Preliminary Study And Application Of FY4A Data In Model Hot Start

In this paper,aiming at the problem of "SPIN-UP" in the numerical model,the research and application of FY4 A data with high spatial and temporal resolution in the hot start of the model are explored based on the information initialization technology in the three-dimensional cloud.The quantitative relationship between lightning frequency and radar echo was established,and the LMIE full flash data provided by FY4 A was converted into the proxy radar echo.Combined with the cloud top pressure and cloud top temperature data of FY4 A,it was first used in the GRAPES cloud analysis system to adjust the initial cloud field of the numerical model,and make the numerical prediction by hot start,so as to analyze the applicability of FY4 A data in the prediction of severe convection.The main conclusions show as follows:(1)The information initialization technology in the three-dimensional cloud has significantly improved the content and distribution of the precipitation(cloud water,cloud ice,cloud snow,etc.)in the initial cloud field of the model,effectively alleviated the problem of "SPIN-UP",and improved the short-term forecasting ability of numerical models.(2)In the process of cloud analysis,the cloud top pressure and cloud top temperature provided by FY4 A cannot be significantly adjusted for the content of water vapor and precipitation in the initial field;Only LMIE lightning data was used to adjust cloud field information,which had a weakly positive effect on precipitation particle adjustment,but it was not enough to support the numerical model to improve precipitation prediction;After adding LMIE lightning data on the basis of multiple observations,the location of the convective cloud region can be reflected.More small and medium-scale information is provided for the initial cloud field,which can effectively improve the "SPIN-UP" problem and improve the ability of short imminent prediction of strong convective weather.(3)LMIE data was supplemented to make the distribution of the model simulated radar echoes closer to the actual observed radar echoes.In addition,the quality of radar echo and the intensity and range of precipitation forecast in the severe convection area are effectively improved within the 1-3h prediction time.It can capture strong convection signals and feedback the precipitation closer to the actual observation on a short time scale.The "SPIN-UP" problem was improved to increase the accuracy of short-term forecasts.(4)The distribution of water vapor and precipitation in the initial cloud field can be effectively adjusted by LMIE data.Its horizontal distribution is basically consistent with radar echoes,cloud top brightness temperature and actual precipitation area.Its vertical distribution is also more consistent with the strong convective center,and the precipitation forecast within 24 hours is improved.Moreover,the prediction results of heavy rain and rainstorm magnitude can be more effectively improved on the short-term intertemporal scale.