Lower Mantle Seismic Anisotropy Generated by Subduction Body Force Stresses Beneath the 660km Phase Transition.

Stuart Nippress1, Nick Kusznir1, Michael Kendall2

1Department of Earth Sciences, University of Liverpool, 4 Brownlow Street, Liverpool, L69 3GP, UK.

2School of Earth Sciences, University of Leeds, Leeds, LS2 9JT, UK.

nippress@liverpool.ac.uk

Observations of seismic anisotropy can provide insights into the style of mantle dynamics near the 660km discontinuity. Wookey et al. (2002) report up to 7 seconds of shear wave splitting for rays generated by deep focus events from the Tonga subduction zone and recorded in Australia. The results suggest a transversely isotropic symmetry with the symmetry axis in the vertical plane, perpendicular to the ray direction. Thus, for horizontally travelling waves this would imply horizontally polarised shear waves (SH lead SV). They show that a topmost lower mantle model with anisotropy between 660-900km could produce theoretical shear wave splitting similar to that observed. Therefore, the seismic anisotropy observed by Wookey et al., can be explained by an anisotropic region between 660-900km, with only a minimal contribution from above the 660km phase transition.

Kusznir (2000) has shown that large deviatoric stresses (maximum values ~ 40 MPa) are generated in the topmost lower mantle when the subducting slab encounters an increase in viscosity at the 660km phase transition. These stresses may induce mineral alignment in a broad region (lateral wavelength » 800km) in the topmost lower mantle below the slab. Perovskite may therefore be aligned with a rotated symmetry axis conformal to the shape of this region of high deviatoric stress. Aligned Perovskite rotated more than 30° predicts SH-waves faster than SV-waves for horizontally travelling S-waves. The goal of this study is to test whether subduction generated deviatoric stresses within the uppermost lower mantle can create the necessary anisotropy to explain the observed shear wave splitting in the topmost lower mantle in the Tonga region.

We use finite element (FE) modelling to calculate slab-induced models of fluid flow, total stress and deviatoric stress. A simple 2D subduction zone model with a prescribed viscosity structure and slab density is used. Calculated fluid flow vectors and the formulation of Malvern (1969) are used to determine the finite strain accumulated by a mantle parcel as it traverses a streamline. After 40My of subduction (» the age of the Tonga subduction zone), the strain ellipsoids align in a similar region (wavelength » 800km and depth » 660-900km) as that of the deviatoric stresses. The strain fields are then mapped into seismic anisotropy. We initially assume that the anisotropy has hexagonal symmetry, with a symmetry axis aligned with the major axis of the finite-strain ellipse. The magnitude of the anisotropy is scaled by the degree of finite-strain ellipticity. Theoretical shear wave splitting is then calculated by ray tracing through the seismic anisotropy model. Modest amounts of anisotropy can explain the seismic observations.