The Study Of The Initial Perturbation Method Of Meso-scale Convection Using WRF-EnSRF Assimilation System

Data assimilation is important to provide high quality initial field of numerical weather prediction. With the augment of observation data and the improvement of observation quality, the description of the uncertain information of background and observation is more and more important. The ensemble Kalman filter (EnKF) can predict the evolution of error covariance and obtain the flow-dependent background error covariance by statistic of the ensemble samples. So, it is very important to construct reasonable ensemble using appropriate initial perturbation in EnKF. How to produce effective perturbation fields is a key problem in storm scale assimilation of EnKF. This study tested the different initial perturbation using WRF-EnSRF system for storm scale assimilation. The main results show as follows.Storm scale convection is a process of energy release. The instability of environment field and the strength of wind shear significantly affected the strength and movement path of the storm, also affect the growth rate and distribution structure of the error. Therefore, in the initial perturbation phase, to find more accurate unstable region (heat, power) becomes more important. It is a ideal solution that radar data is used to determine the convection region and the perturbation area. With the development of observation techniques and the assimilation system, other observation data, such as satellite data, sounding data and so an can be used to determine the convection area.If perturbation location is not ideal, especially in the unstable area, but the moment has no convection area (such as the downstream of real convection), will produce false convection. Therefore, the unstable region can not be generalized to the disturbance, but should be combined with observations and wind to determine.The range of perturbation in the convection zone is narrower, usually caused the spread less than in the assimilation process. Using spatial smoothing formula to smooth the perturbation area can improve the spread, and can reflect the error distribution, improve the quality of assimilation.The significance of dynamic and thermodynamic of ideal case is relatively clear, the quality of assimilation is better than the real case. The quality of assimilation of real case affected by many factors, the situation is more complicated. Different meso-scale systems, the meaning of power and thermal are also different. The perturbation method should also be changed, needs further study.This paper discusses the method of random disturbance and it's improvement in the WRF-EnSRF assimilation system to construct ensemble members. Random perturbation method is only one of many perturbation methods. As long as the perturbation method can reflects a reasonable distribution of possible errors at atmosphere and to maintain the spread between members, the perturbation method can be used in the assimilation system. In addition to the variable on the grid, use of the environment field, the terrain, the different micro-physical processes can be used to construct ensemble members, should be considered in the future study.