Research On Recognition And Forecast Methods Of Severe Convective Weather

Severe convective weather is a kind of the most destructive weather and it is a serious threat to people's life and property safety. Doppler weather radar is one of the most important tools for monitoring and forecasting of severe convective weather. It can achieve the recognition, tracking and forecasting of severe convection weather by using radar data.Based on the radar echo reflectivity image, this paper determines the severe convective weather regions in the image. Extract the related echo features through analysis, then establish objective recognition model by using data mining technology. Based on the recognition, it purposes the storm cell tracking algorithm and obtains continuous temporal process of the severe convection weather. The ultimate task is to realize the effective prediction of severe convective weather.In this paper, the main work is as follows:(1) Transform the echo reflectivity factor of radar base data into reflectivity image. By analyzing the 1197 samples about 85 hail processes and 81 rainstorm processes, extract the image features of storm cells as well as conventional meteorological features in order to establish the sample feature database.(2) By using the method of rough set theory in data mining, the classification rules of hail and storm weather can be generated through the data mining of sample feature database, then establish the objective recognition model of the severe convective weather. This algorithm uses the standard voting decision method to realize the automatic recognition of the severe convective weather.(3) Through the analysis of the two storm tracking and prediction algorithms: TITAN algorithm and SCIT algorithm, we propose the storm cell tracking algorithm for Beijing-Tianjin-and-Hebei region. It contains the internal structure and the whole information of the storm and solves the problem of false merge between storms.(4) Using the weighted least square method, we fit the storm cell centroid trajectory in continuous scan of temporal process. It realizes the storm cell position prediction by computing the average motion vector of storm cell.In conclusion, this paper realizes the automatic recognition of the severe convective weather and improves the recognition accuracy. It completes tracking and prediction of severe convective weather based on the recognition.