Modeling Extreme Rainfall Using the Principle of Maximum Entropy

Abstract This research work analyzes extreme rainfall events in a mining-metallurgical region of North Greece, using available data from four meteorological stations recorded between 2006 and 2021. The Maximum Entropy method was applied to estimate the parameters of various probability distributions and identify the optimal fit for accurate risk prediction. Model evaluation was conducted using Akaike’s Information Criterion (AIC). Results indicate that the Generalized Extreme Value (GEV) distribution provides the best fit for modeling rainfall extremes. The entropy-based approach effectively captures the empirical probability distribution and demonstrates robust performance in modeling extreme rainfall across the stations. These findings highlight key periods of heightened rainfall activity, offering a valuable basis to increase the reliability of future risk assessment studies and offer valuable benefits for the mining site, such as optimum planning of mining-related activities, increased productivity, improved health and safety at the workplace, reduced probability for infrastructure damage, and development of an effective prevention strategy.