Evaluation of satellite precipitation retrievals and their potential utilities in hydrologic modeling over the Tibetan Plateau

In this study, we evaluate four widely used global high-resolution satellite precipitation products against gauge observations over the Tibetan Plateau (TP). We also investigate the capability of the satellite products in streamflow simulations using the VIC hydrological model. Results show that the 3B42 and CMORPH perform better than the 3B42RT and PERSIANN at both plateau and basin scales. The 3B42RT and PERSIANN considerably overestimate the gauge precipitation estimates almost over the entire plateau, and the PERSIANN fail to capture the spatial and temporal pattern of the gauge precipitation estimates. For different satellite estimates, the error sources are systematically different for various seasons. For the 3B42, the miss bias is the main problem. The CMORPH exhibits obvious negative hit bias and miss bias in the rainy season and false-rain bias in the non-rainy season. The total bias in the 3B42RT and PERSIANN mainly attribute to positive hit bias in the rainy season and false-rain bias in the non-rainy season. The 3B42RT and PERSIANN show little capability for streamflow simulations over the TP, while the CMORPH exhibits an encouraging potential for hydrological applications in this regions in spite of the general underestimates. The 3B42 shows comparable performance to the CMA (China Meteorological Administration) data in both monthly and daily streamflow simulations mostly due to the monthly gauge adjustment involved in it. (C) 2014 Elsevier B.V. All rights reserved.