| Xinjiang Uygur Autonomous Region is located in Eurasia hinterland. The Southern Xinjiang Region is one of the important agricultural regions in China, but it is severely influenced by meteorological disasters, especially thunderstorms and hails which caused the serious economic loss to Southern Xinjiang every year. Based on the data of 24 meteorological stations in the Southern Xinjiang from 1980 to 2013, the climatic characters of spatial and temporal distribution of thunderstorms and hails were analyzed by using methods of Mathematical Statistical Analysis. It was surveyed, analyzed and classified that the atmospheric circulation situation resulted in the occurrence of thunderstorms and hails from 2002 to 2012 in the Related Areas of Southern Xinjiang(Aksu and adjacent areas), and 4 main weather types have been summed up. Moreover, the optimality criterions of the 4 types which can be used to identify weather type automatically were determined. On this basis, convective parameters correlated with thunderstorms and hails were screened out as the forecast factors, and then the forecasting models of regional thunderstorm and thunderstorm at single station were respectively built by three methods of the comprehensive multi index superposition, the multiple stepwise regressions and the BP neural network techniQue. The forecasting models were tested by using historical data, and their forecasting effects were evaluated. Finally, by adopting the idea of Perfect Prognostic Method, the operational forecast system of thunderstorms was esta BLIshed based on the T639 numerical forecast products. The main research results are as follows:(1) The analysis on temporal and spatial distribution characteristics of thunderstorms and hails in Southern Xinjiang showed that: The thunderstorms and hails were mainly in the Tianshan Mountains, and had a banding distribution along the southwest-northeast direction. Southern Xinjiang could be divided into 4 climate zones: Aksu and adjacent areas、the western mountain areas、the eastern areas and Hotan district. Aksu and adjacent areas was the high-occurrence disaster zone for thunderstorms of Southern Xinjiang, and the thunderstorms and hails often occurred in April to October, focused on June to August, the peak was in July. The thunderstorms and hails concentrated in afternoon to the first half of the night in a day. The number of thunderstorm day of the other three climate zones less than Aksu and adjacent areas, but the characteristics of inner-annual changes and diurnal changes were similar to Aksu and adjacent areas.(2) It was surveyed, analyzed and classified that the atmospheric circulation situation resulted in the occurrence of thunderstorms and hails from 2002 to 2012 in Aksu and adjacent areas and the results showed that: the weather situation was mainly divided into 4 types: The Balkhash Lake low trough type, the jet with vertical shear type, the northwest airflow before Urals type and the temperature trough type. By constantly revise the criterion of the 4 types which can be used to identify weather type automatically to achieve optimal results. The result by using the data of 2002-2012 showed: The number of samples decreased from 1683 to 876, empty elimination rate is 48%, and the non-response rate is 2.4%; the result by using the data of 2013 showed: The number of samples decreased from 153 to 80, and empty elimination rate is 48%, without omission. According to the analysis of the weather situation and typical case, the study found that, the occurrence of thunderstorms and hails in Aksu and adjacent areas had its uniQue origin, when there was cold trough near the Balkhash Lake, as the movement of the weather system, the low trough guided the cold air across the Tianshan Mountains reach to over the Aksu and adjacent areas, at this time, low still warmer due to the Tianshan block, an unstable stratification in which it was dry and cold at the upper level, wet and warm at the lower level formed, eventually it led to the occurrence of thunderstorms and hails.(3) The screening results of forecast factors showed that: The physical Quantities such as BCAPE, DCAPE, BLI, CIN, K, TT, SI, SHR, Q, SITASE, TMPADV, TD, TDD, RH and PWAT were good predictors for thunderstorms and hails in Aksu and adjacent areas, while different weather types had different kinds of physical Quantities.(4) The forecast results of the regional thunderstorms forecasting model which was built by three methods of the improved comprehensive multi index superposition method, the multiple stepwise regression method and the BP neural network techniQue showed that: These three methods had their own advantage. In general, the BP neural network techni Que was better than the other two methods, with the TS score up to 50.0% and the lowest false alarm ratio. Exploring the method of thunderstorms probability forecast which turned out that: The more the numbers of factors met the forecast indexes or the greater the forecast values, the greater the probability of the occurrence of thunderstorms.(5) The forecast results of the thunderstorm at single station forecasting model which was built by two methods of the multiple stepwise regression method and the BP neural network techniQue showed that: For these meteorological stations with higher climatic probability, the TS scores of the two methods could reach to 30.0%; For these meteorological stations with lower climatic probability, the BP neural network techniQue was better than the multiple stepwise regression method.(6) Based on the above research result, the operational forecast system of thunderstorms was esta BLIshed in the end. The operational forecast system could automatically obtain the medium and short term numerical prediction product data, also perform the data Quality control and preprocessing. The operational forecast system could call the forecasting models automatically and provid the result of weather types identifying. The generation and display of the forecast result were all performed by the computer.The research results in this paper could provide the reference and technical support for the medium and short term potential forecasting of thunderstorms in Aksu and adjacent areas. |
PDF Full Text Request