On the accuracy of gravimetric geoids and the recovery of oceanographic signals from altimetry

Abstract

Any attempt at recovering the surface expres‐sion of time‐averaged circulation from altimetric measurements of sea surface topography requires both an independent geoid and a realistic assessment of its accuracy. It is shown that publicly available marine gravity data over the North Atlantic yield expected geoid accuracies ranging from 30 to 260 cm (root mean square, lower bound), depending on location. Most of the error is due to unsampled short wavelengths in the gravity field. These expected errors underestimate the difference between sea surface (measured from SEASAT) and the geoid computed here, by a factor of about two. Most of this discrepancy can be attributed to a failure of the power spectral model used to describe the geoid at short wavelengths. The difference between altimetric and geoidal surfaces is dominated by geoid errors. It is shown that oceanographic information can only be identified after the expected signal and the expected errors are included in an optimization scheme.