利用ARMA模型短期预报北极地区电离层TEC

发布时间：2018-02-13 23:18

本文关键词： ARMA 北极地区电离层TEC 线性最小方差预报　出处：《测绘工程》2016年03期 　论文类型：期刊论文

【摘要】：北极地区电离层结构分布较为特殊,存在梯度变化。利用时间序列分析中的自回归移动平均模型(Autoregressive Moving Average,ARMA)对欧洲定轨中心(CODE)发布的北纬67.5°~87.5°以及利用反距离加权插值法得到的90°的格网数据逐点进行建模,分别利用7d、10d、20d、30d、40d、50d的电离层TEC值为样本数据采用线性最小方差法进行预报分析。结果表明:90%以上的预报绝对误差小于3TECU,预报精度随TEC样本序列长度的增加而提高,但样本序列增加到一定值后,相对精度提高不大;相同样本数据的预报精度随预报时间长度的增加而降低,起初不是很明显,超过20d后精度降低明显且波动幅度较大。尽管北极地区存在梯度变化,ARMA模型在北极地区具有较高的预报精度,是一种比较理想的预报方法。
[Abstract]:The distribution of ionospheric structure in the Arctic is relatively special. The autoregressive Moving average model is used to model the grid data of 67.5 掳(87.5 掳) N and 90 掳(90 掳) published by the European orbit determination Center (CODED) in time series analysis. Using the ionospheric TEC values of 7d ~ 10d ~ 20d ~ 30d ~ 40d ~ 50d as sample data respectively, the linear minimum variance method is used to forecast and analyze. The results show that the absolute error of prediction is less than 3TECU for more than 90% of the samples, and the prediction accuracy increases with the increase of TEC sample length. However, the relative precision of the same sample data decreased with the increase of the prediction time, but at first it was not obvious that the relative precision of the sample sequence increased to a certain value, and the prediction accuracy of the same sample data decreased with the increase of the prediction time. After more than 20 days, the precision decreases obviously and the fluctuation range is large, although the ARMA model with gradient variation in the Arctic region has higher prediction accuracy in the Arctic region and is an ideal prediction method.
【作者单位】：信息工程大学;中国科学院测量与地球物理研究所;北京卫星导航中心;
【基金】：国家自然科学基金资助项目(U1431115;41374042) 大地测量与地球动力学国家实验室开放式基金资助项目(SKLGED 2014-3-5-E)
【分类号】：P228.4

【参考文献】