An Integrated County-level Community-aware Analytics-driven Risk Evaluation (CARE) Index for Wildfire Emergency Management

Though wildfires have long posed a threat across the southwestern United States, their frequency and intensity are increasing due to climate change, poor land management, and the uncontrolled sprawling of population into the wildland-urban interface (WUI). For efficient wildfire mitigation and post-wildfire recovery, prioritizing resource allocation based on community resilience, wildfire severity, and experienced economic losses is essential. For example, rural communities who often lack basic capabilities and adaptive capacities need more resources and recovery efforts to manage the disaster than their wealthier counterparts. As resource and recovery efforts are managed by local authorities across counties, understanding the county-level risk in wildfire emergency management (WEM) is instrumental to advance risk-informed community- wildfire resource allocations. Therefore, this study aims to develop a county-level Community-aware Analytics-driven Risk Evaluation (CARE) index for WEM leveraging advanced machine learning algorithms. To that end, we collected data from multiple publicly available sources from 2015 to 2022 on wildfire hazard exposure (e.g., information on past wildfire incidents), social vulnerability of communities (e.g., socioeconomic factors, demography) and physical/economic damage caused by wildfires (e.g., structural damage, post-wildfire resource allocations) to predict WEM-CARE index under various scenarios. For our case study, we considered the southwestern US states including California (CA), Arizona (AZ), Colorado (CO), Nevada(NV), Utah (UT), and New Mexico (NM) that are exposed to frequent and severe wildfire events. Our preliminary insights suggest that although urban counties like Los Angeles are highly exposed to wildfire events, the remote, mountainous counties like Plumas, CA are more vulnerable to wildfire and require greater attention in recovery efforts due to their socio-economic attributes. The outcome also identifies socio-economic and demographic factors that contribute to counties' vulnerability to wildfire, which is an understudied aspect of current research.

Keywords: Wildfire emergency management risk, Targeted resource allocation, Statistical learning, Communityfocused decision-making.