Subduction of the Philippine Sea slab in view of P-wave anisotropy

Three-dimensional (3D) structure of isotropic and anisotropic velocities in the Chugoku and Shikoku districts, southwest Japan, is retrieved from the first P-arrival times from local earthquakes by assuming P-wave azimuthal anisotropy with hexagonal-symmetry axis. The main result of isotropic velocity structure is that the Philippine Sea (PHS) slab is represented by a high-velocity body with the thickness of 20–30 km lying at the around depth of 40 km beneath the Shikoku and southern Chugoku districts. The PHS slab subduction appears to be interrupted by a small-scale low-velocity body under middle Chugoku and is not clear under north Chugoku. These features of 3D isotropic velocity structure are basically similar to those obtained in previous tomographic studies using recent seismic data. This result suggests that seismic anisotropy has no significant effect on estimation of isotropic velocity structure from the P-arrival time data. On the other hand, fast propagation directions of P-wave azimuthal anisotropy in the crust are basically in close agreement with fast polarization directions estimated from polarization anomaly of shear waves. Azimuthal anisotropy in the crust of Shikoku district is characterized by the fast propagation direction with the E–W orientation trend, which is nearly parallel to trajectory of maximum principal stress acting on the southwest Japan arc and strike directions of zonal geological structure. Crustal anisotropy in the Chugoku district is consistent with the orientation of the geological structure, and not with the maximum principal stress direction. These results indicate that the crustal anisotropy is attributable not only to alignment of the stress-induced cracks but also to lineament of geological structures. We could also point out the difference in the upper mantle anisotropy between the Shikoku and Chugoku districts; i.e., P-wave velocity tends to be fast in the N–S directions beneath Shikoku and southern Chugoku and in E–W directions under northern Chugoku. Since the N–S axis trend of P-wave anisotropy is highly correlated with high-velocity anomaly of PHS slab, it is regarded as slab anisotropy. The reason why the P-wave anisotropy shows the high propagation velocity in the E–W direction beneath northern Chugoku is discussed on the example of two cases in which the PHS slab is absent and present in the upper mantle beneath the region. Although we could not locate exactly the north leading edge of the PHS slab in view of anisotropic velocity structure, if the PHS slab is defined as a high-velocity body with fast propagation axes in the directions of N–S system, it turns out to exist at least beneath Shikoku to south Chugoku.