Validation of Oceanic Mass Changes derived from GRACE Gravimetry: A comparison with independent data from Bottom Pressure Recorders in the Southern Ocean

The satellite mission GRACE (Gravity Recovery and Climate Experiment) provides monthly solutions of the global gravity field of the Earth. Temporal variations in gravity reflect mass redistributions over the Earth. Those mass movements are caused by physical phenomena such as ice sheet dynamics, accumulation of hydrological masses over land, the dynamics in the solid Earth and the dynamics in the atmosphere and ocean. This thesis work focuses on the validation of the oceanic mass variations derived from GRACE. The goal of the project is to compare the oceanic estimates from the space-borne instrument with independent data from pressure recorders (BPRs), deployed on the bottom of the deep ocean in the southern ocean. Although those datasets seem to be of a completely different nature they measure the same quantity. Namely, the variations of mass in the atmosphere and ocean, which in their turn influence the dynamics of the ocean. In this study the GRACE data is compared to the following two quantities: - Local bottom pressure at the BPR stations - Pairwise differences of local bottom pressure (a measure of the oceanic transport) The processing methodology applied to the BPR records is as follows: The raw BPR data is de-tided using a harmonic analysis method. The data is then averaged (over a month) to make them equivalent to the GRACE solutions. GRACE data from three different computing centers1 are compared. The spherical harmonic coefficients are converted to equivalent water height and geostrophic flow, indicators of the variations in the ocean and atmosphere. The solution is filtered to remove correlated errors, after which the de-aliasing models are added back to the solutions. The final solution is obtained by applying a spatial averaging filter. The bottom pressure comparison showed good agreements at locations in the southern Indian ocean, the south Atlantic and the Argentine basin. The best correlations obtained values over 0.8 and typical errors of below 1.5 cm. The analysis showed an apparent connection between phenomena in the Argentine basin and near Crozet. A two monthly mode appeared to be present in both areas. The capabilities of GRACE to detect this mode is marginal and further investigation is required. The comparison for the geostrophic flow (pressure differences) showed a significant agreement in the south Indian ocean, with correlations over 0.7 and errors as low as 7 mm/s. Correlations in other areas were weak or even non existent. We conclude that, for dedicated areas, GRACE shows the ability to measure realistic large scale variations in bottom pressure and geostrophic flow on spatial scales yet barely explored.