Oceanic gravity by analytical inversion of Hotine's formula

An analytical inversion of the Hotine formula is developed using fast Fourier transform techniques. Detailed mathematical derivations are used to explain the concepts behind the inverse transformation. Three modifications of the analytical inversion of the Hotine formula are compared and tested using both synthetic data from the OSU91A geopotential model and real GEOSAT altimetry data from the Exact Repeat Mission. The stability of this inverse Hotine approach is investigated using simulated data, and numerical tests are done to quantify the stability of this approach. The approach seems to be numerically stable without employing any stabilization technique. Estimated gravity information from GEOSTAT altimetry data is compared to marine gravity data from shipboard measurements in the Orphan Knoll area. The standard deviations and mean values of the differences between satellite and marine gravity disturbances are 8.2 and 2.9 mGal for the planar approximation, 9.2 and 3.7 mGal for the spherical approximation, and 9.5 and 1.9 mGal for the Molodenskii‐like approximation, respectively, indicating that latitude‐dependent errors affect the latter two approximations. Such errors could be eliminated by performing the calculations by the rigorous one‐dimensional fast Fourier transform (FFT) technique, and any data noise could be filtered out by utilizing covariance knowledge about the input geoid undulations and their errors. Simulation studies also showed that the accuracy of the techniques (for all approximations) can reach a root‐mean‐square (RMS) level of only a few mGal when proper treatment of FFT edge effects is employed and a rather wide area of results is disregarded around the edges.