Seismic anisotropy across the Australian continent

Continental anisotropy is typically explained by shear at the plate's base, with surface observations (i.e. shear wave splitting) mirroring the direction of plate motion and mantle flow. However, in some regions there is considerable mismatch between plate motion and fast polarization directions obtained from shear wave splitting (SWS), likely the result of regional dynamic processes or local lithospheric anisotropy. The Australian plate is the fastest moving continental plate at 6.75 cm/yr, yet SWS results across the continent do not align with plate motion, instead exhibiting lateral variations. Additionally, anisotropic tomography models of Australia suggest layered anisotropy in the lithosphere, implying the complexity of SWS results is tied to heterogeneous lithospheric structure from a long and complex deformational history.

To better understand the nature of seismic anisotropy, we will examine SWS and Ps receiver functions (RFs). Reconciling regional differences in fast directions begins by characterizing the depth-dependence and variation of Australian anisotropy with shear wave splitting. We have calculated frequency-dependent SWS results at 14 Australian using Splitlab (Wüstefeld et al., 2008) and following methodology from Eakin and Long (2013). SK(K)S phases were measured repeatedly over different frequency bands (6 in total) from epicentral distances 90 - 130°. There is an average of 35 non-null splits per station. Our results are generally consistent with previously published ones, but display a wide variation in fast direction and delay time, even from the same or similar backazimuths. Ps RFs were calculated at the same stations using multiple-taper spectral correlation (Park and Levin, 2000) and were depth-migrated using the local tomography model AuSREM (Kennett et al., 2012; Salmon et al., 2012). Events were selected from epicentral distances 30 - 95°, with an average of 204 waveforms per RF. Most stations show significant amounts of negative energy on the radial component at depths above the predicted LAB. Most transverse component RFs show small amplitudes varying as a function of backazimuth, suggesting complex anisotropy. To better constrain structure responsible for both phenomena, RFs were forward modeled using the code RaySum (Frederiksen and Bostock, 2000).