Evaluating global precipitation datasets over Sicily: From daily estimates to extreme events

Sicily, the largest island in the Mediterranean, exhibits complex hydro-climatic variability due to its diverse topography and proximity to contrasting maritime and continental influences. Accurate precipitation data are therefore essential for reliable hydrological modeling, disaster risk management, and climate-related assessments in this region. This study evaluates the performance of 11 widely used global daily precipitation datasets, satellite-based (GPM IMERG, TRMM, PERSIANN family), reanalysis (ERA5-Land, GLDAS-2), and blended products (MSWEP, HydroGFD), against ground-based observations across Sicily for the 2003–2023 period. A suite of statistical metrics (MAE, RMSE, KGE, Pearson’s r ) was used to assess accuracy, precision, and bias. Additionally, the capability of MSWEP to represent extremes was analyzed using the Generalized Extreme Value (GEV) distribution and Peaks Over Threshold (POT) method. MSWEP consistently delivered the highest overall accuracy, followed by ERA5-Land and HydroGFD. All datasets effectively captured general precipitation patterns but struggled to reproduce high-intensity events, particularly in mountainous areas such as the northeastern slopes of Mount Etna. While MSWEP accurately represented moderate extremes, it underestimated the magnitude and frequency of severe events. These results emphasize the importance of local calibration and validation when applying global datasets in hydro-climatically complex regions. Although the 2003–2023 record limits long-term climatic inference, the findings provide practical guidance for short-term hydrological studies and dataset selection in Mediterranean settings, contributing to improved flood assessment and water resource planning. • Evaluated 11 global daily precipitation datasets across Sicily. • Compared datasets against dense ground-based monitoring network. • MSWEP showed highest overall accuracy, ERA5-Land and HydroGFD followed closely. • Datasets underestimated high-intensity rainfall areas. • MSWEP's extremes were analyzed via GEV and POT methods to test peak rainfall skill.