Multi-parameter Polar Motion Prediction Based on Effective Angular Momentum Function

Abstract

The change of polar motion is closely related to a variety of excitations. These excitations include atmospheric surface pressure and atmospheric wind, seabed pressure and ocean currents, land water distribution, and sea level changes caused by climate warming, and can be estimated by the effective angular momentum function. In the polar motion prediction, the effective angular momentum function is included through the Liouville equation, and combined the least square and autoregressive method for data fitting and extrapolation. At the same time, more options are set for the adjustable parameters of the autoregressive model. In different phases of polar motion prediction, the predictions of each components are matched with optimized parameters, which effectively improves the prediction accuracy of polar motion. In 441 polar motion prediction experiments from 1 to 90 days, the short and medium term predictions were improved more obviously. In the 1–6 day and 7–30 day of the polar motion X (PMX) prediction results, there were 56.9% and 53.5% respectively better than the IERS (International Earth Rotation Service) prediction; in the 1–6 day and 7–30 day of the polar motion Y (PMY) prediction results, 66.5% and 59.7% are better than the IERS prediction, respectively. In General, the performance of PMY prediction is better than that of PMX. Taking IERS EOP (Earth Orientation Parameters) C04 as a reference, the MAE (Mean Absolute Error) of the polar motion X prediction on the 1st-day and 5th-day is improved by 2.6% and 33.0%, respectively compared with the IERS prediction. Compared with the IERS prediction, the MAE of Y prediction on the 1st-day and 5th-day is improved by 20.8% and 49.0%, respectively.