Abstract
<title>Abstract</title> <p>This study revisits the application of Extreme Value Theory (EVT) in managing a hydroelectric dam, with a focus on the Akosombo Dam in Ghana. EVT is a statistical tool used to analyse extreme deviations from the centre of probability distributions, which is crucial in modelling rare events such as floods and earthquakes. The Akosombo Dam is a significant source of electricity generation in Ghana; however, its water levels require careful management to prevent flooding or water scarcity. The study aims to develop a model that accounts for natural truncation at high water levels, and estimates exceedance probabilities and return periods to enhance dam management decision-making. The Right-Truncated Peaks-Over-Threshold (RT-POT) model is compared with the Generalized Extreme Value (GEV) and Generalized Pareto Distribution (GPD) models. Results indicate that the RT-POT model provides more precise parameter estimates and yields a realistic upper endpoint below the dam's maximum operating level.</p>