Spectral analysis of gravity field data in view of sub-decimeter geoid determination.

摘要

Recent developments in the modern geodetic, geophysics and oceanographic applications require a geoid with absolute accuracy of 10 centimetre or better and a relative accuracy of 1 part per million (ppm) of the inter-station distance. Gravity field data in Canada are spectrally analysed with the view of refining geoid estimation methods that will yield the above-mentioned accuracy requirements. The analysis is based on estimates of empirical covariance functions and degree variances derived from local gravity observations, a global geopotential model (EGM96), and topographic heights.; Numerical results for selected areas in mountainous, flat and marine areas of Canada show that the empirical signal and error covariance functions are non-uniform and they are highly correlated with the roughness of the topography. Gravity data and topographic heights with 1′ spatial resolution are required for 1 cm geoid in the mountainous areas while the same level of geoid accuracy can be achieved with 5′ data set in the flat areas. In addition, rigorous modelling of the topographic effects with actual topographic density values, as well as combination of the solution from the GM and Stokes's integral using gravity data in a cap size of 10° x 10° is required for accurate geoid estimation.