Development of precipitation retrieval algorithm over land for a satellite-borne microwave sounder

We develop the rain/no-rain classification (RNC) method for an over-land rainfall retrieval algorithm for the Advanced Microwave Sounding Unit (AMSU) based on the Global Satellite Mapping of Precipitation microwave radiometer algorithm. The current RNC method over land based on the window channel tends to underestimate warm rain areas due to high land surface emissivity. The sounder channels are less affected by surface emission. Therefore we developed a new RNC method using the sounder channels in order to detect the warm rain which is missed by the current method. The proposed methods are evaluated using the Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) data. The proposed method can detect the warm rain which is missed by the original method and about 84% of the AMSU-PR matched-up cases during Jan.-Dec. 2007 is improved.