A preliminary model for Geosat altimeter data errors

Noise power spectral densities (PSDs) are derived from the analysis of GEOSAT altimeter data from segments of nearly collinear repeat tracks. The difference time series, which are analyzed using autoregressive (AR) modeling techniques, contain primarily time‐varying oceanographic signals and altimeter noise, since geoid content is canceled by the differencing process. An average noise PSD, obtained from eight independent repeat‐track PSDs, is fitted closely by a first‐order Markov model with a white‐noise floor. The GEOSAT average noise PSD is compared with analogous spectra derived earlier for GEOS‐3 and SEASAT. In general, the GEOSAT noise PSD is the same as the SEASAT noise spectrum at wavelengths greater than about 100 km, but has a lower noise level at shorter wavelengths.