Envisioning Adaptive Strategies to Change : Participatory Scenarios for Agropastoral Semiarid Systems in Nicaragua

Historically, the semiarid social–ecological systems of the dry Central American corridor have proven resilient to pressures. However, in the last century, these systems have experienced huge environmental and socioeconomic changes that have increased the vulnerability of local livelihoods to shocks. New approaches are needed to capture complex, uncertain, cross-scale and nonlinear relationships among drivers of change and vulnerability. Therefore, to tackle this challenge, we have applied a participatory and interdisciplinary methodological framework of vulnerability assessment to a case study in northern Nicaragua. We triangulated a range of information and data from participatory and scientific research to explore historical and current drivers of changes that affect the system’s components and indicators of vulnerability, represented in a 3-dimensional space in terms of ecological resilience, the socioeconomic ability of individuals to adapt to change, and an institutional capacity to buffer and respond to crisis. A projection of climatic changes combined with a participatory scenario analysis helped, then, to heuristically analyze tendencies of vulnerability in the future and to explore what policy options might enhance the system’s adaptive capacity to face new pressures. Our work primarily contributes to an empirical understanding of key factors that influence vulnerability and learning about local strategies to adapt to change in semiarid agropastoral systems in Central America. We also make a methodological contribution by testing the use of a multidimensional vulnerability framework as a way of stimulating discussion among researchers, local stakeholders, and policy makers.


INTRODUCTION
Dryland regions provide livelihoods for almost half of the world's poorest people (Millenium Ecosystem Assessment 2005) and are vulnerable to global changes including climate change (Eriksen and O'Brien 2007).For example, 75% of drylands in South America, Central America, and the Caribbean are seriously degraded and threatened by desertification (Scherr 1999, United Nation Convention to Combat Desertification 2004).In particular, in Central America, the "dry corridor" that stretches between Mexico and Panama (World Food Program 2002) remains largely unstudied, despite holding 25% of the region's population (Reynolds et al. 2005).In addition to the stress of climate change, the area's livelihood systems are also threatened by war, cultural and demographic changes, and political and economic instability (Martí i Puig 2004).Specifically, in the centralnorthern semiarid region of Nicaragua, 75% of the farmers live on less than U.S. $2 per day (Marín and Pauwels 2001) and 27% of the population is undernourished (Food and Agriculture Organization 2004).However, there are no detailed empirical studies about how livelihoods in this region are vulnerable to a changing climate, or how multiple threats interact.
In general, there is an extensive body of literature that can be drawn upon to assess the vulnerability of livelihoods to both climatic change and other http://www.ecologyandsociety.org/vol16/iss1/art20/threats.Theoretical work draws on concepts like resilience and adaptive capacity (see Gallopin 2006, Fraser et al. 2010).More empirical studies tend to use one of two approaches.Either they use top-down quantitative biophysical modeling, and are criticized for failing to integrate methods and missing key local factors that determine vulnerability, or they use bottom-up qualitative case studies that provide a vivid contextual understanding of people's adaptation strategies in the face of multiple stresses (Fussel and Klein 2006) and their future aspirations (van Aalst et al. 2008) but lack a common and integrated analytical assessment framework to enable a structured comparison between context-specific cases and to extract broader lessons.In light of the gap between topdown and bottom-up vulnerability assessments, as well as the need to conduct detailed work on drylands livelihood systems in Central America, our work has two objectives: 1. To evaluate how the multidimensional vulnerability of livelihoods to pressures in a semiarid agropastoral system in Nicaragua has been affected over time by multiple drivers.

2.
To combine knowledge systems and methods, such as participatory methods with climate models, to develop scenarios for the future and link these scenarios with locally relevant adaptive strategies.

STUDY AREA Biophysiographic Components
The case-study area is located in the department of Estelí on a semiarid plateau in northern Nicaragua (13º09'N-86º14'W).It is located in the Miraflor-Moropotente Terrestrial Protected Landscape (Fig. 1), which consists of three ecological zones: fluvial valleys, mountainous cloud forests, and the intermediate semiarid plateau that was dominated by tropical savannah, oak woodlands, and dry forests in the 18 th century.

Climatic Features
The climate of this semiarid region is influenced by the North American monsoon system.Annual mean temperature is 23.5 ºC and annual total rainfall is 773 ± 195 mm.Rainfall occurs mainly between May and October, followed by a marked dry season.An intermediate period of scarce precipitation and high temperature ("canicula") between July and August determines the first ("primera") and the second ("postrera") crop-growing season.Rainfall is concentrated into convergence zones and is also influenced by topography, thus it shows an extreme interannual variability of up to 750 mm.The region is also exposed to droughts and floods related to the El Niño-Southern Oscillation (ENSO) phenomenon (see Appendix 1 for details).

