Extreme event counterfactual analysis of electricity consumption in Brazil: Historical impacts and future outlook under climate change

Abstract

Climate change and other disruptive events have a significant impact on the electrical grid, affecting both power supply and consumption. With the rise in frequency and severity of extreme events, like heatwaves and droughts, the stability and operations of the system are increasingly at risk. The COVID-19 pandemic, unrelated to climate, has also brought about dramatic shifts in global energy patterns. We apply machine learning to model electricity consumption counterfactuals for Brazil, one of the largest hydropower producers, to understand the effects of these events. By training our model on 23 years of data (1999–2021), we achieved a .848 R2 and 2.6% MAPE. This enabled us to assess the impact of historical events on electricity consumption at both hourly and daily levels. Next, we use climate change scenarios to forecast electricity consumption and find that Brazil’s capacity is unlikely to meet demand from 2070 on-wards. Our research provides much needed insight into the impact of extreme events on Brazil, with implications for understanding energy system responsiveness and resiliency. The counterfactual approach proposed is also transferable to other countries and contexts, with the potential for new application areas given interactions between extreme events, climate change, and transitioning energy systems.