A decade of GPS geodesy in the Australian region : a review of the GDA94 and its performance within a time series analysis of a 10 year data set in ITRF 2000

  • Russell Colin Tiesler

Student thesis: Master's Thesis


The University of Canberra (UC) has been involved in GPS processing since the late 1980s. This processing commenced with the GOTEX 1988 campaign and progressed through a series of project specific regional campaigns to the current daily processing of a distributed set of continuously operating sites for the determination of precise GPS station positions for user applications. Most of these earlier campaigns covered only short periods of time, ranging from a few weeks to multiple occupations of a few days to a time over one to two years. With software developments, these multiple occupations were able to be combined to produce results from which crustal motion velocities could be extracted. This first became feasible with the processing of the Australian National Network (ANN),which yielded realistic tectonic velocities from two occupations (1992 and 1993) of sites 12 months apart. Subsequently, this was successfully extended by a further 12 months, with re-occupation of certain sites for a third time in 1994. Analysis of the results indicated that the accuracy of determining the earth signals improved as the time span from first to last observation was increased. The same was true also for the determination of the position of global references sites. However, by current standards the results achieved were poor. Consequently, the process was extended to combine the results of subsequent campaigns with the original ANN data set. From 1995 to 1999,campaigns were conducted across Australia, covering many State and tide gauge sites included in the original ANN solution. These provided additional multiple occupations to improve the determinations for both position and velocity. UC has maintained a data set of the global IGS sites, commencing with the IGS pilot campaign of 1992. Daily data sets for those global sites, which contained days common to the regional campaigns, were processed to produce our own independent global orbit and reference frame connection. The motivation for doing so was fourfold. .Firstly, to see if historic data could be reprocessed using current modern software and thus be able to be incorporated in this and other analysts research programs. .Secondly, to compare the results of the reprocessing of the original data set using modern software with the original ANN solution and then validate both the solutions. .Thirdly, to extend the timespan of observations processed to include more recent campaigns on as many original sites as possible. This to achieve a stronger solution upon which to base the determination of an Australian tectonic plate velocity model and provide quality assurance on the solution comparisons with re-observed sites. .Fourthly, to develop a set of transformation parameters between current coordinate systems and the GDA94 system so as to be able to incorporate new results into the previous system. The final selection of regional and global sessions, spanning from mid 1992 to late 2002,contained almost 1000 individual daily solutions. From this 10 year data span a well determined rigid plate tectonic motion model was produced for Australia. This site velocity model was needed to develop a transformation between the thesis solution in ITRF00 an the GDA94 solution in ITRF92. The significant advantage of the plate velocity model is that all Australian sites can now have computed a realistic velocity, rather than being given a value which has been interpolated between sites whose velocities had been determined over a one or two year span. This plate velocity model is compared with the current tectonic motion NNR-NUVEL-1A model and other recently published models. To perform the comparison between the thesis solution in ITRF00 and the GDA solution in ITRF92 a transformation was developed between the two reference systems. This set of transformation parameters, in conjunction with the plate velocity model developed, enables site solutions at any epoch in the current ITRF00 to be converted onto the GDA94,and vice versa, with a simple, non-varying seven parameter transformation. The comparisons between the solutions are analysed for both horizontal position and height consistency. There were 77 sites whose differences were compared. The horizontal consistency was within estimated precisions for 75 of the 77 sites. However, the vertical comparisons revealed many of the single epoch sites, especially in 1992,have inconsistent results between the two solutions. The heights from this thesis for some West Australian sites were compared with analysis done by DOLA and the height recoveries are very similar, indicating a weakness in the GDA94 solution for some of the single epoch sites. Some of these differences have been resolved but others are still under investigation. This thesis describes the reprocessing of the original ANN data set, the addition of later data sets, the results obtained, and the validation comparisons of the old and new solutions. As well as the plate velocity model, transformation is provided which enables the user to compute between the GDA94 system, and any epoch result in ITRF00. Recommendations are made as to which sites need additional work. This includes sites which only need further analysis or investigation and those which require further observations to achieve a result which will have acceptable accuracy and reliability.
Date of Award2005
Original languageEnglish
Awarding Institution
  • University of Canberra
SupervisorPeter Morgan (Supervisor) & John Rayner (Supervisor)

Cite this