Simulating and discussing alternative landscape futures by means of agent-based models: insight from place-based explorations in the Mediterranean

Cultural and behavioral motives are significant root influences to landscape transitions, yet efforts to address and manage these changes are often hampered by a failure to account for the heterogeneous decision-making nature of its agents of change and the inherent complexity of socio-ecological systems. Novel techniques are required to further disentangle responses to multi-level drivers and discuss alternative landscape development trajectories.
Agent-based models present increasingly applied tools in this context, credited with the ability to dynamically explore aggregated individual and collective behavior through learning, adaptation, anticipation and interaction processes under a range of scenario conditions. Participatory approaches for model and scenario development in particular provide a promising yet underrepresented avenue. We present own findings from explorative and empirical agent-based modelling studies undertaken in the Mediterranean that address landscape transitions occurring within contexts of increased rural depopulation and climate variability. The studies aimed to identify existing socio-cultural determinants guiding land managers in their farm-based decision-making. We adopted behavioral frameworks expanding on the Theory of Planned Behavior, specifically scoping the determination of individual intentions towards land abandonment, farm commercialization and the adoption of climate-adaptive actions and practices. The spatially explicit agent-based models integrated extensive surveying of local farming communities as part of wider, iterative processes of model refinement and stakeholder consultation.
We present the farmer typologies identified in the case study areas, the role of “bottom-up” collective initiatives, institutions and agricultural policy in relation to individual decision-making, and the resulting landscape transitions modelled under contrasting scenario storylines. The agent-based models explicitly capture and illustrate landscape changes emerging from presently occurring farming discourses, rather than parameterizing variables from historical census or remote-sensing datasets. We reflect on the role of this implicit bias, and conclude with a discussion on the role and implications of stakeholder consultation and participative agent-based model refinement in shaping landscape perceptions and exploring alternative rural futures.