Abstract
Climate change is altering the intensity, frequency, and characteristics of tropical cyclones, amplifying risks to educational infrastructure across the U.S. Gulf and Eastern coasts. Schools are increasingly vulnerable to Prolonged Unplanned School Closures (PUSCs), defined as unexpected shutdowns lasting seven or more instructional days. While previous research has examined localized school disruptions, few studies have linked evolving hurricane dynamics to large-scale projections of future closure risk. Building on prior work that developed a machine learning framework for characterizing hurricane-related PUSCs between 2011 and 2019, this study integrates climate projections to estimate how closure probabilities may shift under late-century warming scenarios. Using synthetic tropical cyclone data generated through dynamic downscaling of global climate models under a high-emissions scenario, we assess 5,431 school districts along the Gulf and Eastern coasts. The model incorporates hazard intensity, district capacity, and socioeconomic indicators to estimate two metrics: (1) the average annual probability of at least one PUSC and (2) the anticipated cumulative number of prolonged closures over a 30-year period. Findings highlight regional hotspots where increasing cyclone intensity could drive more frequent and prolonged school closures. Across the six climate models, localized coastal hotspots show absolute increases exceeding 0.30 in the annual probability of at least one PUSC, including districts in Florida. Consistent with these probability shifts, the projected cumulative burden of prolonged closures over a 30-year period increases substantially in the most exposed coastal regions, with model-specific median increases on the order of tens of additional PUSCs in high-risk states. This study presents the first probabilistic, climate-informed assessment of future PUSC risk, offering a scalable, data-driven framework to support resilience planning for the nation’s educational systems.