ETKF—3DVAR Hybrid Data Assimilation Scheme In An Application Of Meiyu Front Rainstorm Forecast

The severe rainstorm along meiyu front is one of the main weather disasters over the Yangtze River area. In this paper, we combine GTS conventional observation data with microwave radiance data AMSU-A&B.Based on the WRF model and its ETKF------3DVAR system, we choose a research case,the experiments of choosing different data assimilation method, such as 3DVAR or hybrid, has different impacts on the simulation of the heavy rain.The results show as follows:Firstly, in the single point test, compared with the traditional 3DVAR assimilation scheme, Hybrid single-point test fully embodies the "flow" of the set covariance characteristics, with the evolution of the situation and the change in the weather. And 3DVAR background error covariance although meet the geostrophic balance and static equilibrium, it has the characteristics of static, isotropic, and does not change with the weather situation, hi the end of the cycle of hybrid data assimilation, the ensemble members are produced by ETKF data assimilation. We examine these members by setting the average collection and the collection of discrete degree, and we find that in 200hPa, the average member field of wind can reflect the trend of the weather, and the spread member field of wind and temperature also have a good performance, the large value of the variance is consistent with the meiyu front, where the area of the convective instability.Secondly, on July 5,6,24 h precipitation rainfall simulation, control experiment, 3DVAR assimilation and Hybrid assimilation tests are simulated well in China and the rainfall distribution of the heavy rain. However, Hybrid assimilation test compared to the control experiment and 3DVAR assimilation, Hybrid assimilation test can better simulate the heavy rain center. In TS precipitation grade examination, Hybrid assimilation test also shows its good heavy rain forecast ability.In the end, hybrid assimilation method can more effectively improve the various meteorological elements (such as wind field, temperature field and humidity field, etc.), effectively adjust the initial field, and makes it more in line with the actual weather conditions. In comparing the RMS error of different assimilation schemes, we find that hybrid data assimilation method of various meteorological elements in the lower troposphere improvement is more obviously, compared with it be used to improve the quality of the meteorological elements in the upper troposphere atmosphere. In 24 h ensemble forecast, compared with 3DVAR data assimilation scheme, the hybrid data assimilation scheme also shows its superiority, especially in the simulation of vorticity and the jet axis. Besides, in the simulation of the circulation filed of 500hPa, compared with 3DVAR data assimilation, the simulation of the weather situation and strong convective weather process will look more alive. In the simulation of the first semester, compared with 3DVAR, the hybrid data assimilation can more easily reflect the characteristic of the wind field and the weather situation field. However, by comparing the differences between 3DVAR and hybrid data assimilation in the RMSE of forecast, we find that in the early 12 hours, the hybrid data assimilation is prior of 3DVAR, but after 12 hours have passed, the advantages of the hybrid data assimilation has gradually been reduced.