Olivine anisotropy suggests Gutenberg discontinuity is not the base of the lithosphere.

Tectonic plates are a key feature of Earth's structure, and their behavior and dynamics are fundamental drivers in a wide range of large-scale processes. The operation of plate tectonics, in general, depends intimately on the manner in which lithospheric plates couple to the convecting interior. Current debate centers on whether the transition from rigid lithosphere to flowing asthenosphere relates to increases in temperature or to changes in composition such as the presence of a small amount of melt or an increase in water content below a specified depth. Thus, the manner in which the rigid lithosphere couples to the flowing asthenosphere is currently unclear. Here we present results from laboratory-based torsion experiments on olivine aggregates with and without melt, yielding an improved database describing the crystallographic alignment of olivine grains. We combine this database with a flow model for oceanic upper mantle to predict the structure of the seismic anisotropy beneath ocean basins. Agreement between our model and seismological observations supports the view that the base of the lithosphere is thermally controlled. This model additionally supports the idea that discontinuities in velocity and anisotropy, often assumed to be the base of the lithosphere, are, instead, intralithospheric features reflecting a compositional boundary established at midocean ridges, not a rheological boundary.