Shear wave anisotropy in the Ryukyu subduction zone
Su, Chin-Tsai
Abstract
Shear(S) wave entering in anisotropic mantle and crust can split into two waves that travel with different speeds. They are polarized in mutually orthogonal directions if the anisotropic structure is approximately transverse-isotropic. [e.g., Savage, 2000]. The phenomena of S-wave splitting have been documented in numerous subduction zones and used to investigate seismic anisotropy caused by either the strain-induced preferential orientation of olivine aggregates in the mantle wedge or the frozen-in lithospheric anisotropy on both the back arc and seaward sides of subducting slabs or both [e.g., Fischer and Wiens, 1996; Fouch and Fischer, 1998].
In the offshore eastern Taiwan, the Philippine Sea plate obliquely subducting northwestward beneath the Eurasian plate along the Ryukyu trench formed the island arc and back-arc basin. To determine the location, strength, and orientation of seismic anisotropy in the Ryukyu subduction zone beneath Ishigaki, we evaluated shear wave splitting from both local and teleseismic shear waves recorded at ISG and IGK, YNG stations from the Japan OHP and F-net broadband networks during 1998-2003. We used the method of Silver and Chan (1991) that searches for the most singular covariance matrix to obtain the optimum polarization direction f of the leading shear phase and the delay δt between the fast and slow arrivals for each observable S arrival between 1998-2003. The best splitting parameters, f and δt, and their 95% confidence region are statistically estimated by the smaller eigenvalue of the covariance matrix indicative of the linearity of particle motion [Hung and Forsyth, 1999].
The result shows that on average the fast direction, f, from teleseismic S arrivals is NW48±13°, roughly parallel to the direction of absolute plate motion. The δt constrained by local S waves tends to increase linearly with the focal depth of events. The maximum delay from the deepest event of 256 km is about 1.65 s comparable to the average value 2.1±0.6 s, from teleseismic S waves, suggesting that observed seismic anisotropy is probably confined within the upper 300 km. The fast direction based on a group of local events from the north normal to the strike of the trench is NW41±14°, while the fast direction from the southwest events with backazimuths subparallel to the trench is NW61±15°. The splitting time dt constrained by local S waves approximately increases linearly with the propagation distance (L). The strength of seismic anisotropy estimated from the slope of d(dt )/dL is 2.3±0.4% assuming one layer of uniform anisotropy. Because the ray paths of local S waves are mostly within the subducting lithosphere, we infer that the observed shear wave splitting is probably caused by the frozen-in anisotropy that imprinted the fossil spreading direction while the Philippine plate was formed.
Reference
Fischer, K. M., and Wiens, D. A., 1996, The depth distribution of mantle anisotropy beneath the Tonga subduction zone, Earth Planet. Sci. Lett., 142, 253-260.
Fouch, M. J., and Fischer, K. M., 1998, Shear wave anisotropy in the Mariana subduction zone, Geophy. Res. Lett., 25, 1221-1224.
Hung, S.-H., and Forsyth, D. W., 1999, Anisotropy in the oceanic lithosphere from the study of local intraplate earthquakes on the west flank of the southern East Pacific Rise: Shear wave splitting and waveform modeling, J. Geophy. Res., 104, 10695-10717.
Silver, P. G., and Chan, W. W., 1991, Shear wave splitting and subcontinental mantle deformation, J. Geophy. Res., 96, 16429-16454.
Savage, M. K., 1999, Seismic anisotropy and mantle deformation: what have we learned from shear wave splitting?, Rev. Geophys., 37, 65-106