Modeling pressure-dependent seismic anisotropy in the lower mantle reveals anisotropic discontinuity at 1000 km

There is growing evidence, both from a modelling perspective and seismic observations, that seismic anisotropy in the lower mantle is localized around penetrating slabs where large straining is anticipated. It is believed that the high stresses experienced near the slab activate dislocation creep mechanisms that drive the crystallographic preferred orientation (CPO) of bridgmanite aggregates. Still, deformation mechanisms in bridgmanite remain enigmatic. In recent years, deformation experiments in bridgmanite subjected to mantle temperatures and pressures suggest that its microstructures evolve with pressure, providing another perspective on the debated structure and deformation in the lower mantle. Using this information, we develop a numerical technique that calculates pressure-dependent large-scale seismic anisotropy in a pyrolitic mantle with variable velocity gradients. As a first test, we use the method to predict seismic anisotropy by calculating anisotropic reflection coefficients of underside reflections off a depth corresponding to 50 GPa where pressure-induced slip transitions in bridgmanite are expected. For this, we consider two simple deformation styles: (1) uni-axial compression, akin to vertically penetrating slabs, and (2) simple shear associated with corner-type flows. Finally, we demonstrate a multiscale approach that calculates large-scale seismic anisotropy from a fully time-dependent thermo-chemical model of free subduction with latent heating and phase transitions. The result is a long-wavelength equivalent azimuthal and radial anisotropy maps that are actually comparable to a seismic tomography model. We demonstrate how such an approach can create discontinuities in anisotropy at ~1000 km and provide insights as to how it relates to the heterogeneous distribution of the 1000-km discontinuity.