The aim of this work is to analyse the orbit errors of the TOPEX/Poseidon orbits delivered by the project. Results are given in term of orbit corrections computed using SLR data. As a by-product SLR residuals are given to assess the quality of each SLR stations.
The laser-based short-arc technique has been described in detail in Bonnefond et al. (1995). In this section, we will only recall its basic concepts and will describe the various parameters and data used (geographical areas, input orbits, SLR data, and set of station coordinates).
The approach we use for the short-arc orbit determination strategy is to assume that a long-arc orbit is available covering several days (e.g., one repeat cycle) determined from a given global tracking data set (in practice, essentially SLR and DORIS data). We determine corrections to this dynamic orbit for short arcs that are typically of 10 to 15 min. duration and so of length up to about 4000 km. Let us note that the corrected tracks of the satellite are no longer exact solutions of the differential equation system for its motion. Instead of dynamically fitting short arcs, we determine, in fact, kinematic corrections representing local orbit errors as well as station coordinate errors or systematic errors in the tracking data. The values of these corrections to be applied to the input orbit are estimated in a least squares procedure from the intensive SLR tracking data that is assumed to be available along the short arcs. Moreover, criteria on the geometrical configuration - between the tracking network and the passes - have been determined and selected in order to guarantee a short-arc radial precision better than 2 cm. This imply that some passes can never be corrected, decreasing, as a consequence, the number of possible determinations over a given area by a factor of 3 to 4 generally [Bonnefond et al., 1995].
The short-arc technique has been applied over two areas. One is located in Europe and roughly centered to the Mediterranean (+15°<latitude <+60° and -25°< longitude <+72°). The other, called hereafter the US area, contains California and a part of the Pacific ocean (0°< latitude<+50° and -180°< longitude <-80°). The main reason for choosing such areas is largely due to the geographical configuration of the permanent SLR tracking network. In addition, the spatial characteristics of the T/P radial orbit error, as assessed through orbit comparisons or perturbation analyses, exhibit long wavelength patterns over the two areas (Figure 1). Their amplitude can be assessed independently by our analysis. As an example, the mean values of the radial orbit differences only generated by gravity modeling differences (JGM-2 minus JGM-3) over the Mediterranean and US areas are expected to be at the level of +1.7 cm and -1.7 cm, respectively [Exertier and Bonnefond, 1997].
Figure 1. Localisation of the studied areas. Colors correspond to Geographically Corralated orbit Errors (GCE) in the radial component due do differences between JGM-2 and JGM-3.
The Precise Orbit Ephemeris are provided twice, by the Goddard Space Flight Center (GSFC) and Centre National d'Etudes Spatiales (CNES) groups [Tapley et al., 1994a; Nouël et al., 1994], and both used JGM-3 gravity field. But let us note that the JGM-3/CNES solution is not purely dynamic. It uses a DORIS-based reduced-dynamic technique (ELFE solution, [Barotto et al., 1996]) permitting certainly to reduce a part of the geographically correlated orbit errors and thus explaining this temporally invariant orbit difference.
15-sec normal points (NP), determined routinely at GSFC for the T/P POE, have been used including the nominal calibration already applied by the SLR stations. Because of the limited SLR coverage it is possible, statistically, to adjust only a limited number of short arcs. There are 6 and 11 passes per cycle in average which have been corrected by the short-arc technique for the Mediterranean and US areas respectively. This represents about 35% and 54% of the selected passes, for which the geometrical criteria have been reached, showing differences in the tracking capabilities of the two areas.
ITRF 97 has been used for these computations [Boucher et al., 1999]. See Table1 for the stations used. Because SLR stations used can changed, depending on the strategy, always refer the List of Stations for an updated list.
Table 1. SLR Stations used in the computations
| Station No. | Site Name | Geographic Location | Region | Site Plate | Type† |
|---|---|---|---|---|---|
7080 |
McDonald |
McDonald Observatory, Fort Davis, TX |
North America |
North American |
BASE |
7105 |
Greenbelt |
GGAO, GSFC, Greenbelt, MD |
North America |
North American |
MOBLAS |
7109 |
Quincy |
Quincy,CA |
California |
North American |
MOBLAS |
7110 |
Monument Peak |
Mount Laguna,CA |
California |
Pacific |
MOBLAS |
7210 |
Haleakala |
Lure Obs., Mount Haleakala, Maui, HI |
Pacific |
Pacific |
BASE |
7810 |
Zimmerwald |
Bern, Switzerland |
Europe |
Eurasian |
BASE |
7831 |
Helwan |
Helwan, Egypt |
Mediterranean |
African |
BASE |
7835 |
Grasse |
Grasse, France |
Europe |
Eurasian |
BASE |
7839 |
Graz |
Graz, Austria |
Europe |
Eurasian |
BASE |
7840 |
Herstmonceux |
Royal Greenwich Obs., Great Britain |
Europe |
Eurasian |
BASE |
†BASE= Fixed station; MOBLAS= Mobile station
Table 2 describes the parameters used in CALTIM Software to compute SLR residuals.
Table 2. Parameters used for SLR residuals computations (CALTIM software, CERGA)
| Corrections and parameters | Type | References |
|---|---|---|
| Set of coordinates | Boucher et al. (1995) | |
| Solid Earth tides | McCarthy [1992] | |
| Earth orientation parameters | ||
| SLR Corrections | ||
| Troposphere | Marini-Murray |
McCarthy [1992] |
| Relativity | ?? |
?? |
| Signal reflection and RC / CM* (T/P) | Schwartz [1990], Tapley et al. [1994] |
*This refers to the position of the reference center of the laser ring relative to the satellite center of mass. For geodetic satellites constant correction is applied (e.g., Lageos I&II).
To
TOPEX/Poseidon Short-Arc Analysis Home Page