Governance System and Socioeconomic Attributes
In terms of the socioeconomic and political landscape, since 2003 when Miraflor-Moropotente was declared a protected area, a new comanagement agreement has been in effect between local community associations, a group of medium-sized cattle ranchers, and the Nicaraguan Ministry of Natural Resources and Environment.A minority of very large commercial cattle ranchers occupy 73% of the plateau area in the region.One-third of the population remains landless, and small-scale and medium-sized mixed farming systems occupy 25% of the land and provide food, jobs, and income for half of the rural population.Livestock is the most important component of the local economy, and is used for food, as capital, as a resource for employment, and for social status.The local diet consists of maize and beans that are produced in both the primera and postrera crop-growing seasons, and these are supplemented with meat, local dairy products, vegetables and fruits from home gardens, and wild fruits from the nearby dry forests.During the dry season, the local availability of food declines and nearly half of the local population suffers from hunger and malnutrition.

FRAMEWORK AND METHODS FOR ASSESSING VULNERABILITY
Three methodological steps were used to achieve the two objectives listed above (see Fig. 2 and Appendix 2).http://www.ecologyandsociety.org/vol16/iss1/art20/Source: satellite images ASTER and LANDSAT-7; V. Garcia-Millan and G. More, unpublished data.
Step 1: Developing a Baseline Understanding of Vulnerability The purpose of step one was to establish a series of preliminary hypotheses about how livelihoods were changing, to identify mechanisms currently used to cope with climate and socioeconomic changes, to explore the values of ecosystem goods and services, and to discuss indicators of vulnerability.To accomplish this, we conducted a series of key informant interviews and a focus group with village elders.This information was triangulated with aerial photographs (1954, 1971, 1988, and 1996) and satellite images (2008) as well as through a literature review and review of archive material.Using the categories of relevant stakeholders identified by Ravera et al. (2009), we also conducted a perception analysis of local environmental and development issues through in-depth and semistructured interviews, selected through snowball sampling, and a first series of deliberative focus groups.
Step 2: Assessing Historical and Current Vulnerability Interviews and survey data collected during step 1 were transcribed and analyzed using discourse analysis (Gee 1999).By following a grounded theory approach, researchers specifically looked for emergent variables and relationships that described the structure and functions of the agropastoral system (Sendzimir et al. 2007).The findings were organized into a conceptual model, accompanied by a graphical representation that helped to incorporate a diversity of knowledge and perceptions to reflect the multiscale causalities and feedbacks expressed in the transcripts.In parallel, a qualitative narrative http://www.ecologyandsociety.org/vol16/iss1/art20/Fig. 2. Methodological framework for assessing vulnerability to change.
synthesized how local agropastoral systems have become over time more or less vulnerable to perturbations.This conceptual model was then refined through a further set of in-depth interviews with key informants, extended meetings with local stakeholders, and a new series of meetings with experts.To assess changes in vulnerability, we followed Fraser (2007) who argues that multidimensional livelihood vulnerability is a function of three overlapping elements: (1) the ecological resilience of agroecosystems, referring to the extent to which the agroecosystem is able to maintain or recover key functions (Holling 1986, Holling et al. 2001) that are essential for production (Walker and Abel 2001), (2) socioeconomic ability, referring to the extent to which the socioeconomic system helps to provide the resources or assets individuals and households require to reproduce a productive system and sustain local livelihoods in the long term and, thus, to adapt to shocks (Sens 1991, Scoones 1998) without becoming destitute (Corbett 1988), and (3) the capacity of local institutions to provide a social buffer or safety net, e.g., food-security programs or risk-alert systems, to protect livelihoods and help mitigate unexpected crises (see also Fraser et al. 2011).Field research (Ravera et al. 2009, Tarrasón et al. 2010) and literature provided the information to qualitatively http://www.ecologyandsociety.org/vol16/iss1/art20/infer current vulnerability in terms of these indicators.
Step 3: Developing Future Vulnerability Scenarios The purpose of this step was to assess how climatic and socioeconomic drivers might affect livelihood vulnerability in the future.This phase also allowed stakeholders to deliberate policy options.Three activities were undertaken.First, an exploratory exercise was developed to capture conflicting visions about the future, by collecting images (collages and metaphors) and storylines during indepth interviews and deliberative focus groups.Second, a future climate scenario was developed based on monthly observed temperatures and precipitation levels for 1961-2007 from nearby meteorological stations with the same geographical and climatic characteristics of the study area.To develop a temperature and rainfall scenario up to 2050, we projected seasonal changes and frequency of extreme events based on literature references (Rousteenoja et al. 2003, Christensen et al. 2007).
Local impacts of these changes were then identified through the literature (see Appendix 1 for details).Third, during a final series of deliberative focus groups, the conceptual model and climate scenario were used as the basis for a backcasting exercise.Each scenario (and its components) was overlapped to the conceptual model to infer future trends of changes in the performance of vulnerability indicators.Moreover, using a 3D plan of the area and a list of questions about environmental, social, and economic policies related to each envisioned future, researchers facilitated a discussion of shortand medium-term management and policy options.The alternative options were classified by researchers as either being: (1) unplanned responses undertaken spontaneously and regularly by local people as a reaction to crisis, portrayed as coping mechanisms (as defined by Osbahr et al. 2008), or (2) renewed, planned, and mainly longer timescale actions, portrayed as adaptive strategies to be implemented (as defined by Nayak 2004).Next, each scenario and policy option was ranked according to the type of uncertainty being explored, such as incomplete knowledge, unclear tendencies and the unpredictability of change, a plurality of conflicting perceptions, or ignorance.As a result, each scenario was evaluated along the three dimensions of vulnerability, forming "vulnerability trajectories" in relation to the present, or baseline.

