Evaluating the performance of three different gridded precipitation datasets by comparing the measurement dataset in the Susurluk basin

In recent years, the need for long-term precipitation datasets that accurately represent spatiotemporal distribution has significantly increased. To meet this need, various gridded precipitation datasets have been developed. However, before utilizing these datasets, their accuracy must be assessed by comparing them with measurement datasets. For this purpose, the performance of three different monthly gridded precipitation datasets (Climatic Research Unit (CRU), Global Precipitation Climatology Center (GPCC) and University of Delaware (UDEL)) was evaluated in the Susurluk Basin over the period 1966-2017 at monthly, seasonal, and annual scales. The measurement dataset was obtained from six meteorological stations (Band & imath;rma, Bursa, Dursunbey, Keles, Simav, and Tav & scedil;anl & imath;) operated by the Turkish State Meteorological Service. The agreement between gridded and measurement datasets was assessed using statistical indexes such as the Coefficient of Determination (R2), Root Mean Square Error (RMSE), Mean Bias Error (MBE), and Nash-Sutcliffe Efficiency Index (NSE). Among the gridded datasets, GPCC demonstrated the best overall performance, while CRU showed the lowest performance. Additionally, it was observed that the gridded datasets exhibited higher consistency with measurement dataset in low-altitude regions. During wet seasons (winter and autumn), gridded precipitation datasets showed better agreement with the measurement dataset.