Abstract
As extreme wildfires increase in frequency and scale, understanding where and how social vulnerability intersects with fire exposure and impacts is essential for promoting equitable resilience. Compared to other disasters, the social dimensions of wildfire remain less well understood; indeed, risk is often assumed to be associated with affluence rather than disadvantage. We analyzed six domain-specific social vulnerability indices across five California regions using a housing cluster framework to examine fire exposure and structure loss from 2013 to 2022. Social vulnerability was not uniformly linked to greater exposure statewide; instead, patterns differed by region and by the dimension of vulnerability considered. Higher-vulnerability clusters tended to be higher-density and more often located in Interface rather than Intermix Wildland-Urban Interface (WUI), suggesting that landscape context mediates both exposure and consequence. Machine learning models showed that the relative importance of predictors associated with structure loss differed by level of vulnerability: fire history was most influential in lower-vulnerability clusters, while population density, structure density, and vegetation played stronger roles in higher-vulnerability clusters. These findings indicate that socially vulnerable communities occupy spatially differentiated parts of the landscape and face distinct fire risk pathways that depend upon the dimension of vulnerability considered. Integrating human and biophysical correlates of fire risk, including housing patterns, WUI configuration, and multi-dimensional measures of social vulnerability can support more targeted mitigation, preparedness, and recovery strategies for socially vulnerable populations.
