Correcting GSMaP through Histogram Matching against Satellite-Borne Radar-Based Precipitation

Improving the accuracy of global precipitation estimates is practically important to predict hydrological disasters such as floods and droughts. Owing to the Global Precipitation Measurement mission and its constellation satellites, surface precipitation can be estimated globally based on microwave radiometers (MWRs). However, MWR-based precipitation is known to be biased, such as overestimation over land. This study aims to improve MWR-based precipitation of Global Satellite Mapping of Precipitation (GSMaP_MWR) through histogram matching against precipitation retrieval based on satellite-borne Ku-band precipitation radar (KuPR) observations. For that purpose, we first developed cumulative distribution functions (CDFs) of hourly precipitation for GSMaP_MWR and the KuPR data for each 0.1° × 0.1° pixel in four seasons independently using data during 2015-2020. The CDF-based histogram matching successfully adjusted seasonal average precipitation of GSMaP_MWR closer to that of the KuPR products globally. Larger corrections were observed over land especially in the summer hemisphere. Validations against radar/raingauge-analyzed precipitation in Japan revealed that the histogram matching successfully improved GSMaP_MWR precipitation in general. The methodology of this study can also be applied to other agency's MWR-based precipitation estimates and their subsequent products such as gauge-adjusted global precipitation.