Comparison and evaluation of gridded precipitation datasets for streamflow simulation in data scarce watersheds of Ethiopia

Abstract

The objective of this study was to evaluate the hydrologic performance of gridded precipitation datasets for streamflow simulation in Ethiopia. Four products are prominent in the literature, namely Climate Forecast System Reanalysis (CFSR), Climate Hazards Group Infra-Red Precipitation with Station (CHIRPS), Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR), and Tropical Rainfall Measuring Mission (TRMM) 3B42 Version 7 (3B42V7). The datasets from these products were compared statistically with the gauge observation dataset. Furthermore, the hydrologic performance of the products was evaluated using the Soil and Water Assessment Tool (SWAT) hydrologic model in two watersheds of the Lake Ziway basin. Four statistical and three contingency indices were used to compare the precipitation products statistically, and Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS) and the ratio of mean square error (RSR) were used to compare streamflow simulations. All the satellite-based precipitation products (CHIRPS, PERSIANN-CDR, and TRMM) performed well for monthly streamflow simulations for both watersheds. The reanalysis product CFSR performed the worst with high mean error (ME) and relative bias ratio (BIAS). The high ME and BIAS values of CFSR adversely affected the hydrologic performance and resulted in unsatisfactory streamflow simulation. Performance of the CHIRPS precipitation product in capturing daily and monthly streamflows showed that it can provide valuable precipitation estimates for use as input in hydrologic models in data-sparse regions of developing countries such as Ethiopia.