Sensitivity of Lower Mantle Seismic Anisotropy Beneath Subduction Zones to Mantle Viscosity Structure

Observations of seismic anisotropy can provide insights into the style of mantle dynamics near the 660km discontinuity. Wookey et al (2002) report up to 6 seconds of shear-wave splitting for rays generated by deep focus events from the Tonga subduction zone and recorded in Australia. Initial wave propagation modelling through different anisotropic slab regions, using ray theory, shows evidence that the anisotropic region is in the topmost lower mantle, with only a minimal contribution from above the 660km phase transition. Hence, the Wookey et al observations are direct evidence that the mid-mantle can exhibit anisotropy. In this study, we examine the effect mantle viscosity has on the generation of shear-wave splitting and thus, attempt to explain the observations from Tonga of top lower mantle seismic anisotropy. We use finite element (FE) modelling to calculate slab-induced models of fluid flow, total stress and deviatoric stress. Large deviatoric stresses (maximum values 40 MPa) are generated in the topmost lower mantle when the subducting slab encounters a viscosity increase at the 660km phase transition. These stresses could induce mineral alignment in a broad region (lateral wavelength 800km) in the topmost lower mantle below the slab. We model finite strain accumulated by a mantle parcel as it propagates through the FE fluid flow models. The computed strain ellipsoids align in a similar region as that of the deviatoric stresses. Strain fields are then mapped into seismic anisotropy. We ray trace from depths comparable to those observed in the Tonga subduction zone through our anisotropic models along similar travel paths as those observed by Wookey et al. We find that the magnitude of the generated shear-wave splitting is significantly dependant on the mantle viscosity structure. A subduction model with viscosity increases at 410km and 660km generates up to 6 seconds of shear-wave splitting from a source placed at 660km depth and rays traced to epicentral distances of 25\deg-65\deg. The magnitude of this shear-wave splitting is similar to the Wookey et al observations. A sensitivity analysis of shear-wave splitting to mantle viscosity structure requires a 10-fold viscosity increase at 410km and 660km depth and supports a viscosity structure similar to that proposed by Steinberger (2000). Observations of topmost lower mantle shear-wave splitting have the potential to constrain mantle viscosity structure.