Home Sociology Understanding Society and Natural Resources
Land Change Models as a Platform for Social Science Integration
The review and discussion in the previous sections has illustrated the importance of the social sciences for studying land change processes. Often, the social sciences have taken the narrative or empirical approach for studying land change. The modeling perspective is often dominated by natural scientists and in many models the social drivers of land change are underrepresented. This underrepresentation can, to some extent, be attributed to the lack of spatial data representing the social drivers. For the physical factors such data are often better available, e.g. soil maps and climate data. At the same time, the poor representation of social science in land change models is due to the difficulty to generalize social science findings outside the context of a specific case study, and the lack of an overarching theory of land change that includes the social dimensions. Still, there are several advances and prospects that allow land change models to act as a platform for social science integration in natural resource studies.
It is not likely that the complete richness of human-environment interactions leading to land change will easily and completely be described by one single, all compassing theory that can inform the design of land change models. Different existing theories describe specific land change processes and are valid under specific conditions or at a specific scale; together the different theories help explain part of the total variation in human-environment interactions leading to land change. The combination and integration of narrative research with empirical investigations will help to better define the conditions under which certain land change processes occur and when theories and conceptual models are valid. Such understanding will help to define under what conditions land change models based on these conceptual relations can adequately capture the system dynamics. The complementary use of narrative, empirical and model-based explorations requires the interdisciplinary collaboration and exchange of insights across the different research perspectives and disciplines. Land change models may be designed based on the narrative and theoretical understanding of the human-environment interactions in a particular context. At the same time, social science perspectives may be formalized by representing them in simulation models, enabling to test the implications for system dynamics.
Besides interdisciplinary collaboration it is also required to broaden the perspective of the individual disciplinary approaches. The most effective way to reap the benefits of more deductive work is not to rigidly 'go deductive' and stay there. Such a 'process-led approach' may blind the analyst to alternative processes at work (Overmars et al. 2007). Rather, the message should be that researchers will profit most from developing a consciousness of the whole spectrum between the inductive and deductive extremes, and an awareness of the advantages of the variation in research routines, and then seeking the most fertile sequences and interactions between inductive and deductive work. Ultimately, this will contribute to theory development in the field of land change while at the same time helping the development of modeling tools to explore the dynamics in land systems and possible responses to policy interventions.
The lack of social science integration in models of land change is exemplified by the, often, very simplistic representation of human-environment interactions in operational land change models which does not do justice to the complexity of decision making. Especially at larger spatial and temporal scales models assume in most cases profit optimizing strategies at the level of either spatial units or for highly aggregate representative agents. In contrast, at local scales much advancement has been made in the representation of human behavior and decision making in agentbased models of land change. Ignoring spatial and temporal variation in decision making and responses to environmental change leads to inaccuracies in global assessment outcomes and difficulties in using these models to design place-based natural resource management and adaptation and mitigation strategies. The upscaling and/or outscaling of agent-based models of land change is restricted by the lack of empirical data to support the parameterization of the human-environment interactions in these models. Such parameterization requires insight in the diversity of diverging decision making models and the contextual conditions that may explain such diversity. To better include such social science information a promising direction is the re-analysis of existing case-studies and social science surveys to identify commonalities across locations as well as the role of context. The use of meta-analysis to achieve some of these objectives has revealed that information reported in case studies is often restricted and incomplete to make a full comparative analysis possible. Moreover, as narrative and econometric case studies are not conducted following a common structure or reporting protocol the necessary information to make a systematic review across case studies is often lacking in the scientific reports. Common reporting protocols to ensure that information is consistently documented have been successful in the individual-based and agentbased modeling communities. The ODD reporting protocol of individual-based and agent-based models (Grimm et al. 2006, 2010) is now common as supplementary material of all individual-based and agent-based model papers in peer-reviewed journals. A similar documentation protocol has been proposed by Seppelt and others (2012) for documenting ecosystem service assessments. If land change case-studies would apply similar documentation standards a wealth of information on land change processes and the underlying human-environment interactions worldwide would be disclosed.
Another constraint for parameterizing agent-based models is the limited information that standard land change case studies provide on the cognitive aspects of land change decisions (Meyfroidt 2012). Many studies describe the ways in which driving forces relate to land change decisions without considering the underlying cognitive processes and the way in which decision making adapts to changing conditions, including learning. While the investigation of such mechanisms is normally the field of environmental psychology, such insights are essential to understand transitions in decision making as are likely to take place under increasing influence of global markets, changing policy environments and climate change. Land change is happening in a dynamic socio-economic and environmental context, leading to dynamic decision making patterns in which we have yet insufficient insights.
Spatial simulation models are frequently used to reconstruct historic land changes (Klein Goldewijk et al. 2011) and explore future changes or evaluate the land change consequences of alternative policies. The comparison of simulation results with reality provides a measure of the extent to which we understand the humanenvironment interactions resulting in land change (Castella and Verburg 2007; Pontius et al. 2008). The wide diversity in modeling concepts and implementations serves the variation in research and policy questions as well as the different scales of analysis. Adequate land change models require the integration of social science perspectives and multi-agent models are an example of the possibility to do so. However, the challenges for better understanding and integrating human-environment interactions in land change models are still manifold. But, in the end, the development of land change models provides a platform for integrating the different disciplinary perspectives on the complex socio-ecological system governing land change. Advancing land change modeling, therefore, not only requires the efforts of individual disciplinary researchers, it especially takes the courage of all individual researchers to collaborate, contextualize findings and respond to the needs to translate findings across spatial scales.
Acknowledgments The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007-2013)/ ERC Grant Agreement n. 311819 and the project VOLANTE Grant Agreement n. 265104.
|< Prev||CONTENTS||Next >|