Geostationary Microwave Observation System Simulation Experiments(OSSEs)Based On GRAPES?GFS 4D-Var

Because geostationary orbit(GEO)meteorological satellite is static relative to the earth,the GEO microwave(MW)atmospheric sounding has the advantage of high real-time,allweather,and large filed-of-view.Thus GEO-MW radiance data can effectively enhance the effect of numerical forecasting of convective weather conditions with quickly changing,mesoscale,and short-to medium-duration,such as typhoon and heavy rainfall.Although there is no GEO-MW atmospheric sounding payload in orbit so far due to the technical challenge of high spatial resolution in the development of GEO-MW radiometer,it had been included in China's second-generation geostationary meteorology satellite "Fengyun-4" series plan.China has already developed the various ground prototypes of GEO-MW radiometer.Geostationary orbit microwave observation is an unprecedented method of satellite observation,causing the researchers cannot guide the design of its payload from the practical application requirements,and it is also difficult to assess the impact of its observation data on the practical application.Therefore,it is necessary to carry out OSSEs in advance to provide research direction for the development of microwave loads in geostationary orbit.Since the 1960 s,OSSEs has been used in the field of numerical weather prediction(NWP),it is used to perform exploration tasks and instrument parameters in a comprehensive balanced design,satisfies the scientific research demand.The basic principle of OSSEs is to simulate the satellite observation data from the output parameters of the atmospheric numerical model,and then evaluate and verify the impact of the simulated new satellite observation data on the numerical weather forecast through data assimilation.Based on the standard OSSEs framework,this paper uses the GRAPES?GFS numerical prediction model and the GRAPES global 4D-Var data assimilation system,select the typhoon "Maria" and "Mangkhut" in 2018,simulate and generate the geostationary orbit microwave observation data with different parameters configuration to carry out the geostationary orbit microwave OSSEs experiments,take the prediction accuracy of typhoon path and intensity as the standard,analysis of the influence of the brightness temperature data of various frequency bands(50-60 GHz,118GHz,425 GHz,183GHz,380GHz),coefficient of covariance of observation error,system noise,frequency of observation data and other factors on the assimilation results.The experiment results show that the assimilation of atmospheric temperature sounding spectrum(50-60 GHz,118GHz,425GHz)of geostationary orbit microwave observations can effectively improve GRAPES?GFS model of typhoon track forecast accuracy,compared with control experiments,in the 72-hour forecast,the mean error decreased by 43 km and the performance improved by 19.7%,assimilation atmospheric humidity sounding spectrum(183GHZ,380GHZ)of geostationary orbit microwave observations will enhance the forecast of typhoon intensity,to a certain extent,10.5% increased accuracy,however,there was no significant effect in the forecast of typhoon track.The covariance coefficient of the observation error,the noise of the observation data is closely related to the quality of the observation data.The experiment shows that the covariance coefficient of the observation error represents the "confidence degree" of the observation data and is negatively correlated.The smaller the noise of the observation data is,the better the quality of the observation data will be,and the more obvious the improvement of the performance of numerical prediction will be.Finally,the interval time of observation data can also have a significant impact on the assimilation effect,and the high time frequency geostationary orbit microwave observation data can effectively improve the assimilation effect,so it is of great significance to develop the application of geostationary orbit microwave load.