Seismic polarization anisotropy beneath the central Tibetan Plateau

SKS and SKKS shear waves recorded on the INDEPTH III seismic array deployed in central Tibet during 1998-1999 have been analyzed for the direction and extent of seismic polarization anisotropy. The 400-km-long NNW trending ar ray extended south to north, from the central Lhasa terrane, across the Kar akoram-Jiali fault system and Banggong-Nujiang suture to the central Qiangt ang terrane. Substantial splitting with delay times from 1 to 2 s, and fast directions varying from E-W to NE-SW, was observed for stations in the Qia ngtang terrane and northernmost Lhasa terrane. No detectable splitting was observed for stations located farther south in the central Lhasa terrane. T he change in shear wave splitting characteristics occurs at 32 degreesN, ap proximately coincident with the transcurrent Karakoram-Jiali fault system b ut similar to 40 km south of the surface trace of the Banggong-Nujiang sutu re. This location is also near the southernmost edge of a region of high Sn attenuation and low upper mantle velocities found in previous studies. The transition between no measured splitting and strong anisotropy (2.2 s dela y time) is exceptionally sharp (less than or equal to 15 km), suggesting a large crustal contribution to the measured splitting. The E-W to NE-SW fast directions are broadly similar to the fast directions observed farther eas t along the Yadong-Golmud highway, suggesting that no large-scale change in anisotropic properties occurs in the east-west direction. However, in deta il, fast directions and delay times vary over lateral distances of similar to 100 km in both the N-S and E-W direction by as much as 40 degrees and 0. 5-1 s, respectively. The onset of measurable splitting at 32 degreesN most likely marks the northern limit of the underthrusting Indian lithosphere, w hich is characterized by negligible polarization anisotropy. Taken in conju nction with decades of geophysical and geological observations in Tibet, th e new anisotropy measurements are consistent with a model where hot and wea k upper mantle beneath northern Tibet is being squeezed and sheared between the advancing Indian lithosphere to the south and the Tsaidam and Tarim li thospheres to the north and west, resulting in eastward flow and possibly t hickening and subsequent detachment due to gravitational instability. In no rthern Tibet, crustal deformation clearly follows this large-scale deformat ion pattern.