Evaluating the accuracy of the global precipitation products: a time-series analysis in Poland

Accurate precipitation data are essential for the assessment of global and regional hydrological processes. Several global precipitation products are now available, providing time series with high spatial and temporal resolution, using different data sets, methods and models. To ensure optimal use, it is crucial to understand the uncertainties and errors associated with these products to enable informed decision making in a range of Earth and climate science applications. The aim of this study is to evaluate the errors and uncertainties of the leading precipitation products in Poland, including TerraClimate, MERRA-2, ERA5, GPM, PERSIANN-CDR and CFSR, based on a comparison with gauge stations. Precipitation time series from 69 gauging stations were used as reference datasets. Errors in the products were assessed at various time scales, including annual, seasonal, monthly and daily, using statistical analysis. Regardless of the time scale or product, the correlation between the precipitation products and the reference dataset ranged from 0.50 to 0.87. TerraClimate showed the highest performance with an average correlation coefficient of 0.80 across all time scales, while PERSIANN-CDR showed the lowest average correlation coefficient of 0.50. The results suggest that TerraClimate could be a reliable alternative for climate data in regions with sparse ground-based stations. These results have significant implications for climate studies, water resource management and drought monitoring, as the study improves our understanding of the error characteristics of available precipitation products and provides valuable insights for refining precipitation retrieval algorithms in the future.