Separating climate and deep Earth signals in satellite gravimetry: a global assessment

SUMMARY This study aims to evaluate the effectiveness of the remove-restore method applied to GRACE (Gravity Recovery and Climate Experiment) gravity solutions, in which climate-related signals are first removed to allow a more meaningful interpretation of residual gravity signals associated with dynamic processes in Earth’s deep interior. By removing seasonal cycles and long-term trends, the analysis focuses on non-seasonal variations where causal attribution is clearer. Results indicate that climate correction reduces GRACE signal variability by approximately 30 per cent over both oceanic and continental regions, with the strongest impact observed in major river basins. The correction is most effective for temporal scales below 10 yr and spatial scales up to spherical harmonic degree 25. While overall variability decreases, certain frequency bands exhibit increased variability, suggesting a potential degradation of the signal due to model or data limitations. Globally, correlations between corrected GRACE signals and key climate indices largely diminish, confirming substantial removal of climate-related variability. However, the climate contribution to time-variable gravity beyond seasonal scales likely exceeds 30 per cent, indicating incomplete correction and occasional alteration of residual signals that complicate the interpretation of deeper Earth processes. Despite these challenges, climate model-based correction shows promise for advancing source separation and deepening understanding of Earth’s interior dynamics via time-variable gravity data, contingent on future improvements in climate modelling.