Crustal Structure of China from Deep Seismic Sounding Profiles

More than 36,000 km of Deep Seismic Sounding (DSS) profiles have been acquired in China since 1958. However, the results of these profiles are not well known in the West. Here, we summarize the crustal structure of China with a new contour map of crustal thickness, nine representative crustal columns, and maps showing profile locations, average crustal velocity, and Pn velocity. The most remarkable aspect of the crustal structure of China is the well known 70±km thick crust of the Tibetan Plateau. The average crustal velocity of China ranges from 6.15 to 6.45 km/s, indicating a felsic-to-intermediate bulk crustal composition. Upper mantle (Pn) velocities are 8.0 ± 0.2 km/s, equal to the global continental average.

We interpret these results in terms of the most recent thermo-tectonic events that have modified the crust. In much of eastern China, Cenozoic crustal extension has produced a thin crust with a low average crustal velocity, similar to Western Europe and the Basin and Range Province, Western USA. In western China, Mesozoic and Cenozoic arc-continent and continent-continent collisions have led to crustal growth and thickening. A high velocity (7.0-7.4 km/s) lower crustal layer has been reported in western China only beneath the southernmost Tibetan Plateau. We identify this high velocity layer as the cold lower crust of the subducting Indian plate. As the Indian crust is injected northward into the Tibetan lower crust, it heats and assimilates by partial melting, a process that results in a reduction in the seismic velocity of the lower crust in the central and northern Tibetan Plateau.

Deep Seismic Sounding (DSS) profile

Western China is a showcase of complex geological and geophysical features, including sedimentary basins, regimes of continental collisional tectonics, and the thickest crust found on earth.  To be able to accurately monitor western China for nuclear explosions we must first understand as much as possible about the crustal structure there.  Here, we present results of a seismic refraction profile across western China (Wang et al., 2001).  Seismic energy for this profile was provided by twelve chemical explosive shots fired in boreholes. The charge size ranged from 1500 to 4000 kg, sufficient to provide clear first arrivals to a maximum distance of 300 km. The distance between shotpoints ranged from 63 to 205 km, and the interval between portable seismographs was between 2 and 4 km.  The profile was recorded along existing roads, and provided nearly straight profile segments.

During the experiments both P- and S-wave data were acquired, even for data recorded by single-component geophones. The reflection from the Moho was especially strong, and this made it possible for us to derive the crustal composition using the crustal Poisson’s ratio or Vp/Vs ratio which could be obtained from the crustal P- and S-wave velocity structures. In the correlation of phases, reduction velocities of 6.0 km/s and 3.46 km/s were used for P- and S-waves, respectively. The time scales used for S-waves were multiplied by a factor of 0.58 in the S-wave record section so they would match the P-wave arrival times. Because digital filtering introduces a slight time shift, the unfiltered P-wave data was used for phase correlation and travel time picking. In order to improve the signal-to-noise ratio for phase correlation, the S-wave data was filtered with a bandpass of 0-6 Hz.

Based on the phase correlation, the first arrivals of the P_g phase were used to invert the upper crustal velocity structure using the finite-difference tomographic method of Hole (1992). The reflection phases were used to determine the one-dimensional crustal velocity structure using the X²-T² method (Giese, Prodehl and Stein, 1976) and the Reflectivity method (Fuchs and Muller, 1971). With one-dimensional crustal velocity structures from each shotpoint, a crustal P-wave velocity structure was established using a 2-D dynamic raytracing program to model both kinetic and dynamic features of the observed seismic wave field (Cerveny, Molotkov, and Psencik, 1977; Cerveny and Psencik, 1984). The different phases on the record sections were all appropriately fitted for travel times and amplitudes.

The transect shows three-layer stratification with P-wave velocities of 6.0-6.3 km/s, 6.3-6.6 km/s, and 6.9-7.0 km/s.  Upper mantle (Pn) velocities of 7.7-7.8 km/s were found.

The accuracy of the final model is dependent on a number of factors, including the shotpoint interval, receiver density, and thickness of sediments in the shallow crust. Model accuracy primarily depends on the correct identification of the various phases and the density of rays penetrating a particular volume of the model, however. Perturbation of the models has shown that, depending on the uniformity of structure and the density of rays, the resolution of velocity and depth to interface may be accepted as better than 2% and 5%, respectively. By the relationship between the Poisson’s ratio and the Vp/Vs ratio, the Poisson’s ratio is determined to within an uncertainty of less than 2%.