A polar motion prediction method considering the polar coordinates

Conventional polar motion (PM) parameter modelling and forecasting research is directly based on observation data for PM parameters. For any given PM, its periodic characteristics are basically the same since it is affected by the same excitation source. However, the time-varying patterns of PM in different coordinates or with different parameterizations are generally different. Therefore, it is possible for one parameterization to show better regularity than another. Notably, for the first time, daily polar displacement and the angle of polar displacement with respect to the X axis of the terrestrial frame display better regularity than the observed polar motion value. Therefore, improved accuracy could be expected in modelling and forecasting using the above two parameters. Different approach was tested to verify the effectiveness and the superiority of the proposed method based on multiple sets of PM prediction experiments with the data of IERS 14 C04 series which spanned from 1 January 2018 to 26 September 2020. Compared with typical conventional approach, the maximum prediction accuracy of the proposed method in the PMX direction and the PMY direction is increased by 59.9% and 63.2%, respectively, and the corresponding 1-500-day forecasting accuracies are improved by 25.1% and 36.4%, respectively, on average. (C) 2021 Published by Elsevier B.V. on behalf of COSPAR.