UC Berkeley

Working landscapes face intensifying pressures from global and local environmental, socioeconomic, and governance changes. These landscapes represent social-ecological systems composed of natural and anthropogenic ecosystems, the human communities that use them, and interactions between the components of the system and the wider world. Disentangling the drivers of change in these complex systems poses conceptual and methodological challenges, but improved understandings of these systems’ interactions and feedbacks may enable humans to manage working landscapes to provide biodiversity conservation and sustainable livelihoods in the face of rapid environmental change. My dissertation integrates theories and methods from conservation science, land system science, and econometrics to identify and quantify drivers of change in three distinct social-ecological systems: fire-prone forests of the western US, California rangelands subject to livestock grazing and wildfire, and a protected area in Amazon Basin experiencing deforestation.

Across three chapters, my dissertation examines three main questions: 1) What is the effect of land ownership on wildfire probability in forests of the western US? 2) How does livestock grazing impact wildfire probability in California’s rangelands? 3) How does integrating qualitative discourse analysis into land use change modeling affect model outcomes and predicted future forest loss? In the first chapter, I demonstrate that federally-owned forests are more likely to burn in wildfires than privately-owned forests and that these management effects are greater than some changes in climate variables. In my second chapter, I assess the impact of livestock grazing on wildfire probability in three regions of California and three different dominant land cover classes. I find that the impact of grazing on wildfire varies by region and vegetation type, but in some regions and land cover classes, grazing reduces wildfire probability. In my third chapter, I present a framework for integrating qualitative discourse analysis into quantitative land use change modeling and demonstrate the benefits of this methodological integration for understanding deforestation drivers and dynamics in Jamanxim National Forest, Brazil.