Effects of crystal preferred orientation on upper-mantle flow near plate boundaries: rheologic feedbacks and seismic anisotropy

Insight into upper-mantle processes can be gained by linking flow-induced mineral alignment to regional deformation and seismic anisotropy patterns. Through a series of linked micro-macro scale numerical experiments, we explore the rheologic effects of crystal preferred orientation (CPO) and evaluate the magnitude of possible impacts on the pattern of flow and associated seismic signals for mantle that includes a cooling, thickening young oceanic lithosphere. The CPO and associated anisotropic rheology, computed by a micromechanical polycrystal model, are coupled with a large scale flowmodel (Eulerian Finite Element method) via a local viscosity tensor field, which quantifies the stress: strain rate response of a textured polycrystal. CPO is computed along streamlines throughout the model space and the corresponding viscosity tensor field at each element defines the local properties for the next iteration of the flow field. Stable flow and CPO distributions were obtained after several iterations for the two dislocation glide cases tested: linear and nonlinear stress: strain rate polycrystal behaviour. The textured olivine polycrystals are found to have anisotropic viscosity tensors in a significant portion of the model space. This directional dependence in strength impacts the pattern of upper-mantle flow. For background asthenosphere viscosity of similar to 10(20) Pa s and a rigid lithosphere, the modification of the corner flow pattern is not drastic but the change could have geologic implications. Feedback in the development of CPO occurs, particularly in the region immediately below the base of the lithosphere. Stronger fabric is predicted below the flanks of a spreading centre for fully coupled, power-law polycrystals than was determined using prior linear, intermediate coupling polycrystal models. The predicted SKS splitting is modestly different (similar to 0.5 s) between the intermediate and fully coupled cases for oceanic plates less than 20 Myr old. The magnitude of azimuthal anisotropy for surface waves, on the other hand, is predicted to be twice as large for fully coupled power-law flow/polycrystals than for linear, intermediate coupled flow/polycrystal models.