Evaluating spatio-temporal variability of gridded precipitation datasets and their effect on river basin responses

This study presents a detailed investigation into the spatio-temporal consistency and hydrological utility of 18 Gridded Precipitation Datasets (GPDs) (one ground-based, seven satellite-based, two reanalysis, one reanalysis and satellite, and seven multi-source) over varying topography and climate conditions of Turkey. Considering the regional performance of GPDs by direct comparison with 757 ground stations for the entire study period (2015–2019), the consistency increases from daily to monthly time step, where daily KGE varies from 0.18 to 0.52 and monthly KGE ranges between 0.35 and 0.73. Regardless of any regional classification and time step, the tendency of GPD detection ability is high for no-precipitation events and the detection strength decreases as the precipitation intensity increases. Moreover, comparing the streamflow reproducibility of selected datasets over three snow-dominated basins, namely Karasu, Çukurkışla, and Kayabaşı, many GPDs are able to reproduce streamflow with higher accuracy when their optimum model parameter set is utilized. Overall, multi-source GPDs generally perform better both in direct comparison over various classifications and hydrological utility with different model parameterization scenarios. Furthermore, some reanalysis datasets also present relatively high performance compared to other satellite-based, reanalysis or combined GPDs.