基于最小二乘支持向量机的雷暴预报初探  

A PRELIMINARY STUDY ON THUNDERSTORM FORECAST BASED ON LS-SVM

作　　者：王振会[1,2] 张祎[1,2] 朱佳[1,2] 

机构地区：[1]南京信息工程大学/气象灾害省部共建教育部重点实验室,江苏南京210044 [2]南京信息工程大学大气物理学院,江苏南京210044

出　　处：《热带气象学报》2011年第2期161-165,共5页Journal of Tropical Meteorology

基　　金：公益性行业科研专项(GYHY200806014)资助

摘　　要：利用2007年和2008年南京地区NCEP 1°×1°历史再分析资料和江苏省闪电定位资料,探讨最小二乘支持向量机(LS-SVM)方法在雷暴预报中的应用。将NCEP 1°×1°资料作为实况,选取了该资料与雷暴相关性较好的参数作为预报因子,而江苏省闪电定位资料则作为预报量,建立南京地区雷暴预报的最小二乘支持向量机模型,并用独立样本对模型进行预报检验。初步结果表明,该模型雷暴准确预报率为78.26%,虚假报警率为21.74%,预报技术评分TS为0.61。表明最小二乘支持向量机方法在雷暴6小时预报中是有效的。LS-SVM（Least squares support vector machine） method is presented to forecast the occurrence of thunderstorm in Nanjing district by combining NCEP FNL Operational Global Analysis data on 1 °×1 ° degree grids and cloud-to-ground lightning data obtained by lightning detection network in Jiangsu Province.The predictors are selected from the NCEP data during 2007—2008,and the predictand comes from the lightning location data.Based on these data,LS-SVM model of thunderstorm-forecasting is set up.Preliminary study shows that the prediction accuracy of this model is 78.26%,false alarm rate is 21.74%,and forecasting technical score is 0.61.Therefore,LS-SVM is an effective method to make 6-hour forecasts of thunderstorm.

关 键 词：雷暴预报 最小二乘支持向量机 南京地区 闪电定位资料 NCEP资料 

分 类 号：P457.9[天文地球—大气科学及气象学]