Dependency of errors for four global reanalysis and satellite precipitation estimates on four crucial factors

Errors in the reanalysis and satellite precipitation products are closely related to four impacting factors, i.e., seasonality, topography, rainfall intensity, and climate type. Therefore, revealing the dependency of the errors of precipitation products on seasonality, topography, rainfall intensity, and climate type is important for data users and developers to improve the quality of precipitation products and enhance their applications in hydrometeorology. In this study, the links between errors and four impacting factors were used to investigate the dependency of errors in the ERA5 reanalysis and three satellite-only precipitation products (i.e., IMERG-Late, GSMaP-MVK, and PERSIANN-CCS) on the four impacting factors. Results from mainland China show that the errors of the four products depend on the four factors, whereas the dependency of different error indicators on the four factors differs significantly. In all seasons, ERA5 outperformed the three satellite-only precipitation products in most cases. In terms of topographic dependency, the probability of detection (POD) of ERA5 increases with topographic complexity but the opposite is true for the satellite-only precipitation products. Compared with purely infrared (IR) data, the fusion of passive microwave and IR data enhances the quality of precipitation estimates and reduces the topographical dependency of the errors. Regarding intensity-based performance, the performance of the four products in terms of the POD and normalized root-mean-squared error improves as the rain rate increases, whereas that of the three bias metrics (i.e., total bias, hit bias, and miss bias) deteriorates. In terms of the climatic dependency of the errors, the errors of the four products increase as the aridity increases in most cases. False bias contributes primarily to the total bias for the four products. More importantly, we found that the relationship between the false alarm ratio (FAR) and climate type cannot accurately characterize the dependency of false alarms based on climate type because of the interference of hit precipitation events. Based on this finding, the mean number of false alarms is recommended to replace the FAR to objectively investigate the false precipitation of various precipitation products.