基于雷达三维组网数据的对流性地面大风自动识别

应用雷达三维组网数据和地面加密自动站风场资料，统计分析了对流性地面大风的6个主要雷达识别指标：风暴最大反射率因子、风暴最大垂直积分液态水含量、垂直积分液态水含量随时间变率、风暴最大反射率因子下降高度、风暴体移动速度和垂直积分液态水含量密度等参数。根据雷达识别指标和地面大风的相关程度，给出了识别指标的隶属函数和权重系数；采用不等权重法，建立了具有模糊逻辑的对流性地面大风识别方法。并将对流性地面大风的出现概率分为3级：当识别风暴单体的判据小于0.3时，出现对流性地面大风的概率小；当判据在0.3—0.5时，产生对流性地面大风的概率较大；当判据大于0.5时，出现对流性地面大风的概率很大。通过对河北省2012年6月21日线状风暴和2009年7月23日孤立单体风暴引发的灾害大风典型个例的识别效果检验，证明这种方法识别到的风暴单体跟踪效果良好，识别出的大风范围与实况风场基本吻合，命中率、虚警率和临界成功指数分别达81.8%、25.0%和64.3%，利用模糊逻辑原理建立对流性大风的识别算法是切实可行的。

Automatic identification of ground thunderstorm gale based on the radar mosaic 3D data

Based on the radar mosaic 3D data and ground automatic wind data, the six main radar identification indices of ground thunderstorm gale are statistically analyzed: storm maximum reflectivity, storm maximum vertical integrated liquid water content (VIL), time change rate of vertical integrated liquid water, dropping height of the storm maximum reflectivity, storm body movement speed and VIL density. According to the correlation of the radar identification index with surface wind, membership functions and weighting coefficients of each identification index are given. Adopting the unequal weighting method, an identification method of thunderstorm gale based on the fuzzy-logical principle is established. It has been distingished into three levels according to the probabilities in the paper: the probability of the occurrence of thunderstorm gale is low when the criterion used in the method is less than 0.3; the probability of gale is high when the criterion is between 0.3-0.5; the probability leads to being very high when the criterion rises to more than 0.5. Two typical cases in Hebei Province are analyzed and tested, one caused by a line thunderstorm which happened on 21 June 2012, and the other caused by an isolated single-cell storm which happened on 23 July 2009. The results show that the tracking effect of this method to identified storm is good, the identified wind range is consistent with the real one, and the hit rate, the false alarm rate and the critical success index reaches 81.8%, 25% and 64.3%, respectively. It also shows that the automatic identification established by the fuzzy-logic principle is feasible.