| description abstract | Projections indicate that solar energy will constitute 55% of total electricity capacity by 2050 in the US. Despite solar energy’s growing importance, few studies have analyzed the risks of countrywide deployments of solar infrastructure due to extreme weather events such as hurricanes. This paper presents a probabilistic framework to evaluate the performance of solar infrastructure to generate energy during hurricanes, which often cause significant outages in the US. Our novel framework integrates recent data-driven models that capture two critical and compounding factors: transient cloud conditions that decrease irradiance and high winds that can cause permanent panel damage. We apply the framework to the 2,694 counties in the 38 central and eastern US states to elucidate the risk landscape of solar generation during hurricanes. Our results show that hurricane impacts are significant, compounding, and strikingly disproportional in the US. We show that in Florida and Louisiana, clouds rapidly reduce solar generation to 32% and 65%, respectively, of their normal levels with a return period of 100 years. Our results also show that damage to panels can induce more acute and permanent energy losses, especially in rarer storms, e.g., causing 80% more losses than hurricane clouds 2 days after landfall for 200-year events. | |
