THIN-PLATE SPLINE QUADRATURE OF GEODETIC INTEGRALS

Thin-plate splines - well known for their flexibility and fidelity in representing experimental data - are especially suited for the numerical evaluation of geodetic integrals in the area where these are most sensitive to the data, i.e. in the immediate vicinity of the computation point. Quadrature rules that are exact for thin-plate splines interpolating randomly spaced data are derived for the inner zone contribution (to a planar approximation) to Stokes's formula, to the formulae of Venzing Meinesz and to the L1 gradient operator in the analytical continuation solution of Molodensky's problem. The quadrature method is demonstrated by calculating the inner zone contribution to height anomalies in a mountainous area of Lesotho and carrying out a comparison with GPS-derived heights. Height anomalies are recovered with an accuracy of 6 cm.