Publication Date:
2014-08-03
Description:
We present a robust technique for the analysis of shear wave splitting in layered anisotropic media by using converted shear phases. In (weakly) anisotropic media, the P -to- S converted phases ( Ps ) exhibit a distinct variation in arrival time (cosine moveout) as function of backazimuth. This variation can be exploited by time-shifting and stacking radial receiver functions to constrain a range of possible splitting parameters (i.e. the fast-polarization direction and the delay time) for an anisotropic layer. Then, the minimization of the transverse-component energy is used to select the pair of splitting parameters that best describes the anisotropic properties of the layer. This two-step approach stabilizes the inversion process and significantly reduces the time for computing the best splitting parameters. In multilayered anisotropic media, the distinct cosine moveout of the radial component is also observed for Ps phases from deeper discontinuities. The direct application of the method leads to apparent (or effective) splitting parameters for all anisotropic layers above the converting interface. Explicit expressions for the apparent splitting parameters are derived. The splitting parameters for the individual layers can be inferred by a layer-stripping approach, where the splitting effect due to shallower layers on converted phases from deeper discontinuities is successively corrected. In comparison to the approach by Liu & Niu ( 2012 ), the method represents a computationally efficient extension to multiple anisotropic layers. We further investigate the influences of noise, gaps in the azimuthal distribution of events, and interface dip on the splitting analysis. As an example, we apply the method to data from two stations of the Iranian National Seismic Network. At station GHIR the observed azimuthal variations of the Ps phase yield splitting parameters t = 0.26 s and f = 48° for a 48-km-thick crust. At station SHGR anisotropic effects on the Ps phases can be accounted for by two anisotropic crustal layers with a total thickness of 44 km. For the upper 22-km-thick layer we obtain a delay time t 1 = 0.25 s and a fast axis 1 = 35°. Values for the lower layer are t 2 = 0.19 s and 2 = –88°. The delay times correspond to anisotropies of about 4 and 3 per cent in the upper and lower layer, respectively. The method also provides a robust tool to separate the anisotropic properties of the crust and mantle and, thus, to isolate the anisotropic signature of mantle-deformation processes.
Keywords:
Seismology
Print ISSN:
0956-540X
Electronic ISSN:
1365-246X
Topics:
Geosciences
Published by
Oxford University Press
on behalf of
The Deutsche Geophysikalische Gesellschaft (DGG) and the Royal Astronomical Society (RAS).
Permalink