EEMD-Holt-Winters的电离层TEC预报模型  

A prediction model of ionospheric TEC based on EEMD and Holt-Winters

作　　者：周强波 ZHOU Qiangbo(Changsha Uranium Geology Research Institute, CNNC, Changsha 410007, China)

机构地区：[1]核工业二三〇研究所,湖南长沙410007

出　　处：《测绘工程》2021年第3期15-20,共6页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(41761089);中国核工业地质局铀矿地质项目(201917;201930-1;202028-1;202021-1)。

摘　　要：利用IGS中心提供的全球电离层TEC数据,建立集合经验模态分解与Holt-Winters组合的预报模型。根据地磁活动情况,选取地磁平静时段和地磁活跃时段的低、中纬度TEC序列,分别采用EEMD-Holt-Winters组合模型和单一Holt-Winters模型进行建模预报。实验结果表明,地磁平静期文中模型的相对精度最优可达91.71%,均方根误差最优可达1.54 TECu,地磁活跃期文中模型的相对精度最优可达86.83%,均方根误差最优可达2.18 TECu,预报结果较单一模型有显著提高。This paper uses the global ionospheric TEC data provided by IGS center to establish a forecasting model combining empirical mode decomposition and Holt-Winters. According to the geomagnetic activity, the EMD-Holt-Winters combination model and the single Holt-Winters model are used to model and forecast the low and mid-latitude TEC sequences in the periods of geomagnetic calm and activity respectively. The experimental result shows that the relative accuracy of the model in this paper can reach 91.71%, the root mean square error can reach 1.54 TECu, the relative accuracy of the model can reach 86.83%, and the root mean square error can reach 2.18 TECu during the period of geomagnetic activity. Compared with the single model, the forecast result is significantly improved.

关 键 词：电离层 集合经验模态分解 Holt-Winters 总电子含量 

分 类 号：P228[天文地球—大地测量学与测量工程]