One-Centimeter Orbit Determination for Jason-1: New GPS-Based Strategies

Abstract

The U.S./French Jason-1 satellite is carrying a state-of-the-art GPS receiver to support precise orbit determination (POD) requirements. The performance of the Jason-1 “BlackJack” GPS receiver was strongly reflected in early POD results from the mission, enabling radial accuracies of 1–2 cm soon after the satellite's 2001 launch. We have made further advances in the GPS-based POD for Jason-1, most notably in describing the phase center variations of the on-board GPS antenna. We have also adopted new geopotential models from the Gravity Recovery and Climate Experiment (GRACE). The new strategies have enabled us to better exploit the unique contributions of the BlackJack GPS tracking data in the POD process. Results of both internal and external (e.g., laser ranging) comparisons indicate that orbit accuracies of 1 cm (radial RMS) are being achieved for Jason-1 using GPS data alone.

Keywords:

• precise orbit determination
• GPS
• satellite altimetry
• Jason-1

Acknowledgments

We are grateful to Marek Ziebart and his colleagues at University College London for providing the high-resolution surface force model for Jason-1. We thank Jean-Paul Berthias and the CNES POD team for providing satellite-specific models for Jason-1 and for supplying the DORIS data. We are grateful to John Ries at The University of Texas for providing the precise SLR/DORIS orbit solutions, and SLR station coordinate and eccentricity information. Finally, we thank Scott Luthcke and his colleagues at NASA Goddard Space Flight Center for insightful discussion on their experiences with the processing of the Jason-1 BlackJack data. The results described in this article were enabled by a long-term investment in GPS infrastructure and technology by the NASA Offices of Solid Earth and Natural Hazards and Earth Science Technology. Additional support was provided by the NASA Physical Oceanography Program and the Jason-1 Project at JPL. The work described was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

Notes

¹ The CNES model is described at ⟨http://calval.jason.oceanobs.com/html/calval_plan/pod/modele_jason.html⟩ Note that a new, high-resolution model based on pixel-array techniques (Ziebart et al. 2003) was used in limited tests (see text).

¹ One of the BlackJack receiver design philosophies is to advance new technology. This philosophy carries the implication that many of the parts are not available in a radiation-hardened form. As a consequence, system resets were designed into the autonomous receiver operations as a means of clearing “soft bit” errors induced by cosmic rays and the occasional “latch up” condition. A significant number of the Jason-1 BlackJack resets are occurring over the South Atlantic Anomaly, strongly suggesting that they are radiation induced. In contrast, BlackJacks in more radiation-benign orbits suffer many fewer resets, e.g., 3–5 per month for GRACE-A.

² Our estimates of the GRACE phase-center variations, discussed later, indicate an offset in the same (zenith) direction, but with significantly smaller magnitude (c.f., Figure 2). It must be remembered, however, that the GRACE POD strategy is much further removed from “dynamic” than that of Jason-1. This might attenuate the expression of the phase-center offset in the postfit residuals.

³ An approximate 2-cm offset of the antenna phase center location relative to the CG in the s/c X direction (Haines et al. 2002) is now recognized as a consequence of a probable error in the modeled CG position (e.g., Choi et al. 2004; Luthcke et al. 2003) and is treated as such.

⁴ We used the following criteria for selecting the six stations: (1) minimum of 100 high-elevation Jason-1 passes for the period spanned by our orbit solutions; (2) short-term range bias stability of better than 16 mm from weekly analysis of LAGEOS data by UT/CSR. (2003 4th quarter, as reported at http://ilrs.gsfc.nasa.gov); (3) normal-point (“quick-look”) data from Crustal Dynamics Data Information System (http://cddisa.gsfc.nasa.gov) holdings require no additional adjustment for range and/or timing biases.

⁵ Sometimes called the geographically anticorrelated error, this refers to errors that are equal in magnitude, but opposite in sign for ascending and descending tracks at a location of interest (e.g., Chelton et al. 2001).