Extreme weather events and energy system resilience: Policy lessons from Brazil

The increasing frequency and intensity of extreme weather events (EWEs) pose growing challenges to electricity systems worldwide, raising critical policy questions about how power systems operate and adapt under climate stress. This study assesses the impacts of EWEs on electricity generation in Brazil, a country with one of the world’s largest interconnected power system and a generation mix largely based on renewable sources. Using an international disaster records database and operational data from the Brazilian electricity system, the analysis combines superposed epoch analysis to quantify generation responses to extreme events with an artificial intelligence approach (XGBoost) to examine how key energy inputs shape system performance during periods of climate stress. The results show that flooding events, the most frequent EWE in Brazil, have produced increasingly asymmetric impacts in recent years. While earlier periods exhibit limited effects, floods since 2019 have significantly increased hydroelectric generation while reducing photovoltaic output. The effects on hydroelectric generation were particularly pronounced in the Northern subsystem of the country. The analysis based on XGBoost modeling further highlights the increased relevance of hydrological inputs during extreme events, particularly for verified hydroelectric generation. The findings suggest that energy system resilience depends not only on the composition of the generation mix but also on how technologies and the availability of their primary energy inputs respond under extreme conditions. From a policy perspective, these results support an international agenda for electricity planning that incorporates climate stress testing, adaptive system operation, and explicit consideration of input sensitivity as central elements of resilience-oriented energy strategies.