基于GM(1,1)和AR组合模型的UT1-UTC超短期预报  

Ultra short-term prediction of UT1-UTC based on the combination of GM(1,1) and AR models

作　　者：雷雨[1,2] LEI Yu(National Time Service Centre, Chinese Academy of Sciences, Xi＇an 710600, China Key Laboratory of Time and Frequency Primary Standards, National Time Service Center, Chinese Academy of Sciences, Xi＇an 710600, China)

机构地区：[1]中国科学院国家授时中心,西安710600 [2]中国科学院时间频率基准重点实验室,西安710600

出　　处：《时间频率学报》2017年第2期65-72,共8页Journal of Time and Frequency

基　　金：中国科学院"西部之光"人才培养计划"联合学者"资助项目(中科院人字[2014]91号)

摘　　要：针对UT1-UTC序列呈现趋势性和随机性变化的特点,提出一种基于GM(1,1)和自回归(autoregressive,AR)的组合预报模型。该模型首先采用GM(1,1)模型预报UT1-UTC序列中的趋势项,然后利用AR模型对GM(1,1)模型残差序列进行建模和预报,最后将GM(1,1)模型和AR模型的预报结果相加获得UT1-UTC预报值。将组合模型的超短期(1~10 d)预报结果与地球定向参数预报比较竞赛(Earth orientation parameters prediction comparison campaign,EOP PCC)结果进行对比,结果表明:组合模型1~3 d的预报效果优于目前国际上普遍采用的预报方法,而4~10 d的预报精度则不及顾及大气角动量(atmospheric angular momentum,AAM)的UT1-UTC预报方法,但仍优于参与EOP PCC的其他方法的预报结果。According to the characteristics of tren d-oriented and random variations in U T1-UTC, a n ew prediction model based on the combination of GM（1, 1） and AR models is proposed in this paper. The GM（1, 1） model is firstly used to fit and predict the trend term of UT1-UTC time-series, and then the residuals of GM（1, 1） model are modeled and forecasted by an AR model. Finally, the predictions of the trend and residual terms are added together to derive the UT1-UTC forecast value. In order to verify the efficiency of the proposed model, the results of the ultra short-term（1~10 days） predictions are compared and analyzed with those obtained by Earth orientation parameters prediction comparison campaign（EOP PCC）. The results show that the prediction precision of 1 to 3 days of this model is better than that of the methods and techniques used nowadays, the error of predictions at the horizon of 4~6 days is higher than that by those prediction approaches taking into account atmospheric angular momentum（AAM） but is lower than that by the other methods and techniques.

关 键 词：地球定向参数 UT1-UTC 预报 GM(1 1) 自回归模型 

分 类 号：P126[天文地球—天体测量] P228[天文地球—天文学]