Power outages and social vulnerability in the U.S. Gulf Coast: multilevel Bayesian models of outage durations amid rising extreme weather

Extreme weather events and disasters, coupled with aging and underfunded infrastructure, threaten critical systems like energy grids. Disruptions in these systems following disasters have cascading impacts on population health and well-being. By combining power outage, disaster, and social vulnerability information (2017–2022) at the county level in one of the most disaster-prone regions of the United States, the Gulf Coast, we created hierarchical Bayesian models to infer profiles of regions experiencing slower restoration following disasters such as tropical storms, heat waves, snowstorms, inter alia. Hurricanes/tropical storms and heat waves were associated with the longest outages regionally. We also estimated the 10 most and least vulnerable counties based on average outages, irrespective of specific hazards. Characterizing community profiles and key vulnerabilities to prolonged post-disaster energy insecurity is critical to prioritizing energy restoration, justice-oriented energy transition policies, and sociotechnical resilience in the region. Comparing observed and predicted outages and accounting for social vulnerabilities, these results can aid planning efforts in community and policy contexts across the region.