Mantle kinematics beneath Southwestern Tibet inferred from direct S-wave splitting measurements

SUMMARY We have analysed direct S waves of teleseismic earthquakes to investigate seismic anisotropy parameters, that is, fast polarization direction (FPD or $\phi$) and splitting time delay (STD or $\delta t$) beneath southwestern Tibet (around the Karakoram Fault), that enable us to comprehend the upper mantle dynamics of the study region. To achieve this aim, we employ the Reference Station Technique, which is proven to be insensitive to source-side anisotropy; hence, it permits the use of teleseismic direct S-wave signals in splitting measurements. A total of 1624 high-quality direct S-wave splitting measurements were obtained from 145 earthquakes (M $\ge$5.5) within an epicentral distance of 30$^{\circ }$ to 90$^{\circ }$, recorded at 31 temporarily deployed seismic stations of the Y2 network. We have found STDs ranging from 1.1 to 1.8 s, indicating a significantly anisotropic upper mantle underneath the study region. Our splitting measurements reveal predominantly E-W FPDs in the western part of the study region, with a slight shift to the ENE-WSW direction in the eastern section. A comparison of our direct S-wave derived splitting measurements with prior SKS splitting measurements indicates a largely analogous pattern at most seismic stations. The seismic stations (WT04, WT05, WT11 and WT18), which previously lacked SKS-derived seismic anisotropy, are now complemented with new measurements with clear anisotropic signatures. Nearly E-W oriented FPDs that exhibit an oblique variation to the main strike of the southeastern segment of the Karakoram Fault (KKF) can be explained by the eastward movement of upper-mantle material beneath southwestern Tibet. The significant discrepancies between the orientation of FPDs and the strike direction of KKF imply that the fault is not a lithospheric-scale fault but rather is confined to crustal depths. Integrating surface deformation derived from geodetic measurements (e.g. global positioning system data) and plate motion vectors of the Indian and Eurasian plates with splitting parameters indicates that the deduced deformation patterns result from both lithospheric deformation and sub-lithospheric mantle dynamics. The FPDs exhibit a significant deviation from GPS data, signifying a decoupling of crust and upper-mantle materials beneath the study area. This suggests that mantle deformation in southwestern Tibet operates in a manner that is distinct from that of crustal deformation. Finally, our novel splitting measurements, enhanced by a greater number of direct S-wave data, provide new insights into the deformation of the upper mantle in the region, elucidating the mechanisms that have shaped the plateau over geologic millennia.