Trends and Drivers in Historical and Current Vulnerability
To analyze historical forces and patterns of changes in the study region, Nicaragua's history during the past century up to the 2006 elections is divided into four periods, summarized in Table 1.The following narrative links historical drivers with current trends of change and is organized based on Fraser's (2007) three dimensions of vulnerability of livelihoods.Key factors and interrelationships that influence the indicators of vulnerability are summarized as a conceptual model in Fig. 3.

Ecological resilience of the agropastoral system
Three key changes stand out as affecting the resilience of the agropastoral system (see the green box in Fig. 3).
First, "native pastures lands" have experienced a slow process of ecological transition from mature grasses (Paspalum notatum Fluggë) to either a degraded state or a bush and woody (e.g., Acacia pennatula) encroached ecosystem.This change has been observed by local land users and confirmed with aerial photographs and fieldwork (Tarrasón et al. unpublished data;Fig. 4) and has been driven by both land intensification and land abandonment.
Land intensification is problematic because although native grasses tolerate trampling and frequent grazing (Primavesi 2004), recurrent droughts combined with overgrazing and the introduction of exotic grass species have spread unpalatable species and inhibited the capacity of P. notatum to recover from climatic shocks and stresses.The historical causes of land intensification are summarized in Table 1.In contrast, some land has also been abandoned because of socioeconomic and political conditions over the last three decades.
In particular, neoliberal macroeconomic policy changes and socioeconomic uncertainty (Table 1) have meant that farming is currently no longer economically viable in some areas.Oral histories and the literature (Gibson 1996, Kaimowitz 1996) suggest that land concentration combined with credit shortage have increased land abandonment.
In fact, according to the interviews, low-income and highly indebted households have first intensified and then abandoned and sold lands to a few "latifundio-minifundio" and new commercial landowners.(The latifundio-minifundio land- tenure system referred, in Latin American countries, to large estates of lands administered by few families with a patronage system [Latin: latus, "spacious" + fundus, "farm, estate"] scattered with tiny land plots [Latin: minus, "minor" + fundus, "farm, estate"]).At the same time, economic stagnation and unclear land tenure have both reduced investments in technology, land management, and labor.As a result, rural workers have passively exercised an opposition to exploitation and inequality by working less efficiently, and have indirectly contributed to pasturelands encroachment.
Other factors have also affected the state of native pastures, such as changes in environmental awareness that have driven the implementation of new environmental priorities and policies.For example, all interviews revealed that since the area was declared "protected," local alliances and rules have been re-established to defend small farmers from being shut out of land access and management by large-scale landowners (Ravnborg 2008).
Although the implementation of new management practices such as fire, a logging ban, and silvopastoral practices has provided in situ benefits mainly during the dry season, this has resulted in bushes and trees spreading drought-resistant species such as A. pennatula, and the consequent inhibition of native grasses (Peguero and Espelta, unpublished data).These changes have affected the resilience of the ecosystem by reducing soil protection, threatening functional diversity such as species control of native grasses (sensu Folke et al. 1996), decreasing productivity of pastures over a range of climatic perturbations (Walker et al. 1999) and favoring landscape fragmentation.Consequently, the degradation of pastureland has decreased dairy and meat production, increased the debt-to-income ratio for small farmers, reinforced household food insecurity, and reduced profits for commercial landowners.
Second, "agricultural lands" have experienced a decrease in yields and profitability, just as the impact of climatic events (e.g., harvest losses) and economic crises, e.g., price volatility, have increased.For a long time, subsistence farmers had maintained a traditional grain system that minimized external inputs to reduce costs and preserved long-term productivity.Nevertheless, traditional grain crops have progressively been replaced with a more intense mixed-farming system, e.g., cash and fodder crops (Table 1).Despite providing a short-term boost of high productivity, this intensification has accelerated a long-term process of erosion in agrobiodiversity and soil quality.In fact, intensified agriculture reduces Ecology and Society 16(1): 20 http://www.ecologyandsociety.org/vol16/iss1/art20/Fig. 4. Bush encroachment trends (1954, 1971, 1988, and 1996) in the study area (the area within the white line