The use of gravimetric data from GRACE mission in the understanding of polar motion variations

P>Tesseral coefficients C-21 and S-21 derived from Gravity Recovery and Climate Experiment (GRACE) observations allow to compute the mass term of the polar-motion excitation function. This independent estimation can improve the geophysical models and, in addition, determine the unmodelled phenomena. In this paper, we intend to validate the polar motion excitation derived from GRACE's last release (GRACE Release 4) computed by different institutes: GeoForschungsZentrum (GFZ), Postdam, Germany; Center for Space Research (CSR), Austin, USA; Jet Propulsion Laboratory (JPL), Pasadena, USA, and the Groupe de Recherche en Geodesie Spatiale (GRGS), Toulouse, France. For this purpose, we compare these excitations functions first to the mass term obtained from observed Earth's rotation variations free of the motion term and, second, to the mass term estimated from geophysical fluids models. We confirm the large improvement of the CSR solution, and we show that the GRGS estimate is also well correlated with the geodetic observations. Significant discrepancies exist between the solutions of each centre. The source of these differences is probably related to the data processing strategy. We also consider residuals computed after removing the geophysical models or the gravimetric solutions from the geodetic mass term. We show that the residual excitation based on models is smoother than the gravimetric data, which are still noisy. Still, they are comparable for the chi(2) component. It appears that chi(2) residual signals using GFZ and JPL data have less variability. Finally, for assessing the impact of the geophysical fluids models choice on our results, we checked two different oceanic excitation series. We show the significant differences in the residuals correlations, especially for the chi(1) more sensitive to the oceanic signals.