The two multi-year gapped periods for which we have DSCs cover (1985-1988) for Geosat and (1992-1998) for Ers1 and T/P, and involve Geosat-T/P and Geosat-Ers1 differences. Since there is no current collinear altimetry extending into the 1980's we can only verify the consistency of our results by checking with 1990's data. We had two choices for this data, from T/P or Ers1/Ers2 (Ers2 altimetry, in the Ers1 tracks, extends from 1995 to 1998).
We chose Ers1/Ers2 collinear data because it was more convenient in terms of the DSCs available but more important because the addition of Ers2 altimetry added another degree of independence to the consistency test.
First we adjusted a 50x50 geopotential from an all-Pathfinder set of SSCs (Geosat, Ers1 and T/P) together with 1 year averages of DSCs for Geosat-Ers1 [(Apr87 thru Mar88) - (Apr92 thru Mar93)] and Geosat-T/P [(Apr87 thru Mar88) - (Apr95 thru Mar96)]. To verify that the residuals of the Geosat-T/P altimetry were consistent with those of Geosat-Ers1 we took the yearly average of collinear differences of Ers2 and Ers1 over the periods in the 1990's between the T/P and Ers1 legs of our DSCs [(Apr95 thru Mar96) for Ers2 - (Apr92-Mar93) for Ers1)].
Formally, to connect our residual DSC data sets, we have:
with G for Geosat and E1, E2 for Ers1, Ers2.
Figure 13 compares sealevel differences between 87/88 and 95/96 from the above transfer of Geosat-Ers1 residuals (with the support of independent Ers2-Ers1 collinear altimetry) with the "observed" residuals of Geosat-T/P DSCs. The agreement is seen to be quite good in spite of the considerable differences in data and processing involved in the two sealevel difference estimates. Aside from the different media corrections involved in Ers1, Ers2 and T/P the collinear differences (Ers2-Ers1) here should involve no geopotential signals, should be a fair transfer of average sealevel between 1995/96 and 1992/93 and thus the agreement says our DSC residuals are also compatible with real sealevel differences.
Of course the above test does not directly address the Geosat altimetry in fundamentally different periods though the actual Geosat altimeter heights used in the two DSC sets (Geosat-Ers1 and Geosat-T/P) of the geopotential solution are not the same, having somewhat different timing and corrections within each same-month window. However, we have extended this test to use DSCs in wholy different periods of Geosat as well and also with Geosat collinear differences to verify the sealevel connections between them. The comparison (not shown) is still good, though not as close as in Figure 13.
We conclude that the residuals of DSC altimetry over multi-year gaps (after removing the geopotential and other smaller signals) are substantial measures of interannual oceanic change across these gaps.