highlights the changes in shrub cover).spatial diversity with effects on soil-nutrient cycling (Ravera and Tarrasón, unpublished data), altered soil food-web composition through herbicide and pesticide use (Wardle et al. 1998), and enhances soil erosion under conditions of extremely strong rainfall and recurrent drought (Stocking and Murnaghan 2001).This affects the stability of production and, consequently, the ability of local household economies to adapt to change.Moreover, this agricultural intensification has increased farmers' dependency on inputs while reducing their capacity to handle debt in times of economic crisis.http://www.ecologyandsociety.org/vol16/iss1/art20/Third, "dry forest lands" are currently degraded and fragmented.This is because of a range of historical forces, including population pressure and changes in agrarian systems.Interview results suggest that, despite new protective rules, the inefficacy of the local governance system has increased the disturbance of dry forests over time.In fact, grazing pressure, firewood extraction, accidental fires, and tree harvesting have increased, and this has reduced natural regeneration rates and tree growth (Tarrasón et al. 2010).Local perceptions confirm observations in the literature (Intergovernmental Panel on Climate Change 2007) that dry forests are currently more exposed to drier conditions and extreme events (e.g., intense drought) at individual (e.g., phenology), communitarian (e.g., distribution, interactions), and ecosystem levels (e.g., water storage) than in the past.However, a renewed sense of belonging has led to an increase in environmentally friendly practices, and this may help decrease future forest degradation (Tarrasón et al. 2010) in the area.

Socioeconomic ability of individuals to adapt to change
In the study region, the extent to which individuals have access to assets that allow them to adapt to change was sharply divided between commercial economies, as opposed to household economies, that is, small farmers, agropastoral, and landless peoples.Key interactions in this system are highlighted in the blue box of Fig. 3.The two types of economies correspond to different objectives and strategies of resource allocation and land management, i.e., maximization of profit and investments as opposed to minimization of risk and food-security maintenance.Individual and household decisions impact the intensity of disturbance that accelerates or slows down the ecological processes of land degradation.Over the past five decades, government policies have favored intensification, and this has reduced the diversity of agricultural commodities being produced in the region (Table 1).Concurrently, local socioeconomic structures have favored land concentration and this has displaced the poorest people onto less productive lands.Paradoxically, although small farmers had obtained access to land through the agrarian reform of the 1980s, they have quickly lost the financial assets to keep these lands productive, and many farmers have fallen under growing debt burdens (Baumeister 2001, Dufumier 2004).Although cattle stocks have increased both locally and nationally for the last two decades (Food and Agriculture Organization 2009a), the cattle raisers' investment capacity has stagnated, and weak organizations have undermined the ability of both commercial and household economies to respond to market dynamics.This has prevented technological innovations from being adopted and has resulted in unequal access to wealth, weakening the ability of individuals to cope with crises (Ravnborg 2003).Interviews show that the most vulnerable households are those composed of elderly parents, single-parent women, or young, landless people who work in the commercial economy.These groups have limited access to land and natural resources such as water and firewood, and lack human, physical, and financial assets, such as wages, animals, technology, and credit.
The perception analysis demonstrates that, historically, local people have shown evidence of mechanisms to cope with risk as a regular component of ongoing management (Table 2).Social capital has also played a key role in supporting the ability of individuals to cope with crises, especially food shortages.However, both individual and collective coping mechanisms, drawing on social networks, have changed and have been eroded in recent years.First, several drivers have affected the extent to which people trust their social contacts.Second, the crisis of household economies has affected initiatives to sustain local management and livelihoods.This has increased the permanent out-migration of local young people to urban areas and foreign countries as a short-term coping response.Demographic changes have meant that the productive population is dwindling, and this itself is driven by poor job opportunities, local living conditions, household indebtedness, and the attraction of an urban lifestyle.Consequently, a renewal in leadership and the long-term ability of individuals to buffer themselves from crisis have been affected.

