SVM方法在北京地区夏季晴雨预报中的初步应用  

Application of SVM method to summer clear-rain forecast in Beijing region

作　　者：何娜[1] 付宗钰[1] 赵玮[1] 吴进[1] 吴剑坤 廖晓农[1] 

机构地区：[1]北京市气象台,北京100089

出　　处：《暴雨灾害》2013年第3期284-288,共5页Torrential Rain and Disasters

基　　金：北京市气象局业务研发专项(BMBYWFZ2011002);北京市气象局预报员创新团队建设计划

摘　　要：以常规观测降水资料、T639模式输出产品为基础,利用支持向量机(SVM)方法分别建立了北京地区延庆、密云、房山和观象台4个站点的夏季晴雨预报模型,并用独立的样本对预报模型进行了业务试运行检验。检验结果表明,经过SVM方法学习后建立的站点晴雨预报效果优于T639模式直接输出的晴雨预报效果,08—20时段和20—08时段4站晴雨预报平均准确率分别提高了20.06%和9.45%。SVM方法对T639的晴雨预报效果有显著的改进,对北京地区夏季晴雨预报有一定的客观参考价值。Based on precipitation data from conventional observations and the T639 Numerical Weather Prediction （NWP） data, the clear-rainy forecast models are established on Yanqing, Miyun, Fangshan and Guanxiangtai in Beijing by the Support Vector Machine （SVM） statistical method. Independent samples are used to validate the clear-rainy forecast models. The results show that, the clear-rainy forecasted by SVM are superior to that of T639 DMO （direct model output）. For the periods from 08 to 20 and 20 to 08, the accuracy of clear-rainy fore- cast is improved by 20.06% and 9.45%, respectively. The accuracy of clear-rainy forecast is remarkably improved by SVM, which provides an objective way in clear-rainy forecast in Beijing region.

关 键 词：支持向量机(SVM) T639 反距离权重 晴雨预报 

分 类 号：P457[天文地球—大气科学及气象学]