Fill the 'blind spot' of satellite observation and greatly improve the accuracy of the Antarctic gravity field model with aerial data

Antarctica hosts the world’s largest ice shelf system, and its gravity field constitutes essential data for investigating ice-shelf mass balance, sea-level variation and polar resource exploration. Nevertheless, gravity field modelling in polar regions faces two major challenges. First, limitations on satellite orbital inclinations (e.g., the GOCE mission lacks observations poleward of 83.3°), which introduces ill-posed problems in satellite gravity inversion. Although efforts have been made to mitigate this issue using Tikhonov regularization, spherical cap regularization and other techniques, the applicability of different regularization approaches over polar areas has not been systematically assessed. Second, multi-source gravity data suffer from inherent bandwidth limitations, and investigations on the spectral characteristics of combined gravity models remain insufficient. To address these issues, this study employs international Antarctic airborne gravity datasets to systematically evaluate the accuracy of various satellite-only and combined global gravity field models over Antarctica from both spatial and spectral domains.
This study aims to quantify accuracy discrepancies among different global gravity field models across Antarctica and explore how regularization schemes and data fusion strategies affect model performance. Special attention is paid to model behaviour within the “polar gap” (regions poleward of 83.3°) as well as the signal contribution of each model across different spectral bands. The ultimate goal is to provide guidance for model selection and gravity field optimisation in Antarctic geophysical research. We utilise international Antarctic airborne gravity datasets, including AntGG, PolarGap and multiple datasets released by the British Antarctic Survey (BAS). These airborne measurements provide Free-Air Anomalies (FAA), with more than 11.07 million data points located south of 60°S. Six satellite-only gravity field models (e.g., DIR_R6, GOCO06s) and five combined gravity field models (e.g., EGM2008, XGM2019e) are evaluated. Spherical harmonic expansion is adopted to compute gravity disturbances from each model at airborne survey heights. Root-Mean-Square Error (RMSE) is used to quantify the misfit between modelled values and in-situ observations. Furthermore, coefficient difference spectra and degree variances are calculated to assess spectral consistency between each model and the high-resolution reference model XGM2019e_2159. We also analyse how RMSE of combined models evolves with the spherical harmonic truncation degree. The evaluation demonstrates that global gravity field models exhibit degraded quality within the polar gap (south of 83.3°S). Incorporating additional observations together with regularisation can substantially improve accuracy. Among all tested models, XGM2019e_2159 achieves the best performance over Antarctica. Coefficient difference analysis reveals that pure GOCE-derived models are strongly affected by the polar observation gap, whereas models assimilating airborne gravity data (such as TIM_R6e and XGM2019e_2159) show remarkable improvements in spectral characteristics.

Our systematic assessment reveals substantial accuracy differences between global gravity field models over Antarctica. Spatially, errors are predominantly confined within −30 to 30 mGal, and higher errors appear along coastlines and inland zones with rugged topography owing to limited resolution of satellite gravity data. XGM2019e_2159, which incorporates airborne gravity observations, outperforms other models with an overall RMSE of 10.52 mGal. Spectrally, models including high-resolution airborne constraints (XGM2019e_2159, TIM_R6e) achieve better spectral consistency with the reference model. This study also indicates that, apart from XGM2019e_2159, simply increasing the maximum spherical harmonic degree cannot effectively improve the performance of other combined models over Antarctica, because they are primarily constrained by low-resolution satellite observations. Over marine areas, the inclusion of satellite altimetry significantly enhances the quality of combined gravity models. Overall, we recommend that combined gravity models assimilating airborne gravity data such as XGM2019e_2159 should be prioritised for high-precision applications in Antarctica.

Title:
Quality assessment of global gravity field models in the Antarctic region based on airborne gravity data
Journal Name: Geodesy and Geodynamics
Funder:
This research was supported by the National Key Research and Development Program, China (Grant No. 2024YFC2813500)
DOI: https://doi.org/10.1016/j.geog.2025.08.001
URL:https://scholarcast.cast.org.cn/top-news/2080484271972810752