Improved estimations of low-degree coefficients using GPS displacements with reduced non-loading errors

• Published in: Geophysical journal international 2018-02, Vol.212 (2), p.1274-1287
• Main Authors: Wei, Na, Shi, Chuang, Wang, Guangxing, Liu, Jingnan
• Format: Article
• Language: English
• ISSN: 0956-540X; 1365-246X
• DOI: 10.1093/gji/ggx357

Summary We investigate and try to reduce the impacts on low-degree estimates of non-loading errors, that is, aliasing of unmodeled loading and Global Positioning System (GPS) draconitic year errors, to improve the sensitivity of GPS observations to the loading mass. Three GPS data sets, ITRF2008−GPS residuals, ITRF2014−GPS residuals and Jet Propulsion Laboratory (JPL)’s residuals, are used and compared in this paper. Results show that the aliasing signals in GPS displacements is an important error source, especially for inferring geocentre motion. The two International Terrestrial Reference Frame (ITRF)–GPS residuals generated in a two-step combination based on Helmert transformation show more complex aliasing errors than JPL’s residuals produced in precise point positions mode. The seasonal variations of geocentre motion derived from JPL thus perform the best among all three solutions, while the higher degree coefficients from the two ITRF–GPS solutions do better. Compared with ITRF2008−GPS residuals, the aliasing errors are indeed reduced, and geocentre motion/${\rm{\Delta }}T_{20}^C$ (degree-2 zonal coefficients in terms of surface mass density) are also much improved for ITRF2014−GPS residuals produced with a six-parameter transformation without scale parameter. Additional translation parameters should be included into ITRF2008−GPS residuals, or else ${\rm{\Delta }}T_{20}^C$ cannot be correctly obtained. The draconitic errors pose another obstacle to accurately studying the seasonal variations of surface loading using GPS data. The draconitic harmonics (first, second and third) are well extracted from ITRF2014-derived ${\rm{\Delta }}T_{20}^C$ and ${\rm{\Delta }}T_{21}^S$ (degree-2 and order-1 sine coefficients), even if the time span is not long enough to independently separate the seasonal variations and draconitic harmonics. These errors account for an increase of about 10 per cent in the annual amplitude of ITRF2014-derived ${\rm{\Delta }}T_{20}^C$ and ${\rm{\Delta }}T_{21}^S$. Removing the found draconitic errors brings ${\rm{\Delta }}T_{20}^C$ and ${\rm{\Delta }}T_{21}^S$ from GPS and Satellite Laser Ranging into closer agreement.