Assimilation of Satellite Track Wind and GNOS Bending Angle Data in Rainy Environment

It is very important to improve global and regional forecast skill by assimilating clear radiance of satellite into numerical model. Since tremendous satellite observations are in rainy circumstance, they could not be assimilated because of contaminated by cloud hydrometeors. The satellite track wind can be calculated through cloud observation in high time frequency by geostationary satellite. The satellite track winds which can depict the atmospheric dynamic states are more useful than those of the satellite radiance used widely in NWP model. Unfortunately, the satellite track wind of FY2 has not been used in WRF-DA at present. So, in this paper, we set up a comprehensive module to assimilate these data into WRF model. Because of quality mask involved in satellite track wind data, strict quality controls have been done for data correlation and continuity in the module. The tracer height of FY2 track wind can be calculated by an optimal matching method based on the spatial scale of motion vector and baroclinicity of atmosphere. These height can also be corrected by the information of the neighbor winds. A RUC platform has been set up to check the impacts of the satellite track wind of FY2 on 9 typhoons' track in 2014. It showed that error of typhoon's track have been decreased by 11.4%. Data of radio occultation for air temperature and humidity at a very high vertical resolution are also very important because of their high quality and less impacts from precipitation. In this paper, a ray tracing method and corresponded quality control have also been established in the RUC system to assimilate bending angles of radio occultation from FY3. A threshold from serious sensitive tests of O-B has been introduced into the quality control to correct abnormal high horizontal gradients of air temperature and humidity. Results show that the RMS of 8 typhoon tracks decrease by 14.6%. Both of the bending angles data of FY3 GNOS and track wind data of FY2 have been assimilated in the RUC platform to see the combined impacts on 2 typhoons cases in 2014. With the two satellite data sets, RMS of the predicted typhoon track in 48 hours for TYPHOON 1412 decreases by 59.1% from 135.33km to 55.38km, 16.4%, from 390.61km to 326.58km for TYPHOON 1415 in 72 hrs.