All-weather Assimilation Technology Of FY-3C Microwave Observation Data And Its Application In Typhoon Forecasting

Satellite observations play an important role in modern numerical weather prediction(NWP).Satellite data assimilation is transitioning from clear-sky to all-sky approach at some operational forecasting centers;the all-sky approach directly assimilates observations under clear,cloudy and precipitating conditions.Since the model cannot simulate cloud and precipitation with a precise enough occurrence time,intensity or location to give a good fit to observations,the first guess departures(O-B)are more non-Gaussian when assimilating the cloud-and precipitation-affected observations.However,data assimilation assumes the O-B is unbiased and Gaussian distributed,thus,the observations in cloudy and precipitating situations could not be directly assimilated.In order to make good use of the cloud-and precipitation-affected observations and improve the forecast of high-impact weather processes associated with the cloud and precipitation,assuming the O-B is state-dependent,Geer and Bauer purposed an observation error model in which the observation error is varies as a function of cloud amount,and it can effectively assign the observation errors to the cloud-and precipitation-affected observations.This achieves the goal of assimilating the satellite microwave observations in all-sky conditions.FengYun-3C Microwave Humidity Sounder-2(FY-3C MWHS-2)instrument contains 118 GHz and 183 GHz channels,which can provide abundant information about the atmospheric temperature and humidity as well as water vapor,and the 118 GHz channels are the first time to be applied on the polar-orbiting meteorological satellite in the world.To assess the impact of all-sky assimilation of FY-3C MWHS-2 observations on the forecasting of the tracks and intensities as well as precipitations of typhoons,the Weather Research and Forecasting(WRF)data assimilation(WRFDA)system is applied and extended in the study to meet the algorithm requirement of all-sky assimilation.It includes:(1)use the fast radiative transfer model RTTOV-SCATT as the radiance observation operator;(2)retrieve the emissivity over land to provide a more realistic backgrounds;(3)develop the FY-3C MWHS-2 quality control scheme;(4)implement the FY-3C MWHS-2 observation error model.In addition,three experiments(i.e.,no radiance assimilation,clear-sky assimilation and all-sky assimilation)are carried out to evaluate the impact by selecting typhoons(i.e.,Typhoon Nida in 2016;Binary Typhoons Haitang and Nesat in 2017;Typhoon Hato in 2017;Typhoon Mangkhut in 2018)which made worse influences on the southeast of China in the last three years.The results show that the distribution of O-B became Gaussian when the observation error model was used to assign the observation errors to the microwave observations in cloudy and precipitating areas.The assimilation of the cloud-and precipitation-affected observations made a positive impact on the description of the vertical structures of typhoons and the forecasting of external surrounding circulations.In all-sky assimilation,the average track errors were reduced 17.9% and 9.6% as compared with the no radiance assimilation and clear-sky assimilation,and it could make more accurate predictions on the landfalls of typhoons in China.Also,all-sky assimilation made a good performances on the intensity forecast of typhoons,especially the period after typhoons made landfall.On the other hand,all-sky assimilation has a significant impact on the forecasting of amounts and areas of rainfall(larger than 100 mm and 250 mm),indicating that it has a remarkable ability in heavy rainfall forecasting.In general,all-sky assimilation has an evidently positive impact on the forecasting of tracks and intensities as well as precipitations of typhoons,which is of great significant to improve the forecast of high-impact weather processes and make a success of disaster prevention in advance.