SLR analysis

For the 11-day period of the CHAMP orbit campaign the available orbit solutions have been used as fixed input orbit in a process to compute the SLR tracking data residuals. The software used at ESOC for this analysis is based on the system that is also used operationally for satellites like ERS and is completely independent of the POD systems that were used for any of the orbit solutions, including the ESA solution. These SLR residuals should therefore form a useful independent check of the orbit quality.

The three-dimensional orbit error at the time of the SLR observation will hardly ever be oriented along the line of sight between tracking station and satellite. This means that the SLR residuals form a lower limit to the 3D orbit error, assuming that the SLR residual is predominantly due to orbit error.

The following data editing has been applied:

  1. The data from stations 1873 and 7237 was excluded because of large resiudals in all solutions
  2. Two bad passes from other stations were identified and rejected
  3. The residuals from the three most precise orbit solutions (according to a first run) were used to compute a common sigma value 7 cm 1-way range), and a 1.5 sigma rejection window was applied to reject individual bad points. Only points that were excluded by all three solutions were rejected as bad points.
  4. The  remaining data set of 1686 SLR observations was used for all input orbit solutions. Any further rejections occurred due to gaps in the orbit files.
The temporal and spatial distribution of the SLR data is of course highly irregular, typically concentrated in a few overlapping passes over Europe, complemented with incidental passes over USA or Australia.

Table 1 below presents the overall SLR residual statistics for each included orbit solution. In total 1908 normal points were available for the campaign period.

The SLR residuals have been plotted in a single figure per contributed orbit solution. Please click on the icons below to obtain the larger Figures. Due to the 11-day time scale, the laser tracking passes are compressed into near-vertical lines in these Figures, which conveniently translates into a mean value with an error bar for each pass. The small images here already show that the SLR residuals increase from the currently most precise solution onwards.

Despite of the careful data editing, there are still some notable differences in the global mean values of the SLR residuals, especially for the solutions of TUM and GRGS that are among the best in terms of RMS.

The results for ASI do not seem very realistic. The obtained SLR residuals of 35 cm are highly inconsistent with the orbit comparison results, which suggest an upper limit to the orbit error of 16 cm. This problem is still under investigation. The only way in which such an inconsistency can be explained from theory would be that the ASI solution is dependent on all other solutions, which seems unlikely. The SLR processing for the ASI solution has been carefully verified and was even repeated manually, but no anomalies were found.
 

centre
nr obs
RMS
MEAN
SIGMA
GFZ
1686
4.81
0.75
4.76
TUM
1686
5.32
2.46
4.72
CSR
1686
4.43
0.04
4.43
NCL
1686
7.44
-0.59
7.42
GRGS
1686
6.80
3.35
5.92
DEOS
1686
8.93
-1.90
8.73
ASI
1686
35.45
0.99
35.45
JPL
1686
10.33
2.46
10.03
AIUB
1686
13.56
2.55
13.33
CNES
1686
13.58
0.31
13.58
ESA
1686
16.83
2.64
16.63
UCAR
1686
17.35
4.68
16.71
UNB
1403
27.37
0.81
27.37

Table 1: SLR statistics per contributed solution


 


CSR GFZ GRGS TUM
JPL DEOS CNES NCL
ESA UNB AIUB UCAR
ASI
Figure 2: SLR residuals per contributed orbit solution