Assimilation Of Multi-source Observations For Severe Convective Weather

In recent years,meteorological disasters caused by severe convection occur frequently in China.Improving the prediction level of severe convection weather is not only an urgent need for disaster prevention and reduction,but also one of the most challenging frontier research topics.With the continuous improvement of weather system observation technology,assimilation of various data is an effective way to improve the accuracy of the initial field of numerical prediction and improve the level of numerical prediction of severe convection weather.Based on this background,the effects of radar,lightning and satellite emissivity data assimilation on numerical simulation and prediction of severe convective weather are studied in this paper.The main research contents and results are as follows:(1)This study evaluated and compared the performance of radar reflectance and lightning data assimilation schemes implemented in the four-dimensional Data assimilation Prediction System(WRF-FDDA)for short-term precipitation and lightning prediction.Six mesoscale convective systems(MCS)that lasted more than 7 hours in Guangdong Province in June 2020 were selected as test cases.Based on the observation results,comparative analysis and quantitative evaluation were carried out.The results show that both radar reflectance data assimilation and lightning data can improve the analysis and short-term forecast of MCS precipitation and lightning.On average,radar reflectance data assimilation is better than lightning data assimilation test for precipitation forecast.However,for lightning prediction,the data assimilation test performed better in the analysis period and 1-hour forecast.This highlights the role of lightning data in short-term lightning prediction.(2)In this study,the combined assimilation of satellite emissivity and lightning data was evaluated to enhance the numerical simulation and prediction of severe convective weather.The WRFDA-3DVar assimilation system was used to integrate three water vapor channels of Himawari-8 into the initial field at the convective scale.Then WRF-FDDA system was used to assimilate lightning data and forecast precipitation.The severe convective system in Guangdong Province on June 25,2020 was selected as an example.The results showed that:The ADA＿3h test that assimilates the three-hour satellite emissivity data can improve the precipitation forecast level to a certain extent,but some regions produce false precipitation.The LDA＿3h tests assimilating the three-hour lightning data produce under-reporting of precipitation in stratified areas and areas where lightning is not observed.The convection simulated by ADA＿3h+LDA＿3h tests combined with the assimilation of the two kinds of data is most similar to the observed convection,which improves the problems of false and missing precipitation reports in the ADA＿3h and LDA＿3h tests,and the precipitation forecast effect is significantly improved.The ADA＿3h+LDA＿3h assimilation test has obtained a higher FSS score by comparing the scores of precipitation forecasting techniques in each group,and it is found that the FSS score of precipitation increases with the enlargement of the covariance window.Compared with the experiment of assimilating only one kind of data,the experiment of assimilating two kinds of data at the same time can significantly improve the precipitation forecast effect.For the assimilation of lightning data which can only be corrected in the area where lightning occurs,it can be supplemented by geostationary satellite emissivity data with a wide range of warm and wet background fields.When there is a cloud area,the observation and inversion of geostationary satellites will be limited to a certain extent.Lightning data can provide three-dimensional graupel particle mass distribution structure in the cloud,and also supplement the assimilation of satellite emissivity data.The complementarity of the two can improve the analysis and forecast results of severe convective weather.