next up previous
Next: 3.1.2. Non-tidal mass Up: 3.1. Static and Previous: 3.1. Static and

3.1.1. Tidal modeling.

Of the temporal variations of the gravity field, tides are best understood for they have the largest effects and occur at well known astronomic frequencies. For near-Earth orbit determination, the tide modeling problem represents two distinct challenges: (1) to improve the long wavelength tidal terms which are in resonance with near-Earth satellites and give rise to long period perturbations and (2) to incorporate a large number of tidal coefficients spanning many tidal families which produce a whole class of short period orbital perturbations. As tracking precision has increased, the geopotential solutions have been extended to include long wavelength tides through direct recovery of spherical harmonic terms in the tidal expansion for major tidal constituents [e.g. Marsh et al., 1990; Shum et al., 1990a]. These simultaneous static/tidal geopotential solutions have provided accurate long wavelength terms to account for tidal mass redistribution in the integrated solid Earth/ocean/atmosphere system.

The selection for terms incorporated in the ``background tidal models'' is made in case (2) on the basis of orbital sensitivity analyses where terms having greater than a certain orbit perturbation cutoff are included. Casotto [1989,1991] used an analytical orbit theory to evaluate the ocean tidal perturbations on the TOPEX orbit. On the basis of this study, a set of spherical harmonic coefficients spanning over 80 tide lines was identified as being TOPEX-sensitive. Because of the large number of tidal terms required, Oscar Colombo [1989, unpublished notes] developed an efficient algorithm to evaluate these tidal models based on a linear scaling of tidal admittances as summarized in Nerem et al., [1993a]. By increasing the number of harmonic coefficients in both the static and tidal gravity models, the omission error on lower orbiting satellites is minimized.

In light of the extensive improvements in orbit accuracy achieved on T/P and the improved ocean tidal models resulting from an analysis of its altimetry [ Eanes, 1994; Pavlis et al., 1994; Schrama and Ray, 1994; Ray, Sanchez and Cartwright , 1994], the tide modeling problem has been recently revisited. Marshall et al., [1994c], by evaluating the spatial and temporal distribution of radial errors on the T/P orbit, observed a geographically dependent signal with a 60-day period which they attributed to mismodeled shorter wavelength tides. This 60-day period closely matches the aliasing period of the principal semi-diurnal lunar and solar tides (M and S) on T/P. Bettadpur and Eanes [1994] showed that a significant improvement (1 centimeter RMS) in T/P radial orbit knowledge can be achieved through incorporation of more complete and accurate background tide models.

The neglect of atmospheric tidal modeling in orbit determination, especially at the solar diurnal (S) frequency, is an area now receiving significant attention. Nerem et al., [1994b] proposed that an anomalous variation seen in the evolution of the eccentricity of the Lageos orbit in part arises from ignoring these tidal effects. The S and S atmospheric tides give rise to dominating long period orbital perturbations on LAGEOS of 561 and 365 days respectively. These correspond to the spectral peaks seen in the in-phase component of Lageos' one-cycle-per-revolution anomalous acceleration discussed in Eanes and Watkins [1991], Nerem et al., [1993b] and Gegout and Cazenave [1993]. However, surface forces, like those described later, are still not ruled out as a possible explanation.



next up previous
Next: 3.1.2. Non-tidal mass Up: 3.1. Static and Previous: 3.1. Static and



U.S. National Report to IUGG, 1991-1994
Rev. Geophys. Vol. 33 Suppl., © 1995 American Geophysical Union