Seismic anisotropy indicators in Western Tibet: Shear wave splitting and receiver function analysis

Abstract

Using recently collected data from western Tibet we find significant variation in the strength, vertical distribution and attributes of seismic wave speed anisotropy, constrained through a joint application of teleseismic shear wave splitting techniques and a study of P-S mode-converted waves (receiver functions). We find that the crust of Tibet is characterized by anisotropy on the order of 5%–15% concentrated in layers 10–20 km in thickness, and with relatively steep (30°–45° from the vertical) slow symmetry axes of anisotropy. These layers contribute no more than 0.3 s to the birefringence in teleseismic shear waves, significantly smaller than splitting in many of the observations, and much smaller than birefringence predicted by models developed through group inversions of shear-wave recordings. Consequently, we interpret models constrained with shear-wave observations in terms of structures in the upper mantle. Near the Altyn–Tagh fault our data favor a two-layer model, with the upper layer fast polarization approximately aligned with the strike of the fault. Near the Karakorum fault our data are well fit with a single layer of relatively modest (~ 0.5 s delay) anisotropy. Fast polarization in this layer is ~ 60°NE, similar to that of the lower layer in the model for the Altyn Tagh fault site. Assuming that layers of similar anisotropic properties at these two sites reflect a common cause, our finding favors a scenario where Indian lithosphere under-thrusts a significant fraction of the plateau. Data from a site at the southern edge of the Tarim basin appear to be inconsistent with a common model of seismic anisotropy distribution. We suspect that thick sediments underlying the site significantly distort observed waveforms. Our ability to resolve features of anisotropic structure in the crust and the upper mantle of western Tibet is limited by the small amount of data collected in a 6 month observing period. We stress the importance of future teleseismic data collection in Tibet over extended periods, to insure better directional distribution of observed seismic sources.