Terrestrial Reference Frame (TRF) Accuracy Evaluation after Unlabeled Jumps Detection and Non-linear Signals Extraction
Abstract
The accuracy of Terrestrial Reference Frame (TRF) is required higher and higher. But there are still some unlabeled jumps and non-linear signals in the TRF residual data. They lead to big residuals and wrong coordinate and velocity estimations. Therefor we need detect these unlabeled jumps and extract these non-linear signals from TRF residual data. We adopt sequential interrupt algorithm which is based on the t-test theory (Rodionov 2004) to check if the interruption occurs at some epoch. After that we also adopt epoch difference method and the human eye method in case false alarm and leak detection. After these jump processing we adopt Singular Spectrum Analysis (SSA) method to model these non-linear signals. This method fitted the time-varying amplitude period signals very well. After these processing we reprocess the data and get the new TRF and EOP. We evaluate the accuracy of TRF and EOP. And find the coordinate residuals of GNSS/VLBI/SLR/DORIS sites become smaller and the accuracy of TRF is improved.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.G23C0769W
- Keywords:
-
- 1229 Reference systems;
- GEODESY AND GRAVITY;
- 1295 Integrations of techniques;
- GEODESY AND GRAVITY;
- 1299 General or miscellaneous;
- GEODESY AND GRAVITY