基于LS+AR模型的三种方式预报地球自转参数的比较

Comparison of three modes of LS+AR model for predicting Earth rotation parameters

为研究最小二乘(least squares,LS)模型和自回归(autoregression,AR)模型的组合(LS+AR)方法用于地球自转参数(Earth rotation parameters,ERP)的预报时,不同的预报方式对预报结果的影响,我们采用递推、迭代和间隔这3种预报方式对ERP进行预报。结果表明,这3种方式对日长变化(length of day,LOD)所有跨度预报的精度相当,而递推方式在极移所有跨度的预报上表现出精度优势,间隔方式次之,迭代方式最差。在数据利用率和计算速度方面,递推和迭代方式的数据利用率高,但前者的计算量明显小于后者,而间隔方式的数据利用率低,但计算速度最快。

In order to study the influences of the different modes of combined LS(least square)+AR(autoregression) model on the prediction of ERP(Earth rotation parameters), we adapted three modes, i.e. the recursion mode, the iteration mode and the interval mode, to predict ERP. The results show that the three modes have equivalent prediction accuracies in forecasting the length of day(LOD) for all prediction spans, while in forecasting the polar motion the recursion mode has the highest prediction accuracy, and the interval mode outperforms the iterative mode. In addition, the analyses indicate that the data-utilization rates for the recursion and iterative modes are higher than that for the interval mode, however, the interval mode shows the fastest computation speed compared with the other two modes, and the amount of calculation for recursion mode is less than that for iteration mode.