Study On Polar Motion Prediction Method With Integration Of Effective Angular Momentum Function

Pole motion is an important part of the earth orientation parameters and one of the key parameters for the conversion between the international celestial reference frame and the international terrestrial reference frame.In the fields of deep space exploration and satellite navigation,high-precision prediction of pole motion has important scientific significance and engineering application value.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.In recent years,the excitation function has been applied to the pole motion prediction.At present,there is little research on using excitation function to improve pole motion prediction in domestic,and there is still a certain gap between the prediction level and the highest level in the world.In order to further improve the accuracy of independent prediction of pole motion in China,based on the prediction method of Dr.Robert dill of GFZ in 2019,this paper integrates the effective angular momentum function into the prediction of pole motion.At the same time,it tries to make beneficial exploration and improvement on the extrapolation method of autoregressive error,which significantly improves the prediction accuracy of pole motion from 1 to 90 day.Firstly,this paper summarizes the status of pole motion prediction at home and abroad,and reviews some important pole shift prediction methods,pole motion measurement methods,and the change excitation mechanism of pole motion on different time scales.Then,for the mathematical description of the geophysical and dynamic causes of excitation function and pole motion change,and the Liouville equation,a more detailed formula derivation and explanation are given.Starting with the Euler equation of rigid body rotation,the mathematical expression of the influence of angular momentum caused by relative motion and material redistribution on the earth rotation under the conditions of rigid body earth and elastic earth is derived.The Liouville equation and Chandler wobble complex frequency formula are further deduced,and the transformation of Liouville equation from complex form to discrete form is completed,which makes a mathematical reserve for the coding of the final prediction program.In the fourth chapter,the AAM,OAM,HAM and SLAM data used in this paper are introduced and analyzed in detail.On this basis,the algorithm and code of pole motion prediction based on effective angular momentum function and LS+AR are completed,and the method of medium and short term(within 90 days)pole motion prediction is improved and optimized.The main innovations of this paper are:(1)The data are fitted and extrapolated by using the combination of piecewise least squares and autoregressive model.And the interval number 7)(62)is set in the autoregressive model,which expands the selection space of parameters,so that the prediction of different pole motion components can match better parameter groups in different pole motion prediction stages.In this paper,441 times of pole motion prediction from 1 to 90 days are carried out and compared with IERS EOP C04 and IERS Bulletin A respectively.The results show that 56.9% and 53.5% of the prediction results of pole motion X from 1 to 6 days and 7 to 30 days are better than those of IERS respectively,and 66.5% and 59.7% of the prediction results of pole motion Y from 1 to6 days and 7 to 30 days are better than those of IERS respectively;Compared with IERS,the MAE of pole motion X forecast on the first and fifth days decreased by 2.6% and33.0% respectively,and the MAE of pole motion Y forecast on the first and fifth days decreased by 20.8% and 49.0% respectively.On the whole,the prediction method proposed in this paper is more suitable for short-term pole motion prediction,and the prediction improvement of pole motion Y is greater than that of pole motion X.(2)This paper makes a further exploration on the evaluation method of pole motion prediction.In addition to the general MAE,we also try to compare and evaluate different prediction methods by using RMSE,Med AE and Max AE.The results show that under all the above evaluation indexes,the method in this paper can get the same results as IERS prediction.In particular,for the prediction of PMX,the Max AE within30 days is significantly lower than the prediction results of IERS.The Max AE is a concept similar to the lowest score and protection level,indicating how big the error is in the worst prediction.The python pole motion prediction program developed based on the research results of this paper has completed the deployment and routine operation of Linux server environment,and has performed well in the second EOP PCC competition held by IERS recently.