LAUSR

Abstract The location and crustal structure of hotspot tracks in the South Atlantic reflect where melts related to sluggishly flowing plume material can reach the plate surface. This raises the paradox of how long-lived, age progressive hotspot tracks can arise in the absence of closely spaced, narrow mantle plumes. Here we show that young hotspot trails in the southern South Atlantic are located above bands of seismically slow material in the asthenosphere, which we interpret as channels of fast-flowing asthenosphere fed by a large scale plume upwelling from the African LLSVP. A broad region of seismically slow asthenosphere in the vicinity of Parana continental flood basalts may be indicative of a long-lived, large scale plume under the South American plate. We propose that hotspot tracks developed above fast flow channels in the asthenosphere that evolved between these large-scale plumes as they migrated apart with the African and South American plates, respectively. A progression from continental flood basalts to broad aseismic ridges (e.g., Walvis Ridge-Rio Grande Rise), to low volume intraplate hotspot tracks (e.g., Tristan-Gough; Discovery; Shona and Bouvet) reflects the interplay between tectonic setting and asthenosphere flow channels driven by waning pulsations from these diverging LLSVP plumes. We link the splitting of the Walvis Ridge into isotopically distinct, age-progressive intraplate sub-tracks about 72 Ma to the first sampling of material rising from the African LLSVP plume, perhaps as weak shallow long-lived plumes. Faster flowing asthenosphere enables melts related to LLSVP plumes to reach the plate surface via spreading and tectonic boundaries, and as low-volume intraplate hotspot (sub)tracks. The concept that asthenosphere flow channels and hotspot tracks evolve together between pulsating deep-seated plumes under Africa and South America suggests that LLSVPs might be a significant force in driving continental rifting and (absolute) plate motion.