突发性强对流天气快速识别预警改进方法

传统多普勒天气雷达强对流灾害性天气监测采用固定阈值判别法给出强风暴的冰雹闪电灾害预警结果,该方法不适用于不同经纬度、季节和复杂地形条件下的强对流天气识别预警。本文利用循环递归的区域生长法对TITAN算法进行改进,从而快速识别三维强风暴单体及其雷达特征物理量;使用多普勒天气雷达和TRMM星载气象雷达的历史观测数据反演河北石家庄地区春夏两季复杂地形条件下的强风暴灾害性天气Logistics多元线性回归概率预警模型。对发生在河北石家庄夏季的一次强飑线天气和发生在春季的一次超级多单体风暴天气进行冰雹闪电灾害性天气识别预警实验,并与传统算法进行误差对比分析。实验结果表明:与传统算法对比,该方法对强风暴天气识别预警的定位精度较高,并且其漏报率和虚报率较低,有助于快速识别预警强对流灾害性天气。

Improved Method for Rapid Identification and Warning of Sudden Severe Convection Weather

A fixed threshold discriminance method was used in the conventional Doppler weather radar for severe convection disaster weather monitoring to get the warning results of hail and lightning disaster in severe storms,which is not applicable to the identification and warning of severe convective weather in different latitude and longitude,seasons and complex terrain. In this paper,the TITAN algorithm was improved by the recursive regional growth method to quickly identify the three dimensional strong storm and its radar characteristic physical quantities. The historical observation data of Doppler weather radar and TRMM space-borne meteorological radar were used to retrievel the severe storm disaster weather Logistics multivariate linear regression probabilistic warning model under the conditions of complex season and topography in Shijiazhuang,Hebei province. Doppler radar identification and warning experiments of hail and lightning were carried out for a strong squall line in summer in Shijiazhuang and a super multi-cell storm near in spring in Shijiazhuang.The experimental results showed that the proposed method has higher identification accuracy based on the azimuth and distance than the traditional algorithm. Moreover,the false negative and false report rates of the proposed method are low,which helps to quickly identify and warn against severe weather.