###
气象:2017,43(9):1110-1116
本文二维码信息
码上扫一扫!
基于集合预报和支持向量机的中期强降雨集成预报试验
黄威,牛若芸
(国家气象中心,北京 100081)
The Medium Term Multi Model Integration Forecast Experimentation for Heavy Rain Based on Support Vector Machine
HUANG Wei,NIU Ruoyun
(National Meteorological Centre, Beijing 100081)
摘要
图/表
参考文献
相似文献
本文已被:浏览 184次   下载 175
投稿时间:2016-03-08    修订日期:2017-07-19
中文摘要: 本文基于欧洲中期天气预报中心(ECMWF)和美国国家环境预报中心(NCEP)集合预报资料和支持向量机(SVM)回归方法建立了多模式集成的动力 统计客观预报模型(SVM 多模式集成预报),继而选用2012年5—9月(共计153 d)发生在淮河流域及其以南地区的大雨和暴雨开展了回报试验,并将所得预报结果与ECMWF的控制预报和集合平均预报进行了多角度比对评估。结果表明:在中期预报时效(4~7 d),SVM 多模式集成预报方法对2012年5—9月大雨和暴雨的预报效果最优,尤其对暴雨预报准确率明显提高,其优势主要体现在对强降雨中心分布范围和强度的预报更接近实况。
Abstract:This paper establishes a multi mode integrated dynamic statistical objective forecast model (SVM multi model integration forecast) based on the European Centres for Medium Range Weather Forecasts (ECMWF) and the National Centers for Environmental Prediction Center (NCEP) ensemble forecast data and support vector machine regression method, then carries out a forecast test for heavy rain process that occurred in the Huaihe River Basin and its south of China during the period from May to September in 2012, and finally the forecast results are compared with the control forecast and ensemble average forecast of ECMWF. The results show that in the medium term forecasting time scale (4-7 days), the SVM multi model integrated forecast method is the best for forecasting heavy rain compared with the control forecast of the ECMWF and the ensemble average forecast during the period from May to September in 2012. Especially for the accuracy of rainstorm forecasting, it is more effective, and the advantage is that its forecast of the distribution and intensity of heavy rain is closer to the observation.
文章编号:     中图分类号:P456    文献标志码:
基金项目:国家重点基础研究发展计划(973计划)(2012CB417204)、国家科技支撑计划(2015BAC03B02)和国家重点研发计划(2016YFC0402702)共同助资
引用文本:
黄威,牛若芸.基于集合预报和支持向量机的中期强降雨集成预报试验[J].气象,2017,43(9):1110-1116.
HUANG Wei,NIU Ruoyun.The Medium Term Multi Model Integration Forecast Experimentation for Heavy Rain Based on Support Vector Machine[J].METEOROLOGICAL MONTHLY,2017,43(9):1110-1116.