Least Squares Collocation with GOCE Gravity Gradients and its Application to Moho-estimation in Central Europe
Rieser, Daniel1; Mayer-Guerr, Torsten2
1Institute of Theoretical Geodesy and Satellite Geodesy, TU Graz, AUSTRIA; 2Institute of Theoretical Geodesy and Satellite Geodesy, Graz University of Technology, AUSTRIA
With the advent of the GOCE gravity field mission, the unique observations of gravity gradients have become available to the users. While the main official products of this mission are global gravity field models in terms of spherical harmonics, the gravity gradient measurements are also a potential in-situ observation type for regional gravity field determination and applications. A very flexible method for deriving regional gravity fields is Least Squares Collocation (LSC). It offers the possibility to combine various input gravity field quantities, like gradients, and in turn also is capable of predicting arbitrary gravity field related quantities.
Using gravity gradients as direct observations in LSC is not straightforward and a consistent stochastic modeling is essential. To achieve this, some specific aspects of GOCE gradients have to be considered. As the gradiometer instrument is measuring precisely only in a bandwith from 5 mHz to 100 mHz, the observations have to be bandpass filtered. Furthermore, the GOCE gradients observed in this bandwith are suffering of coloured noise. The spectral behaviour of these long-wavelength errors can be approximated by building the residuals between the measured signal and a synthesized reference signal using a global gravity field model. In order to reduce the coloured noise, a further filtering step is introduced following the Wiener filter theory, which finally also leads to the required stochastic error information for collocation.
In this contribution, we want to utilize our LSC approach for estimating the Moho-depth in the central European region. By using a topography model the topographic influences can be reduced from the gradient observations. Under the assumption of constant mantle and crustal densities, surface densities are directly derived by LSC on regional scale, which in turn are converted in the respective Moho depths.