The structure and dynamics of the lithosphere beneath Tibet from seismic surface-wave analysis

Abstract

Despite the numerous studies undertaken to investigate the underlying structures beneath the Tibetan Plateau, fundamental questions about the mechanism of lithospheric convergence between India and Asia (establishing if convergence is from: underthrusting, north/south subduction, lithospheric thickening, or convective removal), and, the internal lithospheric dynamics within Tibet (rigid block or continuous deformation), persist until now. This study aims to determine which of the different mechanisms are at work beneath Tibet, keeping in mind that elements of different mechanisms may be present beneath different parts of the plateau. The data set is broad-band Rayleigh- and Love-wave phase-velocities measured from pairs of station across the plateau. A series of rigorous analyses is performed on the data: (1) direct comparison of inter-station dispersion, (2) inversion for 1-D radially anisotropic S-velocity models, (3) inversion for 1-D azimuthally anisotropic S-velocity models (4) modelling of regional averages, and (5) tomographic models. All of these techniques yielded consistent results regarding shear-velocity structure and anisotropic properties of the crust and upper mantle beneath Tibet. The Tibetan crust has expected low shear velocities, however, with strong north-south variations across the plateau. Mid-crustal LVZ (and inferences of melt and low viscosity) and high conductivity are found across large areas of northern Tibet. The consistent lateral distribution of the data and the coherent pattern of anisotropy within the northern regions, strongly suggest that crust is homogeneous and that deformation is diffused across large areas. Strong radial anisotrophy is observed beneath western Tibet and Yunnan, both regions experiencing extension (flattening), whereas north-eastern Tibet has very weak radial anisotrophy, strike-slip fault mechanisms, and no extension. The ongoing crustal thinning in the west probably causes the anisotropic mica crystals to become near-horizontally oriented (Shapiro et al., 2004), whereas in the east, horizontal pervasive flow may align micas in the vertical plane resulting in weak or absent radial anisotropy. The flow direction, derived from crustal azimuthal anisotropy, show W -E and NW -SE fast directions in central and eastern Tibet, respectively. Special focus is given to northeastern Tibet, where the inferred fast directions are aligned southeast rather than northeast, as would be expected from an elevation-gradient induced flow (Clark and Royden, 2000). The fast azimuths are parallel to the extensional component of the current strain rate across Tibet, strongly suggesting similar deformation through the entire crust. Despite the mid-crust’s greater susceptibility to deformation and flow, the correlation of azimuthal anisotropy with surface strain indicates that the mid-crust still holds some degree of coupling with the adjacent layers. The close agreement of anisotropy and extension component of strain with the traces of sutures implies that the dominant deformation mechanism within the plateau has not changed since initiation of continental collision and is still governed by the northward push of India. The upper 75 km of the mantle beneath Tibet is made up of an Indian lithosphere in the west and southwest and a Tibetan lithosphere and asthenosphere elsewhere. Strong, cold, cratonic Indian lithosphere underthrusts southwestern Tibet (up to the BNS at 85°E), and warm Tibetan lithosphere and asthenosphere lay further north, up to the Kunlun Fault. The Tibetan lithosphere and asthenosphere have low-average S-velocities, indicative of warmer temperatures. Although the finer structure of the Tibetan lithospheric mantle remains hard to resolve using surface waves alone, a thick layer of low-average Vs in the uppermost mantle is difficult to explain from high temperatures generated by crustal radioactivity and reconcile with the presence or formation of a thick cratonic lithosphere at depths down to 200 km (McKenzie and Priestley, 2008). Surface-wave data can fit a series of other seismic observations such as Sn, Pn, and a shallow LAB discontinuity, which together add support for a thin Tibetan lithosphere underlain by an asthenosphere. The dynamics of the asthenosphere is revealed by azimuthal anisotrophy beneath central Tibet, characterised by fast SSW-NNE direction. The amplitude of the anisotropy increases from south to north and is parallel to the direction of India’s plate motion, suggesting that asthenospheric flow is pushed outward by India’s northward subduction. Cold Indian lithosphere subducts beneath the Tibetan asthenosphere under the central and eastern plateau. The lithospheric convergence mechanisms varies from west to east; steep-angle subduction of India beneath west-central Tibet and shallow-angle subduction of India in eastern Tibet, with the subducting Indian lithosphere reaching as far north as northern Qiangtang-Songpan-Ganzi Terrane.