RegGRAV - Regional gravity field modeling as multi-resolution representation estimated from the combination of heterogeneous data sets
Abstract:Models of the Earth’s gravitational potential rely on heterogeneous data sets. Typically, regional gravity data with high spatial resolution are combined with low-resolution global observations. We achieve an optimal combination within a regional gravity field modeling approach using radial basis functions, which emphasizes the strengths of each data set. As the resulting regional models contain gravity information from different frequency bands, they can be used as basis for various applications, such as national geoid determination, detection of mass anomalies in the Earth’s interior or determination of mass and height variations at the Earth’s surface.
We developed powerful software called “RegGRAV” that enables the flexible combination of high-resolution terrestrial and airborne data, down to mid-resolution altimetry and low-resolution satellite gravimetry measurements. The combination on normal equation level includes a full stochastic model and error propagation. We set up our approach as multi-resolution representation (MRR), i.e. modeling the gravitational potential and decomposing it into frequency-dependent detail signals. The basis functions are defined as spherical wavelet functions and located in the region of interest in such a manner, that the highest measure of information of the input data is captured depending on its spectral sensitivity. After determining the appropriate scaling coefficients of the functions by variance component estimation, any functional of the Earth’s gravity field can be derived, as e.g. geoid heights or gravity anomalies.
A comparison of the developed MRR to accordingly filtered global gravity field models shows the additional value of our regional approach in the high frequency domain and validates our results. This (1) verifies the consistent estimation of local gravity field parameters for various measurement techniques and (2) enables identifying areas in existing global gravity field models where high-resolution data are missing.