HR: 08:30h
AN: G11A-03    [Abstracts]
TI: On the Accuracy of IGS Orbits
AU: * Griffiths, J
EM: jake.griffiths@noaa.gov
AU: Ray, J
EM: jim.ray@noaa.gov
AB: In order to explore the reliability of IGS internal orbit accuracy estimates, we have compared the geocentric satellite positions at the midnight epoch between consecutive days for the period since November 5, 2006, when the IGS changed its method of antenna calibration. For each pair of orbits, day "A" has been fitted to the extended CODE orbit model (three position and three velocity parameters plus nine nuisance solar radiation parameters), using the IGS05 Final orbits as psuedo-observations, and extrapolated to epoch 24:00 to compare with the 00:00 epoch from the IGS05 Final orbits of day "B". This yields a time series of orbit repeatability measures, analogous to the classical geodetic test for position determinations. To assess the error introduced by the fitting and extrapolation process, the same procedure has been applied to several days dropping the 23:45 epoch, fitting up to 23:30, extrapolating to 23:45, and comparing with reported positions for 23:45. The test differences range between 0 and 10 mm (mean = 3 mm) per geocentric component with 3D differences of 3 to 10 mm (mean = 6 mm). So, the effect of the orbit fitting-extrapolation process nearly always adds insignificant noise to the day- boundary orbit comparisons. If we compare our average 1D position differences to the official IGS accuracy codes (derived from the internal agreement among combined orbit solutions), root-sum-squared for each pair of days, the actual discontinuities are not well correlated with the expected performance values. If instead the IGS RMS values from the Final combination long-arc analyses (which also use the extended CODE model) are taken as the measure of IGS accuracy, the actual orbit discontinuties are much better represented. This is despite the fact that our day- boundary offsets apply to a single epoch each day and the long-arc analyses consider variations over a day (compared to the satellite dynamics determined over the full week). Our method is not well suited to probe the extent to which systematic effects dominate over random orbit errors, as indicated by satellite laser ranging residuals, but eclipsing satellites often display the most problematic behavior. Users must always be skeptical of GPS-based formal error estimates for all quantities, not just orbits, and should be aware that non- Gaussian errors are common. Nevertheless, the IGS SP3 accuracy codes appear to misrepresent the actual orbit errors in many cases. A better metric would probably be the IGS long-arc RMS values rather than the RMS agreement among solution submissions.
DE: 1240 Satellite geodesy: results (6929, 7215, 7230, 7240)
DE: 1241 Satellite geodesy: technical issues (6994, 7969)
DE: 1244 Standards and absolute measurements
DE: 1299 General or miscellaneous (1709)
SC: Geodesy [G]
MN: 2007 Fall Meeting