Résumé
The origin of the hotspot Tristan da Cunha located at the southwestern end of Walvis Ridge in the Atlantic Ocean is still a controversial topic. We especially question on the nature of the involved geodynamical processes and on their origin depth. The latest results based on local seismic and magnetic data (Schloemer et al., 2016; Baba et al., 2016; Geissler et al., 2016) suggest the existence of a plume coming from the mid-mantle in the southwest of the archipelago. Here we give a regional view of mantle dynamics patterns in the area by using the high-quality satellite geoid data. To extract the mantle signature, we estimate the crustal and lithospheric signals of the ocean basin and South American and African continents, which contribute to mid- and long-wavelengths in the total geoid. We pay particular attention to the modeling of continental margins and their effects on the residual geoid signal. In addition, we explore a large density values set derived from petrological and geochemical studies in the calculation of the lithospheric geoid model. After subtracting the lithospheric signature to the EGM2008 geoid, we apply a multi-scale analysis, which unfolds the different components of the geoid residual signal. The analysis underlines a set of positive anomalies at 200-400 km in the study area, notably in north and west of Tristan de Cunha, and a positive anomaly at 700-1100 km scale in the southwest of the archipelago. These patterns do not change by using different lithospheric geoid models, which allow us to evaluate the reliability of the residual geoid anomalies. These results indicate the existence of small-scale density anomalies in the upper mantle and a larger scale density anomaly in the mid-mantle. Our study suggests that a large dome topped by plume clusters could be a good candidate to explain the volcanism of Tristan da Cunha. Schloemer et al., 2016 Hunting for the Tristan mantle plume..., EPSL, http://dx.doi.org/10.1016/j.epsl.2016.12.028 Baba et al., 2016 Marine magnetotellurics imaged no distinct plume..., EPSL, http://dx.doi.org/10.1016/j.tecto.2016.09.033Geissler et al., 2016 Thickness of the oceanic crust, the lithosphere, and..., EPSL, http://dx.doi.org/10.1016/j.tecto.2016.12.013