Evaluation of earth gravity field models used for precise satellite orbit determination through applications of satellite laser ranging data

Abstract

One of the applications of the Satellite Laser Ranging (SLR) technique is the derivation of gravity field models; these models have various geophysical and geodynamical applications. Gravity field modelling has reached a new era where the latest satellite missions (CHAMP, GRACE and GOCE) are thought to provide significant improvement of global gravity field information in terms of quality and spatial resolution. In particular, the recent satellite missions carry on-board Global Navigation Satellite System (GNSS) receivers, accelerometers, K/Kaband microwave system (e.g. in GRACE) and gradiometers (e.g. in GOCE) allowing measurements of gravity field with unprecedented accuracy in contrast to the unsteady and fragmented orbit tracking by unevenly distributed SLR ground stations. Numerous gravity field models have been derived based on the newly available data sets by various research groups globally. Due to the availability of high quality SLR and satellite data, some of the older gravity field models are being updated as new models with higher degree and order are developed. Notwithstanding the significant progress in gravity field modelling, research focusing on assessing the accuracy and precision of the existing gravity field models has largely remained insufficient. The difference between the observed and computed satellite orbit (which is often expressed as the O-C range residuals) is used as a parameter for Precise Orbit Determination (POD) of satellites. Furthermore, O-C range residuals computed during SLR analysis are used as proxy parameters for evaluating the accuracy of gravity field models. The work presented in this thesis firstly reviewed and evaluated the accuracy of gravity field models released between 1990 and 2008. The accuracy of the gravity field models was examined by analysing the O-C residuals computed from LAGEOS 1 and 2 data analysis based on a set of twelve gravity field models. The results demonstrated that in general, there has been an improvement in the accuracy of gravity field models released between 1990 and 2008 by a factor of 2 based on improvements in the O-C residuals. Additionally, the influence of SLR tide parameterization (the IERS 2010 solid Earth and pole tide models) on the O-C residuals across five gravity field models has been assessed and results illustrate that the solid Earth and pole tides parameterization influence on the O-C residuals is dependent on the type of gravity field model. In order to ascertain the significance of mean differences in the Standard Deviations (SD) of O-C residuals based on the tide parameterization options, the student’s t-test was used. Results suggest that in general the O-C residuals derived from SLR LAGEOS 1 data have insignificant mean SD differences across the tide parameterizations. On the other hand analysis of SLR observations of LAGEOS 2 resulted in statistically significant mean SD differences in the O-C based on EIGEN-CG03C, EGM2008 and AIUB-GRACE01S gravity field models. The J2 coefficient forms part of the SLR Data Analysis Software (SDAS) package output products and was investigated in this thesis due to its role in understanding mass-redistribution within the Earth system (i.e. the equatorial bulge due to centrifugal force and rotation). In particular, the J 2 coefficient computed from SLR analysis of LAGEOS 1 and 2 data sets and based on the four selected gravity field models were compared with a priori J2 coefficients from the four models and those published in the literature. The results indicated that the J2 coefficients computed from the SDAS package were in agreement with the published coefficients. For geophysical applications, the relationship between the J2 parameter and LOD and AAM was investigated by use of data adaptive analysis methodology (the empirical mode decomposition). The results demonstrated that some degree of synchronization exists between the signal components of J2 and LOD and J2 and AAM.