Respective Advantages of “Top‐Down” Based GPM IMERG and “Bottom‐Up” Based SM2RAIN‐ASCAT Precipitation Products Over the Tibetan Plateau

Abstract

AbstractThe Tibetan Plateau (TP) is characterized by complex topography and heterogeneous surface cover, which makes it difficult to obtain accurate precipitation data at a regional scale. Therefore, it is important to use dense ground observations to evaluate satellite‐derived precipitation products. In this study, the top‐down‐based Global Precipitation Measurement Integrated MultisatellitE Retrievals (GPM IMERG) and bottom‐up‐based Soil Moisture TO RAIN‐Advanced SCATterometer (SM2RAIN‐ASCAT; SM2RASC) precipitation products were evaluated against 1,584 unevenly distributed rain gauges over the TP from May to October 2015. Results show: (1) The overall performances of IMERG and SM2RASC are comparable at the daily scale, with a higher correlation coefficient (CC) for IMERG and lower root‐mean‐square error (RMSE) for SM2RASC. Regarding the precipitation detectability, IMERG outperforms SM2RASC; (2) Spatially, SM2RASC shows better performance than IMERG in grasslands, but IMERG outperforms SM2RASC in forest areas, due to the low quality of the soil moisture retrieved by satellite in the areas; (3) The accuracy of both SM2RASC and IMERG is closely correlated with altitude and rainfall intensity. The SM2RASC product estimates light to moderate rainfall (5–25 mm/day) more accurately than the IMERG, but the IMERG product shows better performance for drizzle and heavy rainfall (>25 mm/day). Due to relatively frequent fluctuations of the ASCAT soil moisture product, SM2RASC frequently reports false rainfall in dry conditions. This study demonstrates the good performance of the SM2RASC product over the TP and suggests that the impact of vegetation density and topography should be considered for improving the SM2RAIN algorithm.