Projecting climate change effects on coupled natural/human systems at local landscape extents is crucial for land use planning and policy development. We describe an approach to modeling interactions and feedbacks between human and natural systems under the joint influences of climate change and human population growth. We couple a biophysical model of how climate change affects forest succession and wildfire in former prairie/savanna grasslands with an agent-based model of how decision makers on individual land parcels alter land use and management in Oregon’s rapidly urbanizing Willamette Valley. Unlike conventional predict-then-act approaches that seek an optimal solution that performs “best” under expected conditions, our explore-then-test approach allows us to: a) explore large numbers of potential future landscapes; b) seek robust alternatives for reducing wildfire risk and biodiversity loss that perform well across a broad range of plausible futures; and c) identify land management policies to conserve and restore imperiled ecosystems while meeting societal goals for human safety and economic well-being.
Results/Conclusions
We present a spatially-explicit modeling system for investigating interactions and feedbacks among human, successional and climatic processes of landscape change. This required solving theoretical and technical issues to integrate different simulation models. We have 1) Downscaled from the 64-ha resolution predictions of the global dynamic vegetation model MC1 to finer-scale outcomes (0.25-5 ha) for plant growth and succession by relating changes in NPP from MC1 to site index, a productivity measure in the forest succession model FVS; 2) Incorporated effects of longer-term climatic trends and short-term fluctuations on plant growth and succession through a state-transition model parameterized using MC1 and FVS; 3) Surveyed local landowners and used statistical analyses to parameterize ENVISION agents to simulate landowner behavior on different land parcel types in response to climate change, land use regulation and incentives, land markets, perceived fire hazard, management costs, and aesthetic preferences; 4) Created a vector-based GIS data layer of ~160,000 polygons for two 100-km2 study areas. The geometry accommodates potential changes in spatial pattern due to parcel subdivision, land management, succession and fire over a 50-year time frame. It includes a suite of attributes for modeling interactions and feedbacks among these processes; 5) Incorporated the fire model FlamMap as a plug-in to ENVISION to model feedbacks between human, climatic and successional influences on fire behavior; and 6) Developed a simulation framework that incorporates multiple types of system uncertainty due to variable climate model predictions, potential ecosystem responses, and human decisions.
