The Development And Optimization Of METKF Initial Perturbation Scheme For GRAPES Regional Ensemble Prediction System

In order to improve the initial perturbation's quality and the forecast level of GRAPES (Global and Regional Assimilation and Prediation System) regional ensemble prediction, and describe the structural characteristics of meso-small scale weather system prediction error accurately, the tset based on the isobaric GRAPES-ETKF initial perturbation scheme is designd to optimize this scheme by introducing the atmospheric motion information in the lower troposphere. To resolve the incoordination between the existing isobaric initial perturbation scheme and the mode-space assimilation system, we also constructe a mode-space ETKF initial perturbation scheme which consistent with the GRAPES-m3DVAR system. This scheme is better to improve the quality of the ensemble disturbance and the rationality of the growing disturbance along with the forecast.. The test results suggest that the optimized scheme is effective in improving the initial perturbation's quality of regional ensemble prediction, and can reflect the characteristics of micro-meso scale atmospheric motion more accurately and reasonably. The mode-space mETKF scheme has a significant improvement on the prediction results of the model physical quantities and the performance of ensemble prediction.The main conclusions are obtained as follows:(1) The scheme which introduce the lower layer atmospheric information can not only capture the perturbation information of the micro-meso scale weather system reasonably, but also accurately describe it's structural characteristics of the prediction error and better improve the ensemble spread. This scheme reflects the uncertainty of the ensenmle prediction by giving more likely states of the atmosphere,and provides more helpful references for the forecasting. The ensemble forecast has a good effect on the precipitation forecast and the quality of the physicl field.(2) The total perturbation energy of the mode-space mETKF initial perturbation scheme increased rapidly at the initial time of prediction, and the ensemble spread is more reasonable than the original schem. It could capture the fast-growing analytical error in time and better show the characteristics that the forecast error increases with the mode foercast time. This initial perturbation scheme has obvious advantages in all kinds of forecasting scores. The results of identifiability and credibility all show that the improved scheme can improve the original ensemble forecast performance to a certain extent level.(3) The coordination between mode-space mETKF initial perturbation scheme and GRAPES assimilation prediction system effectively avoid the interpolation error,and further improve the performance of the ensemble prediction system which based on the mETKF initial perturbation scheme. It lays a scientific foundation for the upgrading and improvement of China's GRAPES-REPS operational system.