Mantle plume capture, anchoring and outflow during ridge interaction

Abstract:

Geochemical and geophysical studies have shown that >40% of the world’s mantle plumes are currently interacting with the global ridge system and such interactions may continue for up to 180 Myr^[1]. At sites of plume-ridge interaction up to 1400 km of the spreading centre is influenced by dispersed plume material but there are few constraints on how and where the ridge-ward transfer of deep-sourced material occurs, and also how it is sustained over long time intervals. Galápagos is an archetypal example of an off-axis plume and sheds important light on these mechanisms. The Galápagos plume stem is located ~200 km south of the spreading axis and its head influences 1000 km of the ridge. Nevertheless, the site of enriched basalts, greatest crustal thickness and elevated topography on the ridge, together with active volcanism in the archipelago, correlate with a narrow zone (~150 km) of low-velocity, high-temperature mantle that connects the plume stem and ridge at depths of ~100 km^[2]. The enriched ridge basalts contain a greater amount of partially-dehydrated, recycled oceanic crust than basalts elsewhere on the spreading axis, or indeed basalts erupted in the region between the plume stem and ridge. The presence of these relatively volatile-rich ridge basalts requires flow of plume material below the peridotite solidus (i.e.>80 km).

We propose a 2-stage model for the development and sustainment of a confined zone of deep ridge-ward plume flow. This involves initial on-axis capture and establishment of a sub-ridge channel of plume flow. Subsequent anchoring of the plume stem to a contact point on the ridge during axis migration results in confined ridge-ward flow of plume material via a deep network of melt channels embedded in the normal spreading and advection of the plume head^[2]. Importantly, sub-ridge flow is maintained. The physical parameters and styles of mantle flow we have defined for Galápagos are less-well known at other sites of plume-ridge interactions, e.g. Tristan, Amsterdam. The observations require a more dynamically complex model than proposed by most studies, which rely on radial solid-state outflow of heterogeneous plume material to the ridge.

^[1] Whittaker JM et al (2015) Nature Geosci 10.1038/ngeo2437

^[2]Gibson SA, Geist DG & Richards MA (2015) Geochem Geophys Geosyst 10.1002/2015GC005723