Evaluation of the IMERG version 05B precipitation product and comparison with IMERG version 04A over mainland China at hourly and daily scales

Integrated Multi-satellite Retrievals for Global Precipitation Measurement mission (IMERG) is a quasi-global precipitation product to provide high-quality and fine-scale precipitation estimates. Recently, there have been some significant improvements of IMERG's retrieval algorithm, generating new half-hour precipitation estimates (Version 5B, termed as IMERG 05B) with 0.1 degrees x 0.1 degrees spatial resolution. However, there are few studies on revealing the characteristics of the errors of IMERG 05B at high spatial and temporal scales. In this study, the performance of IMERG 05B is assessed at the hourly and daily scales against dense rain gauge measurements in mainland China from May to September 2015. Furthermore, IMERG Version 04A (IMERG 04A) is introduced as a comparison of IMERG 05B to test whether it can provide better precipitation estimates. At hourly scale, IMERG 05B is weakly related to the rain gauge measurements (coefficients of determination (R-2) < 0.12). The bias mainly consists of false bias and missed bias due to mis-classified rain areas. On the other hand, IMERG 05B provides better precipitation estimates at daily scale (R-2 = 0.49). And the bias is mainly hit bias caused by wrongly estimated precipitation intensity over correctly classified raining regions. Compared with IMERG 04A, the IMERG 05B provides more accurate precipitation intensity estimates, but contains more false alarmed precipitation at both hourly and daily scales. Therefore, IMERG 05B shows improvements on estimating precipitation in mainland China compared with the previous version. We believe an analytical study of its errors will provide useful information for further improvement of its retrieval algorithm. (C) 2019 Published by Elsevier Ltd on behalf of COSPAR.