Institutional capacity to buffer and respond to crisis
Over time, a lack of coordination across administrative levels has resulted in there being little in place in the way of disaster planning or earlywarning systems in the region.This problem has been widely studied in the literature over different periods (Pyner and Strachan 1976, Biondi-Morra 1993, Sahley et al. 2005).The drivers of change for each historical period are summarized in Table 1.As a result, extreme-weather events, such as Productive mechanisms: • transhumance migration systems of animals and people between semiarid and lowlands or humid areas • renting of farmland in the humid area to extend the growing season, to minimize risk of losses (the "apante" is the growing season during the dry period) Land-use changes and resources management mechanisms: • agropastoral diversification and management to reduced impact of market volatility • preservation of traditional seed varieties to maintain a range of resistance characteristics • retaining easily disposable assets, such as small livestock, to be sold during times of stress Labor allocation and intensification mechanisms: • temporary migration, both seasonal to obtain cash, and to urban or foreign areas for remittance • allocation of work within the extended family Collective mechanisms: • a sharing system ("a media" means "to share") in which landowners either rent grazing land in the dry season in exchange for half the milk production or dung, or purchase inputs for sharecroppers who provide labor • in-kind transfers of goods and services between farmers • family and social networks as a source of food or cash in crisis periods (e.g., seeds gathering by landless people during the dry season when landowners' cattle has migrated) • informal markets of dairy products through social networks in the town Destitution mechanisms: • permanent out-migration • borrowing food and money from merchant and financial organizations at high rates of interest • transfer of capital stock to financial capital (i.e., selling animals or land) Hurricane Mitch (1998), extreme droughts (such as in 2001 and 2005), and socioeconomic shocks, such as the global economic crisis (2007-2008;Food and Agriculture Organization 2009b), have all resulted in famine in this region.In particular, decentralization programs have failed to build human capacity and enhance investments, and this has disempowered local governments (Martí i Puig 2004).Moreover, the World Bank's welfare programs, implemented by a neoliberal government to act as buffer for food security, have been ineffective (Sahley et al. 2005).Similarly, cooperative unions, small farmers' syndicates, and historically powerful ranchers unions, have all shown internal divisions at national and local levels since the collapse of the cooperative system, and have been unable to respond to crises.The state's incapacity to coordinate and handle crisis relief has been partly offset by NGOs that have played a key role in reinforcing social safety nets and providing assistance in crisis-stricken areas.On the other hand, the activities of NGOs have reinforced a dependency amongst locals on these programs (Sahley et al. 2005).Recently, the new socialist central government has implemented a commandand-control regime of natural resource management and food production, distribution, and storage.This has been criticized as it overlaps with local community authorities and risks undermining social relations (Muñoz 2007).At the local level, new schemes and rules around the use of and access to natural resources implemented in protected areas (e.g., fire bans, restrictions on agrochemical use, restrictions on logging activities), have triggered new resource-based conflicts (Nygren 1999, Ravnborg 2008).According to interviewees, the comanagement framework still lacks transparency and legitimacy.Consequently, local tensions have increased, resulting in a loss of trust and the fragmentation of social ties.This has created a vicious cycle: the erosion of social networks has destroyed informal exchanges of goods and services, and this exacerbates social conflicts.The red box in Fig. 3 demonstrates the current links in this dimension of vulnerability.http://www.ecologyandsociety.org/vol16/iss1/art20/

Summary of trends in current vulnerability
Taken together, the degradation of the landscape, driven by both land abandonment and intensification; the loss of assets available to poorer households, driven by economic changes and agricultural specialization; and a reduction in the capacity of formal institutions to provide an effective social buffer or safety net, suggests that the vulnerability of livelihoods in this region has increased.From this, we may infer that future climatic and socioeconomic perturbations, such as drought events or financial crises, may have a commensurately larger impact than past ones.

