Experiment And Research Of Storm Scale Ensemble Forecasting Based On Ideal Case

For the storm scale systems, the "single" determinate forecast is very difficult at the present time. Considering the tremendous success of using the ensemble forecast technique to predict the synoptic scale systems, it is a reasonable choice to introduce the ensemble forecast technique to the storm scale systems.This study is based on an ideal storm simulated by WRF model, we firstly produce a control prediction by adding random perturbations to the initial wind and temperature fields of the ideal storm, then study the error structure of the storm and the development and evolution of initial perturbation errors. The results show that, the development and evolution of initial perturbation errors have a close relationship with the regions of the convective instability and the horizontal wind fields.Secondly we do the ensemble forecast experiments by applying the Monte-Carlo method and BGM(Breeding of Growing Modes) method to control prediction respectively. Furthermore, we compare and analyze the results of these experiments with the ideal storm and control prediction. The ensemble experiments with Monte-Carlo method show that, the prediction effects of different ensemble schemes are different according to the ensemble mean and the ensemble spread of rainfall, some schemes can improve the control prediction, but the others can not. The prediction effects of ensemble schemes depend on the choice of perturbation of different variable and amplitude. The results of BGM methods are better than Monte-Carlo methods and control prediction, and reflect the uncertainty of location and strength of the rainfall better. Like as the Monte-Carlo method, the prediction effects of BGM method also depend on the choice of variable and amplitude of perturbation.Finally make an evaluation to these two methods, show that the BGM method is better than the Monte-Carlo method on the storm scale ensemble forecasting. Because the storm is very sensitivity to the perturbation of different variable and amplitude, so it is very important to choose an appropriate perturbation variable and amplitude for storm scale ensemble forecasting.