Accuracy of regional geoid modelling with GOCE
20/09/2016 | 17:15 | Session 3: Local/regional geoid determination methods and models
Author(s): Christian Gerlach and Vegard Ophaug
Christian Gerlach and Vegard Ophaug
Regional geoid models are based on the combination of satellite-only gravity field information and terrestrial data. Today, satellite information is based on the combined adjustment of observations from the GRACE and GOCE satellite missions and is usually provided in terms of a spherical harmonic global gravity field model, e.g., GOCO05s. Terrestrial information is usually provided as a grid of block mean gravity anomalies in the region of interest. Optimized combination of satellite and terrestrial information may be based on Stokes integration, applying the spectral combination method, where error degree variances of satellite and terrestrial gravity information are used for providing a Wiener-filter type weighting schema for the integral kernel. We combine terrestrial gravity field information with GOCE-based global models for selected study regions in Germany and Norway and compare the results to GNSS-levelling. As trade-off between optimized data weighting and simplicity/efficiency of computation we consider the block-diagonal elements of the variance-covariance matrix of the global model. The final aim is not only to allow optimized data combination, but also to provide realistic formal error bounds for the combined regional model. This is important for validating or combining the outcome with external/independent data. It allows, e.g., quantification of systematic distortions in GNSS-levelling networks, or to set up a realistic error budget for a geodetic model of the ocean's mean dynamic topography.