Future Vulnerability Scenarios
Researchers and local stakeholders jointly developed scenarios for 2030 to reflect how livelihood vulnerability may change due to future multiple drivers in this agropastoral system.The main socioeconomic components of the four scenarios suggested by the stakeholders are summarized in Table 3, and their short storylines appear in Appendix 3.
In particular, local stakeholders were asked to debate a range of possible adaptation strategies to the added stress of increasing the annual mean temperature by 3.5°C, as compared with the 1980-99 average, by 2050, accompanied by a shift in seasonality and a high variability in the intensity of rainfall events (Fig. Conflicting perceptions emerged, and unclear and unpredictable cause-effect relationships were debated.Stakeholders agreed that innovations in green technology could enhance land productivity and would likely be ecologically friendly and economically efficient.However, other stakeholders argued that such innovations could have unexpected impacts, and would likely only be beneficial during climatically "normal" years because, as agroecosystems become more intensively managed, they would be more sensitive to drought and other climatic stresses.This would likely benefit commercial ranchers who have huge land holdings (and are thus protected from small-scale climatic problems), and they would be able to increase their assets, leading to greater inequities and conflict.This would create a further feedback with mechanized agriculture reducing the demand for labor, thus increasing migration and dismantling social mechanisms that buffer poor households against food insecurity.Thus, safety nets would be driven by private and external aid support, resulting in high levels of uncertainty about the accountability and transparency of local governance.

Summary of trends in future vulnerability and adaptation
The implications for vulnerability to multiple stresses (Fig. 5) are summarized as follows: 1.
Under scenario I, in terms of future agroecological resilience, the ability of individuals to adapt and the capacity of institutions to provide buffers will diminish.Uncertainty about the trends is low, and primarily related to either ignorance or incomplete information and knowledge with regard to the extent of possible changes.

2.
Under scenario II, it is unclear whether agroecological resilience will rise or fall.The assets available to individuals will likely increase for household-level economies, whereas commercial landowners will redistribute assets.The capacity of institutions to respond to crisis will increase, but with high levels of unpredictability due to possible contradictions about the effects of a neosocialist system.

3.
In scenario III, all three dimensions of vulnerability appear to be set to improve.Uncertainty about trends is low, and there is a confluence of visions with regard to the effects of drivers.

4.
The final scenario (IV) has unclear and conflicting implications regarding all three dimensions.
As a final point, the consequences of different development pathways visualized in a systematic way helped local stakeholders to design a multiscale bundle of strategies across these different scenarios (Table 4).

FINAL REMARKS Empirical Remarks about the Drivers of Vulnerability
This study focuses on understanding the key driving forces behind changes in semiarid agropastoral systems in Nicaragua, providing some evidence of a larger picture of future vulnerability and suggestions with regard to adaptive strategies across space and time.

Economic forces and the uncertain environment
The literature shows that both global environmental and economic changes have an effect on livelihood vulnerability (O'Brien and Leichenko 2000, Stringer 2009).Therefore, anticipating future vulnerability in semiarid environments must take into account varying degrees of ecological and socioeconomic uncertainty.Our analysis shows that historical inequities in land and wealth distribution were driven by macroeconomic policies, and this affected the local management of natural resources.
In terms of how this may play out in the future, both scenarios I and IV suggest that inequities may contribute to the continued destitution of small farmers, and this may increase vulnerability (Ohlsson 2000) and environmental degradation (Rahman 2004).The study also suggests that the relationships among resilience, economic policies, and instability, are complex.Historically, incentives that stimulated both commercial farmers in the 1960s and cooperatives in the 1980s disrupted important local-level social-ecological functions (Kaimowitz 1997).Top-down agricultural policies may erode ecological resilience even where political conditions are stable (Fraser and Stringer 2009).
Conversely, in unstable situations, such as in neoliberal regimes, land abandonment, short-term investments, and opportunistic behaviors in landowners are common (Albers and Goldbach 2000).Finally, the research presented here shows that when local economies depend on few resources and activities, they are vulnerable to the "boom and bust" nature of markets (Adger 2000), and this accentuates the likely impact of future climatic variability.In contrast, macroeconomic policies that favor equal access to land and diverse markets that include tourism and fair-trade markets can cushion households during periods of food insecurity.These more optimistic futures are the subject of scenarios II and III.

Leadership, social capital, and the governance system
This study also confirms the literature (e.g., Folke et al. 2002)   Table 4. Individual and institutional adaptive strategies and policies proposed by stakeholders, their scale, and dimension of intervention.

