Effects of high-speed train traffic characteristics on seismic interferometry

Abstract

SUMMARY Train traffic has been recognized as a powerful noise source for subsurface imaging and monitoring due to its properties of being strong, economic and repeatable. Compared with the traditional train, the high-speed train (HST) runs faster and generates vibrations that also manifest as periodical signals but with a more striking feature of sharp equidistant spectral lines in the frequency spectrum. Like the vibrations generated by the traditional train, the HST-induced vibrations can be utilized after the retrieval of interstation impulse responses using seismic interferometry. However, little attention has been paid to the characteristics of HST-induced vibrations during seismic interferometry. The aim of this study is to investigate the effects of train traffic characteristics on the seismic interferometry method. We first present an alternative derivation of seismic interferometry that is not based on the assumption of uncorrelation between sources at different locations. The derivation is valid for scalar waves in 3-D and for elastic surface waves. After analysing the effects of several HST-related factors, including HST moving direction, HST carriage number, rail structure, HST speed and the stacking number of HST records, on the seismic interferometry, we find that the seismic interferometry using HST traffic noise generates crosstalk that can be effectively attenuated by stacking the interferometric results of several HSTs with slightly different speeds. This finding has been validated by the seismic interferometry of both synthetic and field HST traffic noise. We further demonstrate that the retrieved surface waves can be used to estimate near-surface velocities.