Three-dimensional Distribution of Azimuthal and Radial Anisotropy in the Japan Subduction

Abstract:

Seismic anisotropy has close relationships with past and present tectonic and dynamic processes. Therefore, detailed description of seismic anisotropy of subduction zones provides important information for our understanding of the subduction system.

The most common method of detecting anisotropy is the S-wave splitting measurement. However, conventional S-wave splitting analysis is not an appropriate way to investigate anisotropy in the mantle and slab because the technique has no vertical resolution. Thus, we have improved common traveltime tomography to estimate three-dimensional anisotropic structures of P-wave, assuming that the modeling space is composed of weakly anisotropic medium with a hexagonal symmetry about a horizontal axis (Ishise & Oda, 2005, JGR; Ishise & Oda, 2008, PEPI). Recently, we extended the anisotropic tomography for P-wave radial anisotropy with vertical hexagonal symmetry axis (Ishise & Kawakatsu, 2012 JpGU).

In this study, we expand the study area of our previous regional analyses of P-wave azimuthal and radial anisotropic tomography (Ishise & Oda, 2005; Ishise & Kawakatsu, 2012, JpGU; Ishise et al., 2012, SSJ) using Hi-net arrival time data and examine the subduction system around the Japan islands, where two trenches with different strike directions and plate junction are included.

Here are some of the remarkable results associated with the PAC slab and mantle structure. (1) N–S-trending fast axis of P-wave anisotropy is dominant in the PAC slab. (2) the mantle wedge shows trench-normal anisotropy across the trench-trench junction. (3) horizontal velocity (PH) tends to be faster than vertical velocity (PV) in the slab. (4) PV tends to be faster than PH in the mantle wedge.

The characteristics of the obtained azimuthal and radial anisotropy of the PAC slab and the mantle wedge qualitatively consistent with heterogeneous plate models (e.g., Furumura & Kennet, 2005) and numerical simulations of mantle flow (Morishige & Honda, 2011; 2013). In addition, the azimuthal anisotropy in the PAC slab that we obtained is subparallel to that in the PAC plate before subducting (e.g., Shimamura et al., 1983). Therefore, we suggest that the slab anisotropy is “frozen anisotropy”, which is attributed to the episode before subduction, and mantle wedge anisotropy reflects present dynamics.