6 Positions and motions of the stations

In the solutions S2000 and earlier we set the positions and motions of the stations to constant values deduced from the ITRF94. Increasing the interval of observations, we noticed slight trends of quantities related to the reference frame, in particular in the obliquity of the ecliptic $\epsilon$. Furthermore, with other sets of positions and velocities for the stations, deduced from ITRF96 or ITRF2000, we observed significant changes in the obliquity. Hence, we modified our list of fitted parameters in the analysis, adding the positions and velocities of the stations as new unknowns. As initial values, we adopted the positions deduced from ITRF94 and the velocities of ITRF2000.
The main results are the following:
- The trend in $\epsilon$, when the time interval of observations increases, is significantly reduced.
- The residuals for Haleakala are significantly reduced from 8 cm in S2000 to 6 cm in S2001 (see Table 1);
- The corrections to the positions of the stations are in good agreement with the positions given in ITRF2000.
The last point calls for some comments. The corrections to the positions determined either with a large set of parameters (see Sect. 3) or independently are very close, which supports the relevance of our present determinations. The corrections obtained in Sol. 1 (MCEP) and Sol. 2 (ICRS), are also very close.

In Table 5 we present our fit in Sol. 2 (ICRS) for the positions of the 2 operating stations: CERGA (Yag) and McDONALD (MLRS2). The "references'' are deduced from ITRF publications (Boucher et al. 1996, 1998); in the case of MLRS2 we used corrections of eccentricity to the nearby SLR station (Ries 1999). The "corrections'' S2001 Sol. 2 (ICRS) must be applied to ITRF94 positions (Epoch 1993.0) and to ITRF2000 velocities, and have been fit independently. For the "comparisons'', our corrected positions at the epoch 1997.0 were computed in two different ways: using the ITRF2000 velocities (positions S2001^*) or using the corrected velocities (positions S2001^**). The positions S2001^* show a better agreement with ITRF2000 than with ITRF96.

For the CERGA the results S2001^** are even closer. Nevertheless, the correction to $\dot{Z}$ (0.0036 m/year) seems to be too large compared to the reference value (0.0101 m/year in ITRF2000). New determinations of the positions of LLR and SLR stations at CERGA have been obtained recently by (Nicolas 2000), using the observations of the satellites LAGEOS 1 and 2 from 1997 to 1999. They show local and seasonal displacements of the LLR station, reaching a few centimeters with a probable tendency of a few mm/year. These small displacements are ignored in ITRF2000, but the recent LLR observations are sensitive to this level of accuracy. The order of magnitude of such displacements is in accordance with our corrections.