基于小波分析的客观预报方法在智能网格预报中的应用研究
投稿时间:2020-01-02  修订日期:2020-01-30  点此下载全文
引用本文:
摘要点击次数: 26
全文下载次数: 0
作者单位E-mail
刘新伟 兰州中心气象台 Ljhx6@163.com 
基金项目:国家重点研发计划项目(2017YFC1502002);中国气象局预报员专项项目(CMAYBY2019-122)
中文摘要:基于2017-2018年中国气象局高分辨率数值预报产品、甘肃实时城镇预报产品数据和国家级地面观测站数据,利用小波分析方法,通过滑动训练、最优融合等技术,研发出甘肃省智能网格预报业务城镇站点气温客观订正产品,以2017年数据作为训练样本、2018年数据作为测试样本进行检验分析,结果表明:(1)不同气温预报产品在72h内的预报差异不大,误差均在1.5-2.0℃;(2)不同气温预报产品在甘肃最高、最低气温的预报方面存在差异,其中最低气温的预报与观测较为一致,而最高气温的预报则与观测差异较大;(3)空间误差检验表明,利用最优融合技术,甘南、陇南等系统误差明显地区的预报结果要好于模式客观预报的情况,而预报员制作的城镇报优势则在气温局地性强或者模式客观预报能力差的区域得到体现;(4)基于小波分析的最优融合预报方法生成的产品预报能力已经与现有预报员制作城镇报产品基本持平,初步具备了客观预报替代主观预报的能力。
中文关键词:智能网格,小波分析,气温预报
 
Research on the application of objective prediction method based on wavelet analysis in intelligent grid prediction
Abstract:Based on 2017-2018 high-resolution numerical forecast products from China Meteorological Administration, Gansu real-time town fore-cast product data and national surface observation data,developing the town station temperature objective correction product of Gansu intelligent grid forecast by using wavelet analysis, sliding training and optimal fusion method. Our test analysis used data from 2018 as test samples while used data from 2018 as training samples, the results show that:(1) There is little difference in the prediction of different temperature forecast products within 72h, and the error is 1.5-2.0℃; (2) Different temperature forecast products have differences in the forecast of the highest and lowest temperature in Gansu, the forecast of minimum temperature is consistent with the observation,while the forecast of maximum temperature is quite different from the observation; (3) The spatial error test shows that forecast result with Optimal fusion technology is better than the objective forecast in Gannan, Longnan and other towns with significant systematic errors, the town forecast advantage produced by the forecasters is reflected in the region with strong local temperature or poor objective forecast ability of the model; (4) The product forecast capability generated by the optimal fusion forecast method based on wavelet analysis is basically equal to the existing subjective forecast made by forecasters, and it also has the ability to replace the subjective forecast with objective forecast.
keywords:Intelligent grid,wavelet analysis,temperature forecast
  查看/发表评论  下载PDF阅读器