Evaluations of retrieved cloud liquid water from AMSR-E, and impacts on numerical weather prediction

Japan Meteorological Agency (JMA) uses the observations from the Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E) in the data assimilation systems for the numerical weather prediction (NWP). In the JMA mesoscale model (MSM), retrievals of total column precipitable water (TCPW) and precipitation (PR) are assimilated. A MSC (Meteorological Satellite Center) algorithm is used to retrieve them. In the global spectral model (GSM), radiances of vertically polarized channels in less-cloudy area are assimilated. But cloud-affected radiances are also used. In our NWP experiments, retrievals of cloud liquid water (CLW) from AMSR-E with the MSC algorithm were used to remove the strongly cloud-affected data and to correct the bias which is derived from the thin cloud. The most significant impact on the forecast experiment was the improvement of the temperature at 850 hPa in the tropics.