Deflection of mantle plume material by cratonic keels

Abstract Lithosphere that formed in Archaean and possibly early Proterozoic time is thicker, more buoyant, and geochemically distinct from lithosphere that formed after about 2.3 Ga. Mantle xenolith and seismic data indicate that some cratonic roots, or ‘keels’, extend to depths of c. 250 km, compared with normal continental lithosphere of thickness 150 km or less; yet many cratons have experienced uplift, dyking and kimberlite emplacement, suggesting interactions with hot, rising asthenosphere referred to as mantle plumes. Plumes supply additional heat to the base of the lithospheric plates, whose base can be heated and entrained in the flow (thermal erosion). How have these cratonic keels persisted despite their interactions with mantle plumes? The geometry of cratonic keels during their interactions with mantle plumes is a critical factor controlling keel preservation. To a laterally spreading plume head, cratonic keels appear as major obstacles, and the hot, buoyant plume material ponds beneath thinner lithosphere. Our model simulations show that deep keels deflect mantle plume material and that steep gradients at the lithosphere-asthenosphere boundary between Archaean keels and ‘normal’ lithosphere will focus flow, leading to localized adiabatic decompression melting. Plume processes can lead to a reduction in the breadth of a cratonic root where the plume rises beneath the craton, regardless of the initial breadth of the craton. Where the plume rises beneath a craton the hot plume material will spread laterally beneath the keel and attain thicknesses of tens of kilometres. This transfers heat to the base of the lithosphere and could generate small volumes of melt at considerable depth, depending on the composition of the lower lithosphere. We have used model simulations of plumes beneath Africa to predict the magnitude and direction of seismic anisotropy caused by lateral flow of hot plume material beneath and around a cratonic keel. The shear-wave splitting in our models is greatest at the edge of the cratonic keel, and its azimuth is parallel to the plume flow direction.