Social-ecological drivers of tree growth and diversity: interdisciplinary learning with smallholder carbon agroforestry schemes in Mexico, Uganda and Mozambique
Item statusRestricted Access
Embargo end date26/11/2020
Wells, Geoffrey James
Achieving sustainable landscapes is one of our greatest challenges, not least in the tropics in which almost half of the world’s people live, many of whom experience high levels of poverty. Yet land use policy interventions continue to be impaired by partial knowledge of how these complex systems function, and thus how interventions actually affect land management outcomes on the ground. Increasingly there are calls to recognise these landscapes as variable and complex social-ecological systems, where the drivers of land use are often sufficiently diverse, dynamic and heterogeneous that conventional technical approaches to land use modelling and policy (e.g. those focused around simple environmental and economic factors) will likely only provide partial knowledge and solutions. In particular, land use interventions face two new demands: 1) to better understand and integrate into decision making variability in social-ecological systems, and how social and institutional factors drive this; and 2) to develop more effective interdisciplinary ways of working which leverage the value of not only technical analyses, but also the experience and contextual knowledge of local and other actors. In this thesis I aim to contribute to these two areas through an interdisciplinary analysis of the social-ecological drivers of actual land management outcomes across 1299 farms in three of the world’s oldest tropical smallholder carbon agroforestry projects: the Scolel’te project in Mexico; the Trees for Global Benefits project in Uganda; and the Sofala Community Carbon Project in Mozambique. In doing so I treat our study projects as case studies of payments for ecosystem service (PES) projects, and of rural tropical land use interventions generally. I use primary data I collected during field expeditions in each country, and a range of other existing biophysical and social data. I situate my study broadly within the field of environmental geography, with various chapters drawing on theories and methods from ecology, sociology, economics, political ecology, science and technology studies, and systems science. In Chapter 2 I examine the actual level of variability of land management outcomes across 1299 project farms. I use tree inventories and ecological methods to assess two common metrics used to assess land management outcomes in forest-based PES: aboveground biomass (AGB) accumulation; and diversity of tree species. I find that, while there has been AGB accumulation at the project-level, within each project there is great variability in AGB and tree diversity between farms, beyond that predicted in project designs. Additionally, this variability increases over time. I argue that similar interventions should thus expect and plan for such persistent and increasing variability in their project designs. In Chapter 3 I use a mixed-methods analysis across 492 farms to examine the effects of a series of hypothesised social and environmental drivers on AGB and tree diversity. I include not only drivers commonly cited in theory, but also drivers suggested in interviews with local farmers and project technicians. Using biophysical and social data I employ regression techniques to assess quantitatively the effects of the social and environmental drivers, then supplement this through a qualitative thematic synthesis of views from interviews with local actors. I find that, amongst farmers participating in these projects, while tree diversity is mainly predicted by environmental variables, household wealth is a better predictor of AGB than environmental factors. Additionally, AGB and tree diversity often tend to be similar within the same village, and this appears to be an institutional rather than spatial phenomenon (e.g. institutions supporting farmer training and tree nurseries). I also find that projects with higher levels of administrative control over land management do not necessarily have lower variability in outcomes. I draw two broad conclusions for natural resource management and land use science generally: 1) social and institutional factors can be just as important as environmental factors in determining land cover; and 2) land management outcomes and their drivers cannot necessarily be closely controlled by project administrators managers, so flexible and adaptive approaches are necessary. In Chapter 4 I extend my analysis into a more in-depth examination of how a series of within-project (or ‘intra-institutional’) factors have been associated with farm-level AGB outcomes across 39 smallholdings in Mexico and Mozambique. Situated within a critical institutionalist theoretical perspective, I use fuzzy qualitative comparative analysis combined with a thematic synthesis of farmer perspectives to analyse four institutional attributes that relate to a number of key debates on design in PES projects: organisational centralisation; organisational decentralisation; participation in decision making; and responses to monetary incentives. I find the presence of effective participation by local people, and the perception of long term (monetary and nonmonetary) incentives, to be most commonly associated with high AGB outcomes in our cases. Beyond this, the degree of (de)centralisation is not important in our cases, as long as there is some form of strong (vertical or horizontal) project process. The actual degree of centralisation appears to be culturally and historically contingent, with effective institutions linking to, and evolving with, local contexts. I thus conclude that PES practitioners should not expect to have direct control over the ‘design’ of project institutions, but instead focus on supporting the evolution and adaptation of (centralised and/or decentralised) processes in response to social-ecological diversity and dynamism—a form of ‘intra-institutional bricolage’. Within this, practitioners should focus on maintaining good communication and relations with local participants, and emphasising monetary and non-monetary benefits in both the short and long terms. In Chapter 5 I take a step back to look at how different key groups of actors with shared views (i.e. epistemic communities) perceive that PES causes overall change in a social-ecological system, and to what degree of certainty (or ‘certitude’). I also explore the epistemological origins, interactions, complementarities and tensions between these different perspectives. Drawing on theories and methods from science and technology studies and social-ecological systems theory, through thematic analyses of 65 interviews, 23 project documents and existing academic literature, I assess the views of three key groups of actors already established in the PES literature: actors from theory, market administration and local practice. I argue that, in our cases, local actors perceive and engage with change in PES with the aim of transforming the system to support broader ‘domains of attraction’ and associated ‘regime shifts’, while other perspectives are focused on simpler shifts in state variables, which are perhaps less concerned about the sustainability of changes. I find that, while these approaches have found a way to co-exist, this does not occur without tension, and is largely supported by local actors working double-time (and not always successfully) to satisfy the narrower views of external stakeholders while also dealing with their own complex day-to-day reality. Given the typical nature of our cases, I argue that PES practice would benefit from greater reflexivity by rule-makers, and improved understandings of perspectives between different groups of actors. I also offer ideas on how PES could supplement existing metrics of success (which are focused on particular system states; e.g. tonnes of carbon per hectare), with other metrics that better reflect the things that drive decision making and change at the local level: namely, relative measures of progress in supporting desired domains of attraction, and adaptive capacity. Finally, in Chapter 6, I synthesis the key findings from my thesis and discuss the implications for the PES field, and for land use policy and theory more generally. I argue that my findings provide novel evidence that socio-economic and institutional factors affect AGB and tree diversity as much as environmental factors in forest land use projects. I also argue that my findings move forward the debate on PES design by showing that there is no trade-off to be had between economic design features and community participation or legitimacy—they all appear equally important. Finally, I argue that my application of systems theory to PES provides new insights on different understandings of causal change in land use systems, which may both improve interdisciplinary ways of working in such projects, and offer new ways of monitoring PES success.