A Family Matter: The Use of Agent-Based Modeling and Social Learning to Promote Sustainable Family Forest solutions

Maine has over 5.6 million acres of a very important resource and landowner type: family forests. Family forests are best described as a coupled social-ecological system, having two-way feedback interactions that are complex and dynamic. As a result, they are especially difficult to model. We cannot adequately model family forests without integrating the socio-economic factors as well as the biophysical components such as the forested landscape. In addition, these landowners face pressing decisions such as harvesting, land transfer, subdivision, and public access. These issues complicate an already difficult prediction of landowner behavior, demanding a unique method of integrating all system components in a comprehensive manner. Agent-based modeling is a method of generative social science that models individualistic behavior and interprets patterns that grow from the bottom-up. This allows a unique analysis of landowner behavior, and ultimately allows researchers and stakeholders alike to better understand behavioral trends.

This study presents the use of agent-based modeling through the creation of the Forest Landowner Agent-based Modeling Experiment (FLAME). We analyze the baseline results as well as the effects of two system-wide shocks: a social-economic change (an increased tax rate) and a biophysical change (a disturbance resulting in increased tree mortality). These three scenarios were analyzed using ANOVA and MANOVA tests on harvested acres and landowner goal scores to assess landowner behavior. Finally, we review implications of agent-based modeling use for researchers, policy makers and family forest stakeholders.

In conjunction with FLAME, this study also held social learning activities throughout the model creation with key family forest stakeholders. These activities occurred in the form of three focus groups and surveys at the beginning, middle and end of the FLAME modeling process. Our research revealed four stakeholder model acceptance factors: interest, knowledge, trust, and beliefs. Furthermore, the process of including stakeholders in the development of FLAME has led to an improved model, with a more robust structure, validated trends, and identified data gaps to be addressed. This study has been the first to apply agent-based modeling to family forests, and has only begun to discover the near-endless opportunities of agent-based model applications.