Session: Causal Inference in Global Change Studies: New Approaches and Emerging Opportunities
Quantifying drivers of change in social-ecological systems: Land management impacts wildfire probability in the western US
Thursday, August 5, 2021
Link To Share This Presentation: https://cdmcd.co/zY9G4K
Katherine J. Siegel, Department of Environmental Science, Policy, and Management, UC Berkeley, Berkeley, CA, Laurel L. Larsen, Geography, UC Berkeley, Berkeley, CA, Connor Stephens, Department of Forest & Wildlife Ecology, University of Wisconsin-Madison, Madison, WI, Bill Stewart, Department of Environmental Science, Policy, & Management, University of California, Berkeley, Berkeley, CA and Van A. Butsic, Department of Environmental Science, Policy, and Management, University of California at Berkeley, Berkeley, CA
Katherine J. Siegel
Department of Environmental Science, Policy, and Management, UC Berkeley Berkeley, CA, USA
Background/Question/Methods Sustainable management of complex social-ecological systems depends on understanding the effects of different drivers of change, but disentangling these effects poses a challenge. We provide a framework for quantifying the relative contributions of different components of a social-ecological system to the system’s outcomes, using forest fires in the western United States as a model. Specifically, we examine the difference in wildfire probability in similar forests under different management regimes (federally managed vs. privately owned) in eleven western states from 1989-2016 and compare the magnitude of the management effect to the effect of climate variables. We use pre-regression matching methods and logistic regression to enable causal inference. Results/Conclusions We find a greater probability of wildfires in federally managed forests than in privately owned forests, with a 127% increase in the difference between the two management regimes over the 28 year time period. Furthermore, we find that the effect of the different management regimes is greater than the marginal (one-unit change) effect of most climate variables. Our results indicate that projections of future fire risk must account for both climate and management variables, while our methodology provides a framework for quantitatively comparing different drivers of change in complex social-ecological systems.