Abstract: Precipitation products provides a new data source for hydrological research of data-poor areas. This paper analyzes quantitatively multi-temporal and multi-spatial scale accuracy and application potential of multi-source fusion precipitation products (CMFD and MSWEP) and an atmospheric assimilation dataset (CMADS) combined with the observation data from the ground stations in the Weihe River basin. The watershed distributed hydrological model, SWAT, is used to explore the substitutability of these precipitation products for the ground-measured precipitation data of the basin in simulations of the river daily flow. The results show that the three precipitation datasets can basically reflect the spatial distribution of precipitation over the basin, but a certain degree of deviation exists. MSWEP performs better at various time scales, and its average error at daily scale is only 0.05 mm. Simulation accuracies forced by the precipitation data recalibrated using adjustable parameters are significantly improved over those in fixed parameter scenarios, indicating a practical approach to effective simulation of the basin’s daily hydrograph. Thus, this study shows that in hydrological simulations, precipitation products have potential to replace precipitation measurements at the ground stations in the Weihe River basin.

Key words: remote sensing precipitation products, accuracy evaluation, hydrological simulation, Loess Plateau