The Effects of Varying Cap Sizes on Geoid Computations: Experiences with FFT and Ring Integration

Abstract

Recent computations comparing geoid heights computed with varying cap sizes in both Australia and Scandinavia, using both FFT and Ring Integration techniques, show unambiguously that the quality of the gravimetric solution varies significantly with cap size. The Australian tests show that the two different approaches produce similar results, and that when compared against control of good quality and over all possible combinations of the points, the best comparison (for both techniques) occurred for an integration cap size between 0.2 and 0.4 degrees. Increasing the cap size beyond this often degraded the comparisons. These results have serious consequences for it reinforces the view that the short wavelength contribution to the geoid height varies with cap size, and that this is not an artifact of the Ring Integration technique. It is therefore important to modify either the cap size and/or the kernel function of the integration to achieve optimal results.