基于Logistic模型的干、湿环境下江淮夏季雷暴大风潜势预报研究

利用2009—2015年江淮夏季雷暴大风观测资料和NCEP再分析资料,按整层可降水量将雷暴大风环境划分为干、湿两种环境,结果发现湿环境雷暴大风日约占总雷暴大风日数的86%。基于物理量参数和Logistic回归方法构建了江淮夏季干、湿环境下区域雷暴大风的潜势预报模型。西南区、东南区和北区湿环境雷暴大风的最显著预报因子分别是冰雹指数(CS)、K指数和沙氏指数(SI)。干环境雷暴大风的最显著预报因子是总指数(TT)。相对于大风指数(WINDEX),综合考虑热力学作用和高空水平动量信息的新大风指数(GUSTEX)对江淮干、湿环境雷暴大风的预报指示意义更好。通过历史样本回报确立了预报模型的概率阈值,并利用2016年独立样本试预报检验证明Logistic模型预报效果良好。

A study of potential forecast of thunderstorm gales in dry and wet environment in summer over Yangze-Huaihe River Basin based on Logistic model

By using thunderstorm gales observations in summer in the period of 2009—2015 and NCEP reanalysis data,this study divides the thunderstorm gales environment into dry and wet environment.The results show that the thunderstorm gales days in wet environment account for about 86%of all thunderstorm gales days.Based on physical quantity parameters and Logistic regression method,potential forecast models for regional thunderstorm gales in dry and wet environment during summer over Yangze-Huaihe River Basin are developed.The most significant predictors for thunderstorm gales in wet environment over southwest,southeast and north regions are CS,K and SI,respectively.The most significant predictor for thunderstorm gales in dry environment is TT.Compared to WINDEX,the new wind index(GUSTEX)which considers both thermodynamical effect and high-level horizontal momentum information shows better predictive significance for thunderstorm gales in dry and wet environment over Yangze-Huaihe River Basin.The probability thresholds are determined through historical sample return.The independent samples of 2016 are used to verify the model and the Logistic model is proved to have a good performance.