Regional gravity field recovery using the GOCE gravity gradient tensor and heterogeneous gravimetry and altimetry data

Abstract

A regional approach using Poisson wavelets is applied for gravity field recovery using the GOCE (Gravity Field and Steady‐State Ocean Circulation Explorer) gravity gradient tensor, heterogeneous gravimetry data, and altimetry measurements. The added value to the regional model introduced by GOCE data is validated and quantified. The performances of the solutions modeled with different diagonal components of GOCE data and their combinations are investigated. Numerical experiments in a region in Europe show that the effects introduced by GOCE data demonstrate long‐wavelength patterns on the centimeter scale in terms of quasi‐geoid heights, which may allow reducing the remaining long‐wavelength errors in ground‐based data, and improve the regional model. The accuracy of the gravimetric quasi‐geoid computed with a combination of three diagonal components is improved by 0.6 cm (0.5 cm) in the Netherlands (Belgium) compared to that derived from gravimetry and altimetry data alone, when GOCO05s is used as the reference model. Moreover, the added value from GOCE data reduces the mean values of the misfit between the gravimetric solution and GPS/leveling data. Performances of different components and their combinations are not identical, and the solution with vertical gradients is best when a single component is used. The incorporation of multiple components shows further improvements, and the combination of three components best fits the local GPS/leveling data. Further comparison shows that our solution is the highest quality and may be substituted for existing models for engineering purposes and geophysical investigations over the target area.

The Impact of Noise in a GRACE/GOCE Global Gravity Model on a Local Quasi‐Geoid

“…This question will be answered by comparing the QG model with a QG model computed with the RCR technique. The RCR‐based QG model used in this study is significantly better than the model in Wu et al (). Some of the data sets used in Wu et al () suffer from systematic errors, which were not corrected for.…”

“…Some of the data sets used in Wu et al () suffer from systematic errors, which were not corrected for. Moreover, the GPS/leveling control data set over Belgium used in Wu et al () is a preliminary data set, which comprises a subset of the control data set used in our study, and suffers from systematic errors on the order of several centimeters, which were not known at the time the study (Wu et al, ) was conducted. Does the weighted least squares combination of a GGM with local data sets improve the quality of the QG model? This question will be answered by comparing the two QG models mentioned before with independent GPS/leveling data.…”

A multilayer approach and its application in modeling QGNSea V1.0: a local gravmetric quasi-geoid model over the North Sea

“…More specifically, the low-order details demonstrate the high-frequency signals stemmed from the shallow and small-scale substances. While, the high-order ones with the local topography (the local digital terrain model (DTM) could be found in Figure 1 in Wu et al, 2017b). We mainly attribute this to the uncorrected signals in RTM corrections, which is due to the inaccuracy of the density parameters in RTM corrections and limitation of DTM both in terms of spatial resolution and precision.…”

“…The details for data pre-processing procedures can be found in Wu et al (2017b), where crossover adjustment and low-pass filter were applied to remove systematic errors and reduce high-frequency noise, respectively, and datum transformations were performed on all the data. Moreover, the satellite-only reference model called GOCO05s with a full degree and order (d/o) of 280 15 (Mayer-Gürr et al, 2015) and RTM corrections were removed from the original observations to decrease the signal correlation length and smooth the data within the framework of remove-compute-restore (RCR) framework, and the details for the RTM reduction and residual gravity data could be found in Wu et al (2017b). where i  is computed as the difference of the spherical scaling functions with low-pass filter characteristics between the consecutive levels i +1 and i , but also can be expressed as the SRBF has the band-limited properties in the frequency domain (Schmidt et al, 2007).…”

“…Typically, the widely-used SRBFs method is implemented by the so-called single-layer approach, i.e., the parameterization of gravity field is only based on a single-layer of SRBFs' grid (Wittwer, 2009;Bentel et al, 2013;Wu et al, 2017b). However, one layer of SRBF's parameterization may be only sensitive to parts of signals' spectrum and reduce the quality of the solution.…”