In addition to these orbit signals in the altimetry we also recognize that the tracking stations for the three satellite missions may refer to different geocenters which may affect comparisons of sealevel between them. This possibility was first raised when the crossovers of Geosat and T/P were first examined (e.g., Wagner et al., 1997a) confirming the earlier finding (Rapp et al., 1994) that the mean sea levels of the two missions showed significantly different (1,1) and (1,0) harmonics. At the time it was felt that these differences could not be explained by either natural long term changes in sealevel between the missions (1985-88 for Geosat and 1992-94 for T/P) or significant first order errors in the geopotential used to track these orbits.
Here we also leave open all three possibilities for these effects by including in our crossover conditioning the following geocenter shift between any two missions (affecting all DSCs of the same pair in the same way, see Wagner et al., 1997a):
where the crossovers as well as the coodinate differences
are taken as [mission i - mission j], and dx and
dy are in the equator towards Greenwich and 
East respectively,
and dz is towards the North Pole.