Using Marine Data to Estimate the Accuracy of Free Air Gravity Anomaly Models Derived From Satellite Altimetry

Global gravity models from satellite altimetry missions have widely contributed to the improvement of the knowledge over the oceans. However, few models are available with a validated accuracy model, as the most recent models V21 to V32 from Scripps Institution of Oceanography (SIO). The aim of this paper is to propose a method to estimate the accuracy of any free air gravity anomaly model derived from satellite altimetry from validated marine data sets. The main idea is to correlate the error of the model to the mean roughness of the free air gravity anomaly on precise marine measurement areas. The relationship obtained is then generalized to each point of the model. This error grid is eventually combined with the grid provided by the satellite models, reflecting the errors related to the technology of the different altimetry satellites used. The method has been successfully applied to successive versions of SIO gravity anomaly models (from V18 to V31) to produce continuous accuracy grids over the North Atlantic Ocean. Validation of the results from independent marine data shows that the proposed approach allows to strongly improve the estimation of the used models' accuracy which converges, whatever the tested versions, toward a minimal value of 0.9 mGal over flat areas.