Adaptive strategy Scale Dimension
Harmonize institutional planning responses, integrate participatory decision making processes: • decentralization, without deregulation, that creates, strengthens, and delegates power and economic responsibility to local organizations and institutions • private-public partnerships national-regionallocal institutional Revise international trade policies to improve market access: • take advantage of existing mechanisms for "local products" and special "safeguard mechanisms" to protect national agricultural sectors • establish appropriate food stocks to prevent price volatility • secure access to information and microcredit nationalinternational political Mechanisms and funding to support rural investments: • establish appropriate policies to reduce impacts of food-price inflation • invest in agriculture in low-potential areas as a social investment • diversify rural on-farm and off-farm economies • financial compensation for ecosystem-services protection • infrastructure investments nationalinternational political Strengthen law enforcement for land ownership and rights to natural resources access: • secure land rights • ensure land access for disadvantaged groups • restrict land sales to foreign investors national political Reinforce organizations and networks (governance and adaptive comanagement): • capacity building for communities to achieve self-sustaining projects • strengthen alliances and coordination between comanagers (FORO Miraflor), communities, landowners, ranchers, trade unions, councils, and academic institutions • enhance market competitiveness, e.g., construct warehouses for crop and dairy products local-regional social-institutional Farmer-to-farmer knowledge exchange and extension: • share good farming practices through the establishment of model farms, and strengthen capacity to monitor and assess • exchange native crop varieties through local seed fairs • join national networks, e.g., Farmer-to-Farmer Program; initiatives such as "Seeds for Identity" • farm-planning design with ten-year timespan local-regionalnational social-institutional Innovative agricultural practices: • switch from monoculture to diversified agriculture: use traditional maize-bean intercropping system, cultivar rotation with green manure or farm-cattle manure

Cultural drivers
Local institutional arrangements and land-use behavior are deeply influenced by cultural values.Historical evidence and future scenarios illustrate how international lifestyles and values have changed (e.g., consumer demand for green products) and have influenced land use and management.For example, technological innovations in conventional agroindustry (scenario IV) are perceived to boost rural development and improve living standards for some, while aggravating social inequities.In contrast, the technological changes discussed in scenarios II and III are more socially inclusive.Providing funding that supports ecosystem-services management (Goldman et al. 2008) is an illustration of this, and has been used to combat poverty and enhance nature conservation (Hecht 2004).However, these programs have failed to recognize the roles of the agropastoral system as a provider of ecosystem services (Pagiola et al. 2007) and to support subsistence farmers as conservationists.Therefore, there is a need to develop new context-specific strategies that value knowledge exchanges among local stakeholders and researchers.

Methodological Remarks about Studying Vulnerability
Reflecting on the process undertaken to conduct this research, we have highlighted some theoretical and methodological challenges concerning vulnerability assessments.First, we have demonstrated that conceptual modeling and participatory scenario development can be powerful tools for bringing knowledge systems together, empowering local stakeholders to distinguish opportunities and threats, and enabling negotiation.Second, overlapping the baseline conceptual model with future scenarios and climatic stress allowed us to creatively imagine a proactive and anticipatory, rather than reactive, adaptation window.Framing interrelated drivers and factors into the three dimensions of vulnerability is a manageable format for dealing analytically with multidimensional assessments of vulnerability to change, and helps to identify critical components for making the system more vulnerable or resilient.Finally, we have found that integrated methodological frameworks can deepen our understanding of semiarid livelihood systems as a whole, and our comprehension of hypothetical factors that may reinforce or weaken http://www.ecologyandsociety.org/vol16/iss1/art20/their vulnerability (Knutsson and Ostwald 2006).Additionally, the proposed framework has enabled a comparison between cases to extract broader lessons on multiple facets of change in complex regional social-ecological systems, such as the semiarid agropastoral systems in the Central American dry corridor.In conclusion, the proposed framework of assessment has demonstrated a tool that is helpful for planning processes, for exploring possible future pathways and negotiating the key components of scenarios that can help to prioritize adaptation decisions.However, the vulnerability assessment needs further refinement.Further research will focus on defining irreversibility in maintaining resilience when indicator thresholds are passed, and on building interfaces between social research and mathematical modeling, both theoretically and practically.

