| The objective of this research is to link satellite radar altimeter measurements from the GEOS-3, SEASAT, GEOSAT, ERS-1/2 and TOPEX/POSEIDON (T/P) satellites, using the T/P dynamical and geophysical modeling, T/P altimeter measurement corrections, and T/P altimeter measurements as a reference. The link was accomplished in three steps. First, all altimeter measurements were corrected with T/P measurement and geophysical corrections, and all satellite orbits were computed using the T/P dynamical modeling. Second, dual crossover residuals were computed with the along track T/P mean sea surface based on the geographically organized altimeter measurement database (stackfiles). Each stackfile is organized based on the repeating groundtrack of one satellite altimeter satellite. Finally, a relative bias and harmonic coefficients for the once per revolution radial orbit error were estimated from the corrected along track dual crossover residuals, using an optimal sequential Kalman filter. Application of the filter results to the orignal altimeter measurements completed the link process.; The consistency of the T/P dry and wet tropospheric corrections, based on the European Centre for Medium-range Weather Forecasts (ECMWF), were investigated in detail and their accuracy was found to be a function of time and geographical location. The stability of the T/P altimeter bias over Galveston Bay was investigated, and was found to have a mean of 1 {dollar}pm{dollar} 65 mm with a drift of 8 {dollar}pm{dollar} 4 mm/year. This drift, however, was found to be caused by large drifts in the tide gauges used. The altimeter bias drift was also determined over the Great Lakes and was found to be in closest agreement with the postglacial rebound corrected tide gauge observations (0.2 {dollar}pm{dollar} 1.7 mm/year). The computed relative altimeter biases from the Kalman filter estimates for GEOS-3 ({dollar}-{dollar}55.0 {dollar}pm{dollar} 19 cm), SEASAT (34.2 {dollar}pm{dollar} 2.8 cm), and GEOSAT (12.1 {dollar}pm{dollar} 1.5 cm) were applied to altimeter measurements over the Caspian Sea and the derived water level variation was found to be in close agreement with tide gauge observations. The relative bias for GEOSAT was also verified over the Great Lakes and estimated to be 12.9 {dollar}pm{dollar} 1.6 cm. |
