Regional variations of the shear-wave polarization anisotropy in the crust and mantle wedge beneath the Tohoku district

We investigated the regional variations of the shear-wave polarization anisotropy in the upper crust, lower crust, and mantle wedge beneath the Tohoku district in the northeast Japan arc. The shear-wave splitting parameters (fast polarization direction (FPD) and split time) for the upper and lower crusts are estimated by shear-wave splitting analysis of the Ps phases, which are generated by the conversion of a P wave into an S wave at the Conrad and Moho discontinuity interfaces. The splitting parameters for the mantle wedge are determined by splitting analysis of the slab Ps phase converted at the oceanic crust on the subducting Pacific slab and the direct S waves from deep focus earthquakes located just below seismic stations. The Ps phases are identified on the P-wave receiver functions constructed from the teleseismic records. To accurately determine the splitting parameters for the lower crust from the Moho Ps phase, the Ps phase should be corrected for the shear-wave splitting effect in the upper crust. Similarly, to estimate the mantle wedge anisotropy, the direct S wave and slab Ps phase must be corrected for the splitting effect in the crust. The anisotropies of the crust and mantle wedge are estimated from the corrected Ps phases and direct S waves. The anisotropy of the upper crust exhibits a regional variation where FPD of the split shear wave is predominant in the N-S direction in the Pacific coast area and in the E-W direction in the rest of the Tohoku district. The split time is less than 0.2 s. The upper crustal anisotropy is attributed to the alignment of vertical cracks induced in the upper crust by tectonic stress. In the lower crust, FPD is predominant in the E-W direction with a split time similar to that in the upper crust, and the anisotropy of the lower crust is due to the lattice preferred orientation of rock-forming minerals. In the mantle wedge, FPD is predominant in the N-S direction (trench-parallel) in the fore-arc side of the volcanic front, but in the E-W direction (trench-perpendicular) in the back-arc side. The split time in the mantle wedge is similar to that of the upper crust. The most likely cause for the trench-perpendicular anisotropy in the back-arc mantle wedge is the lattice preferred orientation of A-type olivine along flows of the secondary mantle convection. For trench-parallel anisotropy in the fore-arc mantle wedge, we considered two possible causes: the preferred orientation of B-type olivine along the secondary mantle convection and that of dry olivine along the trench-parallel flow produced by the along-strike dip variation of the Pacific slab. However, we could not determine which one is the origin. (C) 2014 Elsevier B.V. All rights reserved.