Canada’s Red River Valley. Applied Behavioral Science Review 7 (2), 103–124. Eriksen, S., Kelly, P., 2007. Developing credible vulnerability indicators for climate adaptation policy assessment. Mitigation and Adaptation for Global Change 12 (4), 495–524. Fischer, G., Shah, M., Tubiello, F.N., van Velhuizen, H., 2005. Socio-economic and climate change impacts on agriculture: an integrated assessment, 1990–2080. Philosophical Transactions: Biological Sciences 360 (1463), 2067–2083. Frank, E., Eakin, H., Lopez-Carr, D., 2011. Social identity, perception and motivation in adaptation to climate risk in the coffee sector of Chiapas, Mexico. Global Environmental Change Part A: Human & Policy Dimensions 21 (1), 66–76. Gbetibouo, G.A., Ringler, C., Hassan, R., 2010. Vulnerability of the South African farming sector to climate change and variability: an indicator approach. Natural Resources Forum 34 (3), 175–187. Gibbes, C., Southworth, J., Keys, E., 2009. Wetland conservation: change and fragmentation in Trinidad’s protected areas. GeoForum 40, 91–104. Hahn, M.B., Riederer, A.M., Foster, S.O., 2009. The livelihood vulnerability index: a pragmatic approach to assessing risks from climate variability and change—a case study in Mozambique. Global Environmental Change 19 (1), 74–88. Heltberg, R., Siegel, P., 2009. Addressing human vulnerability to climate change: towards a no-regrets approach. Global Environmental Change 19 (1), 89–99. Hinkel, J., 2011. Indicators of vulnerability and adaptive capacity: towards a clarification of the science–policy interface. Global Environmental Change 21 (1), 198–208. Hooke, W., Pielke, R., 2000. Prediction: Science, Decision Making, and the Future of Nature. Hurley, G., 2010. Achieving Debt Sustainability and the MDGs in Small Island Developing States. UNDP Discussion Paper, Poverty Practice, Bureau for Development Policy. New York (October). IPCC, 2007. Fourth Assessment Report. Climate Change 2007: Synthesis Report. . Jankowska, M., Lopez-Carr, D., Funk, C., Husak, G., Chafe, Z., 2012. Climate change and human health: spatial modeling of water availability, malnutrition and livelihoods in Mali, Africa. Applied Geography 33, 4–15. Kacal, S.A., 1999. Social Assessment and Community Action Plan of Nariva Managed Resource Area. Ministry of Agriculture, Land and Marine Resources, Port of Spain, Trinidad and Tobago. Kelly, P.M., Adger, W.N., 2000. Theory and practice in assessing vulnerability to climate change and facilitating adaptation. Climatic Change 47 (4), 325–352. Kelman, I., West, J.J., 2009. Climate change and small island developing states: a critical review. Ecological and Environmental Anthropology 5 (1). Kumar-Range, S., 2001. Environmental Management and Disaster Risk Reduction: A Gender Perspective. United Nations Division for the Advancement of Women. McCarthy, J., Canziani, O., Leary, N., Dokken, D., White, K., 2001. Climate Change 2001: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II

137

to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge, UK. McNamara, K., Gibson, C., 2009. ‘We do not want to leave our land’: Pacific ambassadors at the United Nations resist the category of ‘climate refugees’. GeoForum 40 (3), 475–483. McWilliams, John P., Cote, Isabelle M., Gill, Jennifer A., Sutherland, William J., Watkinson, Andrew R., 2005. Accelerating impacts of temperature-induced coral bleaching in the Caribbean. Ecology 86 (8), 2055–2060. Morrow, B.H., 1999. Identifying and mapping community vulnerability. Disasters 23 (1), 1–18. Morrow, B.H., Enarson, E., 1996. Hurricane andrew through women’s eyes: issues and recommendations. International Journal Mass and Disasters 14, 5–22. Mustafa, D., 1998. Structural causes of vulnerability to flood hazards in Pakistan. Economic Geography 74 (3), 289–306. Nathai-Gyan, N., 1996. Conservation status of the Nariva Swamp. In: Nariva Swamp Seminar Report. Wildlife Section, Ministry of Agriculture and Forestry, St. Augustine, Trinidad. Neumayer, E., Plumper, T., 2007. The gendered nature of natural disasters: the impact of catastrophic events on the gender gap in life expectancy, 1981–2002. Annals of the Association of American Geographers 97 (3), 551–566. O’Brien, K., Eriksen, S., Schjolden, A., Nygaard, L.P., 2007. Why different interpretations of vulnerability matter in climate change discourses. Climate Policy 7 (1), 73–88. Osbahr, H., Twyman, C., Adger, W.N., Thomas, D.S.G., 2008. Effective livelihood adaptation to climate change disturbance: scale dimensions of practice in Mozambique. GeoForum 39 (6), 1951–1964. Pan American Development Foundation, 2006. Felicity Community: Community Emergency Plan. Port-Of-Spain, Trinidad. Preston, B.L., Yuen, E.J., Westaway, R.M., 2011. Putting vulnerability to climate change on the map: a review of approaches, benefits, and risks. Sustainability Science 6, 177–202. Prowse, M., Scott, L., 2008. Assets and adaptation: an emerging debate. IDS Bulletin 39 (4), 42–52. Pulwarty, R.S., Nurse, L.A., Trotz, U.O., 2010. Caribbean Islands in a changing climate. Environment 52 (6), 16–27. Rygel, L., O’Sullivan, D., Yarnal, B., 2006. A method for constructing a social vulnerability index: an application to hurricane storm surges in a developed country. Mitigation and Adaptation Strategies for Global Change. 11 (3) 741– 764. Schneider S.H., 2007. In: Parry, M.L., Canziani, O.F., Palutikof, J.P., van der Linden, P.J., Hanson, C.E. (Eds.), Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge, UK, pp. 779–810. Shah, K.U., 2011a. Organizational drivers of corporate environmental responsibility in the Caribbean tourism industry. Policy Sciences 44 (4), 321–344. Shah, K.U., 2011b. Firm level characteristics as determinants of external stakeholder perceptions of corporate environmentalism in an emerging economy. Corporate Social Responsibility and Environmental Management 18, 80–90. Shah, K., Rivera, J., 2007. Export processing zones and corporate environmental performance – the case of the oil, gas and chemical sectors of Trinidad and Tobago. Policy Sciences 40, 265–285. Slater, R., Peskett, L., Ludi, E., Brown, D., 2007. Climate change, agricultural policy and poverty reduction – how much do we know? Natural Resource Perspectives 109, 1–6. Sookhoo, R., 1987. The Impact of Human Activity on the Caroni River, Trinidad, WI Chemical and Microbial Studies. Watts, M., Bohle, H.G., 1993. The space of vulnerability: the causal structure of hunger and famine. Progress in Human Geography 17 (1), 43–67.