资源环境科技发展态势分析平台

Precipitation products are good choices to complement ground observations in hydrological research, but their accuracy is uncertain in different areas. This study aims to evaluate the systematic error characteristics of four major precipitation products, namely, Climate Prediction Center MORPHing technique(CMORPH), Climate Hazards Group Infrared Precipitation with Stations (CHIRPS), Global Land Data Assimilation System (GLDAS), and Tropical Rainfall Measuring Mission(TRMM) 3B42 v7, over the Yangtze River Basin in terms of estimating precipitation amounts and detecting events. The results show that the precipitation products have high spatial and seasonal heterogeneity of error characteristics, and the capability to capture rain occurrence decreases when rainfall intensity increases. GLDAS demonstrated the poorest performance, with the lowest correlation of 0.08 and the largest relative bias of over 25% underestimation. The possibility of GLDAS missing medium and heavy rains (>15 mm/d) reached 50%, and of falsely reporting light rainfall was up to 40%, while CMORPH outperformed the others with the highest consistency (0.39) against the gauge, the smallest root-mean-square error (RMSE) (10.28 mm), and the highest scores for most subregions. Generally, in this study GLDAS proved its inferiority to satellite-based precipitation products for hydrological applications over the Yangtze River Basin.