Extraction of high-frequency surface waves on the Xishancun landslide from ambient seismic noises

There dimensional structures of landslides provide key information to landslide treatment, disaster prevention, and risk mitigation, as well as the understanding of landslide dynamics, while high-frequency surface wave tomography has been proven to be an effective method to image shallow structures. In this study, we attempt to obtain high-frequency surface waves for 38 stations on the Xishancun landslide in southwestern China. We extract empirical Green's functions from three-component continuous ambient seismic noises recorded on December 1st, 2016. First, we test several parameters that may yield better noise correlation functions (NCFs) , including time-segment length, data filtering, and topography effect. After stacking NCFs from one-day noises with a segment length of 1200 s, we obtain clear Love and Rayleigh waves in the frequency range of 1 similar to 5 Hz with apparent velocities at about 400 m . s(-1) and 700 m . s(-1), respectively. Love waves generally have higher signal-to-noise ratios than Rayleigh waves. Beamforming of Love waves suggests that the seismic noises mainly come from the sediment transport of the river to the south toe of the landslide. The obtained surface waves and location of seismic noises provide a valuable input for 3D tomography of the shallow structures of the Xishancun landslide, as well as their temporal variations.