Impact of climate change on electricity production in the Kaleta dam, Republic of Guinea

Abstract

In Guinea, electricity remains one of the primary challenges faced by the population. The situation may become more complicated due to the impacts of climate change, which can directly influence electricity generation from hydropower dams. This study assesses the potential impacts of climate change on electricity generation at the Kaléta hydropower dam in Guinea, utilizing advanced machine learning models. Two gradient boosting algorithms, CatBoost and XGBoost, were employed to predict production flow and effective electricity production based on historical hydro-climatic data from 2016 to 2024. The models incorporated climate variables including precipitation, temperature, potential evapotranspiration, wind speed, and downward longwave radiation flux. Future projections were generated for the period 2026-2034 under two Shared Socioeconomic Pathways (SSP245 and SSP585) using downscaled CMIP6 climate model outputs. The analysis of historical trends showed significant increasing patterns in both production flow and effective production, while water levels exhibited a non-significant declining trend. Future projections indicate positive impacts of climate change on hydropower generation capacity. Under SSP245, CatBoost predicts increases of 10.28% in production flow and 13.12% in effective production, while under SSP585, increases of 9.92% and 13.25% are projected, respectively. XGBoost showed more conservative but still positive projections with increases ranging from 5.54% to 8.32% across scenarios. Anomaly analysis demonstrated that future projections generally maintain seasonal patterns while showing enhanced production capacity, particularly during wet seasons. These findings suggest that the Kaléta hydropower plant may benefit from climate change in the short term (2026-2034), with increased water availability potentially enhancing electricity generation.