Depth localized azimuthal anisotropy fromSKSandPreceiver functions: The Tien Shan

Abstract

SUMMARY Shear wave splitting in the seismic SKS phase provides a unique possibility to judge on deformations at depths inaccessible for direct observations. Fast S wave polarization direction in collisional belts is often parallel to the trend of the belt, although deformations of the mantle lithosphere in low-angle thrusts would lead to the fast polarization direction normal to the trend of the belt. These considerations suggested that the upper mantle in collisional belts is decoupled from the crust. However, SKS technique is notable by a poor depth resolution, and usually it assumes that the fast polarization direction is the same at any depth, which is hard to justify. Here, to investigate depth dependent azimuthal anisotropy in the mantle, we invert jointly P receiver functions and SKS particle motions at a number of seismograph stations. The technique involves azimuthal filtering of the receiver functions and provides a criterion to discriminate between the effects of azimuthal anisotropy and lateral heterogeneity of isotropic medium. A search for the optimum models is conducted with a technique similar to simulated annealing. Testing with synthetics demonstrates that this approach is robust. The results for 10 seismograph stations in the Tien Shan, the world’s most active intracontinental collisional belt in Central Asia, reveal a pronounced change in the patterns of azimuthal anisotropy at a depth around 100 km. In the mantle lithosphere (at depths less than 100 km), anisotropy is relatively weak and fast wave polarization direction varies laterally in a broad range. This layer is not necessarily decoupled from the crust: its anisotropy can be a combined effect of present day thrusting and of deformations of the geologic past. In the lower layer (asthenosphere) the average azimuth of fast wave polarization is close to the trend of the belt, whereas magnitude of S wave anisotropy is stable and large (between 5 and 6 per cent). This anisotropy is a likely result of recent uniaxial shortening at right angle to the trend of the belt. At some stations the data require anisotropy in the crust. There is no evidence for anisotropy at depths exceeding 150‐250 km.