Resilience in rural social-ecological systems : a spatially explicit agent-based modelling approach

Rural areas are increasingly changed by drivers on large spatial scales such as economic globalization and climate change. These international drivers may bring forth abrupt disturbances, such as high output price peaks and falls, water floods and droughts, and massive outbreaks of animal diseases which negatively affect a rural areas’ environmental and socio-economic development. To understand such effects, approaches are needed that consider rural areas from a complex system perspective, taking into account the interdependence between ecological and socio-economic dynamics. Therefore, in this thesis rural areas are considered as a social-ecological system (SES), characterized by strong links between the social and the ecological component, and by multiple interactions across spatial and temporal scales. The behaviour of such a complex system is highly unpredictable, and the effects of disturbances therefore highly uncertain. In this context, resilience has been promoted as a concept to guide and direct the management of SESs. Resilience is a system property that reflects the capacity of a system to absorb disturbance, undergo change, and retain the same essential functions, structure, identity and feedbacks. So far, resilience has often been explored as a metaphorical or theoretical construct. However, to be able to use the concept in the management of rural SESs, it needs to be made operational and measurable. The objective of this thesis is to explore how the concept of resilience can be operationalized and implemented into the management of rural social-ecological systems. This general objective can be divided into two sub goals. The first objective aims to identify criteria that can be considered as indicators for a rural SESs resilience. The second objective aims at assessing the behaviour of these indicators in an experimental setting, capturing the complex dynamics of rural SESs by means of a spatially explicit agent-based model[1]. In this way, management advice can be provided that takes the coevolving nature of rural SESs into account and supports strategies to cope with uncertainty. In Chapter 2 criteria are developed for a policy objectives evaluation framework that analyses how rural development policies contribute to the resilience of rural areas. Each criterion is described in its rural social-ecological context and specifications are proposed that make each of the criteria applicable in judging and evaluating policy measures for developing resilience in rural SESs. The framework is applied to European rural development policies, specifically focusing on the spending of compulsory modulation budget. The case study signalled the strengths and weaknesses of the framework with respect to the coherence and distinctiveness of the used criteria. In Chapter 3 a spatially explicit rural agent-based model (named SERA) is developed which aims at evaluating how rural policy interventions affect farmers, their land use and the landscape of which they are a part. Using the criteria proposed in Chapter 2, the model provides a way to evaluate how the contribution of these policies to resilience could be made effective, while imposing disturbances to the system. The model constitutes a virtual rural region, comprised by a large number of individually acting farms that operate with each other and with parts of their environment. The model initializes with empirical data on individual farms and existing agricultural spatial structures. According to their behavioural model, the individual farm agents evolve subject to their state of attributes and to changes in their environment. Chapter 4 investigates the sensitivity of the model output to changes in parameters. In this way, the parameters that are the key drivers of the model results are discovered. A mixed methodological approach is used, simulating uncertainties one-at-a-time and then together in a Monte Carlo simulation using random sampling. This mixed methodological approach provides understanding in the model’s behaviour by revealing non-linear relations between parameters and outputs, interactions between parameters and possible conditional terms. Chapter 5 discusses the implementation of the model for assessing ecological resilience under different policy scenarios in the case study area Winterswijk. The model is used to explore how farmers decisions to include biodiversity conservation in their enterprise are affected by fluctuating market conditions, and how they respond to two different policy scenarios aiming at biodiversity conservation. The first policy scenario includes a fixed compensatory payment per hectare, irrespective of the location and spatial configuration of the parcel in the landscape. In the second policy scenario spatially dependent payments are given, depending on the contribution of parcels to species habitat networks. Results show that during periods of large price falls, farmers have the tendency to switch to biodiversity conservation contracts as they try to secure their income. Parcel size, quality and distance to the homestead show to be decisive characteristics in the decision making process. Furthermore, it is shown that whenever policy makers aim at achieving the highest amount of contracted hectares, a fixed payment is preferable. Whenever they aim at the highest contribution to ecological resilience, they should switch to spatially dependent payments. Chapter 6 explores potentials and limitations of operationalizing resilience in rural SESs through measurable indicators. The resilience criteria proposed in Chapter 2 are translated into measurable indicators and their behaviour in a rural SES in times of disturbances is tested using the agent-based model discussed in Chapter 3. The approach is illustrated by comparing two management regimes, namely self-governance and hierarchical governance, in two experiments aiming at conserving biodiversity in the case study area Winterswijk. The first experiment focuses on agri-environment schemes, compensating farmers to preserve biodiversity. The second experiment concerns the compulsory 7% Ecological Focus Areas measure, which is part of the European Commission proposals for the Common Agricultural Policy for the 2014-2020 period. The model and the proposed resilience indicators can assist rural policy makers to evaluate to what extend management strategies contribute to resilience in rural SESs. Chapter 7 provides a discussion of the findings from Chapter 2-6 and provides a reflection of the scientific added value of the collection of chapters and what their main messages are to the rural policy makers and managers.This thesis contributes to scientific literature by operationalizing the concept of resilience for use as a management concept in rural SESs.Only few studies empirically operationalize resilience in agent-based simulations, and this approach builds on existing literature by including indicators that covered the interdisciplinary character of the rural SES,not limiting the scope to ecological indicators only. The spatially explicit agent-based model builds on existing ABMs used to operationalize resilience by emphasizing more on the economic, ecological and social dimensionsof rural SESs, including a more realistic natural environment and capturing more social interactions between actors. Furthermore, the output of the model was validated in a sensitivity analysis comparing one-at-a-time and Monte Carlo approaches; this showed the added value of combined approaches for the understanding of nonlinearities and interactions in the model. To rural policy makers and managers, this thesis provides arguments for acknowledging resilience in the policy making process, while accepting sustainable development of rural areas as a main framework for decision-making in the EU. The resilience criteria proposed in this thesis can provide a starting point for discussion among policy makers, managers, scientists and rural stakeholders about which aspects of the system are of vital importance for the survivability of rural SESs in times of disturbances, and need policy support. The spatially explicit agent-based model can subsequently serve as a useful tool to explore the contribution of potential rural policy scenarios to the resilience of the rural SES, while disturbances are imposed.