Spatial and temporal evaluation of global rainfall products in a data-scarce region: The Dez Basin, Iran

Abstract

The limitation in approachability to rainfall data sources with an appropriate spatial-temporal distribution is a significant challenge in different parts of the world. The development of general circulation models and mathematical algorithms has led to the generation of various rainfall products as new sources with the potential to overcome the shortage in datascarce basins. In this study, the performance of the PERSIANN-CCS and CMORPH satellite-based rainfall product, as well as the ERA5 and ERA-Interim reanalysis, was evaluated based on detection skill and quantitative metrics in a daily, monthly and seasonal time scales in the Dez basin located in the southwest of Iran. The basin has a wide topographic variation and scattered rain gauge stations. Overall results denote that the ERA5 dataset has the best performance in all statistic verification than other rainfall products. Based on the daily evaluation of all rainfall products, the false alarm rate (FAR) is higher than 0.5, so none of the datasets could capture the temporal variability of rainfall occurrence. This study has covered the western parts of the Zagros steep slopes in which the topographic conditions have a significant effect on the activity of rainfall systems. On a monthly scale, the mean value of the correlation coefficient (CC) for ERA5, ERA-Interim, PERSIANN-CCS, and CMORPH was equal to 0.86, 0.85, 0.51, 0.39, respectively. The results of seasonal evaluation suggested that all datasets have better rainfall estimation in autumn and winter, and the capability of all datasets dramatically decreased in the spring. The current paper argues that the ERA5 reanalysis typically outperforms ERA-Interim and can be considered as a reliable rainfall source in the future hydrological investigation in the southwest of Iran.