Updated:  28 June 2005

GA Global VLBI Solution 2004a

About this solution

This global homogeneous solution has been computed using the new facilities of OCCAM (version 6.1). The new version treats unstable radiosources as local or 'arc' parameters through the full set of data. 3081 daily sessions have been used to make up this GA solution (2004a) using a global set of VLBI data from 12-Apr-1980 till 27-May-2004, and include 2.990.979 observational delays from 672 radiosources (571 global and 101 'arc'), observed by 56 VLBI stations. The weighted root-mean-square error of the solution is 0.62 cm (~21 ps). The parameterisation includes three groups of parameters:


[back to top]

Solution strategy

The aus2004a solution strategy follows the strategy for computation of the ICRF-Ext.2 [Fey et al., 2004]. Coordinates of 207 of the 212 defining sources [Ma et al., 1998] were treated as global parameters and with weighted No Net Rotation (NNR) constraints. 107 sources were treated as local and their positions were estimated for each VLBI session. Other sources were treated as global parameters without the NNR-constraints. Station coordinates were also estimated as local parameters using NNR and No Net Translation (NNT) constraints. The long-term coordinate time series of the 56 stations have been established to estimate the corresponding velocities for each station. Due to a limited number of observations, the velocities have been estimated for 52 stations only.

Velocities for the Gilcreek VLBI site was estimated using the points before 3 November 2002, when a strong earthquake displaced the site dramatically. Tectonic motion for Gilcreek after the Denali earthquake is modelled using an exponential function [Titov and Tregoning, 2004]. There are two solutions for Gilcreek, designated as 'Gilcree0' and 'Gilcree1' for periods before and after the earthquake, correspondingly. The adjustment has been completed using the least squares collocation method, which considers the EOP, clock offsets, wet troposphere delays, troposphere gradients as stochastic parameters with a priori covariance functions. Following this approach the matrix of a priori covariances becomes non-diagonal. The method takes into account the mutual correlations between delays within each of the 24-hour sessions, as well as estimating the stochastic parameters for every observational epoch, i.e. time resolution in 5-10 minutes.


[back to top]

Models used

The IMF mapping function [Niell, 2001] has been used for modelling the troposphere and its gradients. The MBH nutation model [Matthews et al, 2002] has been applied for calculations with the exception of the Free Core Nutation (FCN) effect, which is included to the final nutation offset time series.


[back to top]

References

Fey, A., et al., The second extension of the International Celestial Reference Frame: ICRF-Ext.1. AJ, 127, 3587-3608, 2004.

Ma, C., et al., The International Celestial Reference Frame as realized by Very Long Baseline Interferometry, AJ, 116, 516-546, 1998.

Matthews, P.M., Herring, T.A., and Buffett, B.A., J.Geoph. Res., 107, 10.1029/2001 JB000390, 2002.

Niell, A., Preliminary evaluation of atmospheric mapping functions based on numerical weather models, Phys. Chem. Earth, 26, 475-480, 2001.

Titov and Tregoning, Post-Seismic Motion of Gilcreek Geodetic Sites Following the November, 2002 Denali Earthquake, in International VLBI Service for Geodesy and Astrometry 2004 General Meeting Proceedings, edited by Nancy R. Vandenberg and Karen D. Baver, NASA/CP-2004-212255, 2004.

For more information contact: geodesy@ga.gov.au