A Method for Improving the Short-Term Prediction Model for ERP Based on Long-Term Observations

Abstract

Earth Rotation Parameter (ERP) is one of the most important parameters in the area of positioning and navigation, autonomous orbit determination and Earth reference framework. However, due to the restriction of timeliness of data processing, the predicted ERPs are taken into the relevant applications to meet the requirements of real-time or near real-time users. Therefore, given the disadvantages of short-term prediction models for ERPs, such as model mismatch, over parametrization and divergence with time, this study proposed a method for improving the short-term prediction model for ERP based on long-term observations. Firstly, the optimal length of observations was analyzed for the Least Square (LS) prediction model based on the Akaike Information Criterion (AIC). It is found that one year of ERP observations is the optimal data sets to establish the prediction model. Then, two constraints models based on LS, called (Constraint LS) CLS and (Enhanced CLS) ECLS, were discussed to decrease the errors in prediction models, which takes the correlation factors and residuals of prediction model into consideration. The results indicated that the prediction errors of ERP can be significantly decreased for short-term prediction, which improved the accuracy of ERP prediction with 50% for polar motions (PM), and 20% for UT1-UTC. Moreover, considering the divergence of predicted ERP along with the increasing of time and separate prediction of each parameters in ERP of Auto Regressive model (AR), a Multivariable Regression model (MAR) was introduced to correct the residuals of predicted ERP, which combined the PM, and UT1-UTC with LOD into prediction models. And the accuracy of predicted ERP can be improved at least 20% compared with AR model. Finally, according to data experiments, the improved short-term prediction model was analyzed based on different observations. It is suggested that our method can improve the short-term prediction in cases that long-term ERP observations are available.