Imaging the lithosphere beneath NE Tibet: teleseismic P and S body wave tomography incorporating surface wave starting models

The northeastern margin of the Tibetan Plateau, which includes the Qiangtang and Songpan-Ganzi terranes as well as the Kunlun Shan and the Qaidam Basin, continues to deform in response to the ongoing India-Eurasia collision. To test competing hypotheses concerning the mechanisms for this deformation, we assembled a high-quality data set of approximately 14 000 P- and 4000 S-wave arrival times from earthquakes at teleseismic distances from the International Deep Profiling of Tibet and the Himalaya, Phase IV broad-band seismometer deployments. We analyse these arrival times to determine tomographic images of P- and S-wave velocities in the upper mantle beneath this part of the plateau. To account for the effects of major heterogeneity in crustal and uppermost mantle wave velocities in Tibet, we use recent surface wave models to construct a starting model for our teleseismic body wave inversion. We compare the results from our model with those from simpler starting models, and find that while the reduction in residuals and results for deep structure are similar between models, the results for shallow structure are different. Checkerboard tests indicate that features of similar to 125 km length scale are reliably imaged throughout the study region. Using synthetic tests, we show that the best recovery is below similar to 300 km, and that broad variations in shallow structure can also be recovered. We also find that significant smearing can occur, especially at the edges of the model. We observe a shallow dipping seismically fast structure at depths of similar to 140-240 km, which dies out gradually between 33 degrees N and 35 degrees N. Based on the lateral continuity of this structure (from the surface waves) we interpret it as Indian lithosphere. Alternatively, the entire area could be thickened by pure shear, or the northern part could be an underthrust Lhasa Terrane lithospheric slab with only the southern part from India. We see a deep fast wave velocity anomaly (below 300 km), that is consistent with receiver function observations of a thickened transition zone and could be a fragment of oceanic lithosphere. In NE Tibet, it appears to be disconnected from faster wave velocities above (i.e. it is not downwelling or subducting here). Our models corroborate results of previous work which imaged a relatively slow wave velocity region below the Kunlun Shan and northern Songpan-Ganzi Terrane, which is difficult to reconcile with the hypothesis of southward-directed continental subduction at the northern margin. Wave velocities in the shallow mantle beneath the Qaidam Basin are faster than normal, and more so in the east than the west.