The crustal structure of the eastern Walvis Ridge : a classical hotspot example?

Abstract

The Walvis Ridge is a more than 3000km long submarine ridge in the South Atlantic, which consists of single volcanoes and numerous smaller and larger ridges. It stretches from Namibia at the African continental margin to the volcanic island of Tristan da Cunha near the Mid-Atlantic Ridge. At some places this remarkable structure raises from the 5000 m deep sea basins to the sea surface or even above. The hypotheses about its origin are strongly related to the great debate about the existence of deep mantle plumes: The ridge is thought to be either the hotspot trail of the Tristan hotspot or the result of stress release caused by normal plate tectonics. The hotspot hypothesis further links the ridge to continental flood basalts, which are thought to be emplaced by an arriving plume. A temporal and spatial relation between the eruption of flood basalts and continental breakup gave rise to the idea that plumes are an important parameter for controlling continental breakups. In 2011 a large geophysical experiment was carried out at the landfall of Walvis Ridge in order to constrain its origin and the influence of the proposed plume head on the opening of the South Atlantic. Here, I present the crustal structure of the Walvis Ridge and the adjacent continental crust derived from two seismic refraction lines and gravity modelling. One line extends 480km across the ridge in approximately 600km distance to the coastline. The other profile covers the continental margin. It extends 430 km along the ridge axis and continues 290 km onshore. My results show that the Walvis Ridge consists of thickened oceanic crust with a thickness of 18-22 km. The crustal thickness decreases with increasing distance from the continental margin. Magmatic rocks cover a major fracture zone and pre-existing oceanic crust. Therefore, they must have been emplaced after the transform fault became inactive. Furthermore, the fracture zone is in about 100 km distance from the main ridge axis indicating that the ridge formed independent of the fracture zone. I conclude that normal stress release along a transform fault cannot account for the formation of the Walvis Ridge and support a hotspot origin. Abnormally high seismic velocities above 7.3km/s are observed in the lower crust at the continent-ocean transition. This high velocity lower crustal body (HVLCB) intrudes into the continental crust and terminates at the Kaoko fold belt. Similar HVLCB have been observed along the continental margin south of Walvis Ridge. The seaward termination of the HVLCB at Walvis Ridge is comparable to those. In contrast, the landward termination occurs 100 km further in land and is attributed to the presence of a mantle anomaly during the initial rift stage. Complementary seismic studies with a perpendicular line orientation show that this HVLCB is very limited in its width, even narrower than offshore. The influence of the hotspot was therefore very localized and the distribution of intrusive material resembles a narrow conduit rather than a broad plume head. I therefore question the presence of massive plume head during the the opening of the South Atlantic. From this it follows that the continental breakup has unlikely been initiated an arriving plume head.