Comprehensive evaluation of Ensemble Multi-Satellite Precipitation Dataset using the Dynamic Bayesian Model Averaging scheme over the Tibetan plateau

摘要

Highlights?EMSPD-DBMA well captures the spatiotemporal patterns of precipitation across the TP.?EMSPD-DBMA outperforms IMERG and GSMaP-MVK in the summer of TP.?EMSPD-DBMA outperforms MSWEP V2 across the TP.AbstractThe objective of this study is to comprehensively evaluate the new Ensemble Multi-Satellite Precipitation Dataset using the Dynamic Bayesian Model Averaging scheme (EMSPD-DBMA) at daily and 0.25° scales from 2001 to 2015 over the Tibetan Plateau (TP). Error analysis against gauge observations revealed that EMSPD-DBMA captured the spatiotemporal pattern of daily precipitation with an acceptable Correlation Coefficient (CC) of 0.53 and a Relative Bias (RB) of ?8.28%. Moreover, EMSPD-DBMA outperformed IMERG and GSMaP-MVK in almost all metrics in the summers of 2014 and 2015, with the lowest RB and Root Mean Square Error (RMSE) values of ?2.88% and 8.01?mm/d, respectively. It also better reproduced the Probability Density Function (PDF) in terms of daily rainfall amount and estimated moderate and heavy rain