JASON SLR

IGS LEO JASON-1 SLR analysis

*IGS LEO Homepage*	*Associate Analysis Centers*	*LEO Orbit Campaigns*	*SP3 and RINEX acronyms*	*Links*
		*CHAMP campaign* *JASON-1 campaign*
	*Orbit comparisons*	*Tracking data analysis*	*Precision estimates*
		*CHAMP SLR analysis* *JASON SLR analysis* *JASON DORIS analysis*

The available JASON-1 orbit solutions have been used as fixed input orbits in a process to compute the SLR tracking data residuals. The processing software and SLR dataset are exactly identical for all these solutions, so that differences in the residuals can only be caused by differences in the input orbits.

Data editing
The way in which the SLR data is edited at ESOC is straightforward:

All SLR observations are processed for every orbit solution separately, without rejecting any data.
A single table can now be created with all 1-way SLR residuals, i.e. one row per SLR measurement and one column for every orbit solution.
Two further columns are added to the table with the number of accepted/rejected residuals (per observation) after applying a +/- 5 cm rejection window (equivalent to a 2-sigma rejection level, the global RMS over the SLR residuals is 2.51 cm).
An observation is accepted if at least 60% of the available solutions show a residual within the 5 cm window. If the observation is accepted, it is accepted for all solutions, even for those that have a residual outisde the rejection window.
If an observation is not accepted on the above criterion, it is rejected for all solutions.

By consequence, exactly the same subset of SLR measurements is accepted for all solutions. The mean, RMS and sigma over the accepted residuals are included as the last three columns of the table. In total there are 5241 SLR data points for the period from 2002/06/03 to 2002/07/01 (both inclusive). From these, 4609 observations are accepted and 632 are. Of these rejections, 255 observations occur on MJD 52451 (June 26) for which the spacecraft attitude quaternion data is missing. As a result, the offset between the laser retroreflector and the spacecraft centre of mass can not be computed with sufficient accuracy. Only 377 observations (7% of the total set) are actually rejected for being bad data points.

Some additional comments

The followed editing method will not give preference to any particular orbit solution, but solutions that have actually used SLR data in the POD process tend to show smaller residuals than those orbits that are based on DORIS and/or GPS data only. Ideally, the editing of SLR observations should be based on solutions that are independent of the SLR data itself, but unfortunatley that leaves too few solutions for reliable analysis. If in a future update more GPS-based solutions become available, the objectivity of the data editing process may be improved.
The method does not (yet) apply editing by pass, so that occasional bad passes may be partially accepted. However, only those residuals in the pass that are below the 5 cm rejection level for 60% of the orbit solutions are accepted, so that the global results should not be affected by these bad passes in any significant way. An example of such a case may be the first pass over station 7832 on 2002/06/03, half of which has been rejected.
If - exceptionally - there is a gap in an orbit solution, the missing residual is set to 999 so that it will be automatically rejected from all analysis. It will however lead to a situation that for that particular orbit solution the statistics are based on fewer data points, making it more difficult to compare these results with the other solutions. Fortunately only one minor gap occurs, for solution csr___ds, which starts exactly at the boundary of cycle 15 and therefore misses the first pass over station 7105 on 2002/06/03. This will not affect the overall results in any significant way: adding the missing pass for 7105 at its global RMS of 1.635 cm would change the total RMS for csr__ds from 2.023 cm to 2.020 cm, which is a negligible difference.
The esoc__ds solution was obtained from the same POD software that is used here to compute the SLR residuals. The POD process for this solution is based on DORIS and SLR, so that the SLR results for esoc__ds will probably be biased in a favorable way. The actual orbit precision of esoc__ds should be worse than what is suggested from these SLR results.

Results
Table 1 below presents the overall SLR residual statistics for each orbit solution. The solutions have been sorted in order of RMS over the accepted data points.
The accepted SLR residuals are also shown in a separate figure for each contributed orbit solution. Please click on the icons in Table 2 below to obtain larger plots.

solution	nr obs	RMS	MEAN	SIGMA
*csr__gds*	4577	1.703	0.169	1.695
*gsfc_gs4*	4577	1.773	-0.148	1.767
*jpl__gps*	4577	1.916	0.047	1.915
*gsfc_dyn*	4577	1.974	0.091	1.972
*gsfc_red*	4577	2.010	0.023	2.010
*gsfc_gps*	4577	2.072	-0.271	2.054
*esoc_ds2*	4577	2.383	0.293	2.365
*esoc__ds*	4577	2.462	0.191	2.455
*gsfc_rex*	4577	2.493	0.068	2.492
*deos_grm*	4577	2.583	0.189	2.576
*csr___ds*	4577	2.645	0.105	2.643
*deos_jgm*	4577	2.671	0.350	2.648
*cnes_poe*	4577	2.841	-0.075	2.840
*cnes_gps*	4577	2.897	-0.300	2.881
*asi__int*	4577	5.567	0.155	5.565
*asi__ext*	4577	9.038	0.700	9.011

Table 1: SLR statistics per contributed solution (1-way range, cm)

csr__gds	gsfc_gs4	jpl__gps	gsfc_dyn
gsfc_red	gsfc_gps	esoc_ds2	esoc__ds
gsfc_rex	deos_grm	csr___ds	deos_jgm
cnes_poe	cnes_gps	asi__int	asi__ext

Table 2: SLR residuals per contributed orbit solution (2-way range)