Multidisciplinary analysis of the temporal variations of the gravity field acquired during the GRACE and GRACE-FO missions

The GRACEFUL project aims at characterizing time-variable processes occurring at the core-mantle boundary and in the fluid core using a synergy of satellite data of the geomagnetic field, gravity field and Earth's rotation. In order to detect deep Earth's signals, the satellite measurements of the gravity field should be corrected for surface water mass distributions and unrelated geophysical processes. Here, we use an ensemble of state-of-the-art models to correct the effects of the ocean-atmosphere circulation, continental water cycle and glacial isostatic adjustment. Over the ocean, the input of freshwater from land-based ice melting is well captured by a simple parametric model including an acceleration and a linear trend. The residual ocean mass variations are fairly accurately predicted by ocean-atmosphere circulation models, especially in the Southern Ocean, where the variance of GRACE signal is reduced by more than 80%. However, systematic differences are observed over the continents, that are larger than the differences between hydrological models or than the differences between the gravity field solutions from different centers. The amplitude of decadal trends and interannual signals is systematically underestimated in hydrological models when compared to GRACE measurements. The largest differences, characterized by decadal oscillations with amplitudes of a few centimeters of equivalent water height, are observed over all major river basins (e.g. Amazon, Zambezi, Okavango, Volga). Significant residual mass variations are also detected at shorter time-scales, with a singularity around 6 years, also present in the satellite observations of the Earths rotation and the magnetic field. While the origin of such signals is unknown, part of it is still likely to be climate-related and at least partially inaccurately predicted by hydrological models. To be able to interpret potential links between the residual mass anomalies and processes originating from the deep Earths interior, efforts must be made to understand the systematic differences observed between the GRACE measurements and hydrological models at interannual and decadal time-scales.