Deep structure of southern Tibet inferred from the dispersion of Rayleigh waves through a long-period seismic network

Abstract

Our understanding of the tectonic framework for the formation of Tibet and the collision of India with Eurasia¹ depends on our knowledge of the regional structure of the crust and upper mantle. Such basic information can be derived from seismic studies, but lateral inhomogeneities cause problems with interpretation of external observations, owing to the lack of stations inside a hardly accessible landmass. In 1982, as part of the French–Chinese cooperation programme, a network of long-period seismic stations was installed for the first time in southern Tibet itself with the object of obtaining local pure-path seismic data. Previously (with the exception of ref. 2) pure-path Tibet dispersion curves were extracted from mixed-path observations^3–9, which required assumptions about neighbouring regions and focal mechanisms. In this study, local Rayleigh phase-velocity dispersions from the network were inverted to yield S-wave velocity models. As the period range of the data extends beyond 120 s, the information so obtained constrains deep-seated structures down to 200 km. The crust is shown to be very thick (65–70 km), in accordance with previous estimates^2–9, and is divided into an upper part with a velocity of 3.2 km s⁻¹ and a lower part with a steep velocity gradient. The underlying mantle has a high velocity lid (V_s = 4.7 km s⁻¹), beneath which the structure resembles that of western Europe, with a relatively weak velocity minimum (compared with active tectonic regions) of 4.36 km s⁻¹ at a depth of 150 km.