We've combined a single 17-day cycle of Geosat data from November, 1986 with 4 days of SDR data from May, 1998. Since the two missions are in the same exact repeat orbit, one can use collinear methods to remove the large orbit errors present in the GFO data (with operational Doppler orbits). Each pole-to-pole pass of GFO data has been fitted with a constant plus once-per-rev sinusoid, and this "orbit error" signal has been removed. The resulting height differences (May '98 minus Nov. '86) were gridded and are plotted for the tropical Pacific in the figure below. By removing the large-scale orbit error (with an amplitude on the order of 5 meters), we are able to discern the height changes in the tropics. The reference Geosat time period was intentionally chosen to be in the opposite season from the GFO data, so that the seasonal signal would be evident in the height change. Changes as large as +/- 30 cm are apparent, and the structure in the near equatorial regions is consistent with the seasonal cycle there. Preliminary checks of tide gauge differences between these two time periods confirms the magnitude and sense of the observed changes.