Future Projections and Potential Impacts
The future climate scenario for this study was developed from the two Special Reports on Emissions Scenarios A1B-projections for the 21st century with 1980-99 as a baseline.First, annual mean temperature and rainfall for the 2050s builds on Ruosteenoja et al. (2003).The annual changes were calculated as the mean changes for the dry and wet season, respectively.Second, seasonal changes for 2080-99 build on scenarios for Central America developed by Christensen et al. (2007).The projected differences in minimum, maximum, median, 25%, and 75% quartiles between the baseline and 2080-99 periods were used to modify the distribution of the observed baseline.The two original sources included seven and 21 General Circulation Models, respectively.
The resulting scenarios indicate a change in annual mean temperature by 0.8°C for 2050.Minimum temperatures may increase by 1.4-2.0°Cand maximum temperatures by as much as 4.6-5.5°C.In terms of seasonal changes, the scenario indicates that winters may have more strong increases in both minimum and maximum temperatures, leading to frequent heat waves and dry spells.The change in annual mean precipitation may be more irregular and may range from -13.5% to +4% by 2050.This translates to a range of about -110 to +30 mm as minimum rainfall decrease by 45%-57%, whereas maximum rainfall increases by up to 24%.These scenarios further indicate that the frequency of dry seasons may increase by 15%-25%.In contrast, the seasonal scenarios, project winters with higher intensity rainfall and stronger and/or more frequent tropical storms.Figure A.1.2shows the ranges of change in projections of average total annual rainfall for 2090.To illustrate the likely increase in extreme events (both floods and droughts), the maximum May rainfall between 1980-99 was 565 mm, whereas by the 2090s it may reach 650 mm.However, the median rainfall was only 80 mm and may decline further to 67 mm.The most likely immediate impact expected in the region is a more intense and recurrent drought risk in both seasons.Moreover, flood and wind-risk damages are expected in winter under the scenario studied.Modest shifts in the seasonality can lead to remarkable ecosystem changes.Seasonally, Central American dry forests are considered severely threatened by global warming (Halpin et al. 1995, Intergovernmental Panel on Climate Change 2007).Drought recurrence can also alter grassland/shrubland boundaries (Intergovernmental Panel on Climate Change 1998), species composition (Sala et al. 2000), reduce plant physiological functions due to heat stress (Battisti and Naylor 2008) and decrease water content in topsoil and soil erosion by wind (Magrin. et al. 2007).Grain-crop yield losses are expected to increase under extreme climatic events (Lobell et al. 2008).In particular, according to studies undertaken in Latin America's dry regions (Giménez 2006), higher minimum temperatures, combined with water limitation in the fall, may shorten the growing season.Other studies show that in medium-altitude semiarid regions of Central America, maize yields may decrease between 6% and 17% by 2055, and other staple crops, e.g., beans, may be badly affected as a consequence of drought and other extreme climatic events (Jones and Thornton 2003).Species characterized by high reproduction rates are generally favored by temperature increases, leading to a potential rise in the distribution and occurrence of pest infestation and pathogens (Magrin et al. 2007).Consequently, increasingly fluctuating weather patterns could have a strong negative economic impact on agriculture, by increasing labor time and production costs (Intergovernmental Panel on Climate Change 1998).Similar studies in dry regions show that the expected key impacts of change in rainfall and drought recurrence and intensity on livelihoods (particularly smallholders) are: food insecurity, through diminished crop production and increased food prices (Hertel et al. 2010), declining survival rates of livestock (Richardson et al. 2007) and increasing spread of diseases, e.g., dengue/dengue hemorrhagic fever (Rosenzweig et al. 2001, Patz et al. 2005).Human migration from drought-affected areas and tenser social relations due to scarcity of land and natural resources (Barnett and Adger 2007) may also be expected in the study area.

Fig. 1 .
Fig. 1.Map of the Miraflor-Moropotente Terrestrial Protected Landscape and land cover in 2008.The study area is located in the semiarid zone (within the black box).

Fig. 5 .
Fig. 5. (a) Heuristic illustration of trends of vulnerability indicators within the four scenarios, and (b) 3D space representation of the resulting hypothetical pathways from the actual baseline (t=0) to 2030 (adapted from Fraser 2010).

Fig. A. 1 .
Fig. A.1.1b.Trends in annual total rainfall from 1959 to 2005 (the data have been standardized to facilitate comparison).

Table 1 .
Historical drivers of change and vulnerability trends.

Table 2 .
Local mechanisms to cope with disturbance in the study area.

Table 3 .
Overview of the main differences between drivers/components of the four participatory socioeconomic scenarios for the area for 2030.