Distributed acoustic sensing for shallow structure imaging using mechanical noise: A case study in Guangzhou, China

Abstract

Distributed Acoustic Sensing (DAS) is a novel seismic detection technology that transforms fiber-optic cables into densely sampled seismic stations, showing immense potential in both urban seismology and the oil and gas industry. We deployed a ~ 182-m-long single-core fiber-optic cable in a campus under construction to record human activity and demonstrated the application for seismic imaging. The spectrogram analysis revealed significant differences in ambient noise between the daytime and nighttime at the construction site. During the daytime, the noise energy exhibited dominant frequencies ranging from 2 to 11 Hz, approximately one order of magnitude higher than that at night, which had frequencies ranging from 2 to 5 Hz. Beamforming analysis demonstrated that noise sources exhibited distinct directional characteristics and propagation velocities during daytime and nighttime periods. We extracted the cross-correlation functions between DAS channels from the mechanical noise generated by a large excavator and enhanced signal-to-noise ratio by using the phase-weighted stacking method. The multichannel analysis of surface waves (MASW) method was then employed to obtain 2D near-surface shear-wave velocity. Based on five 60-m-deep boreholes drilled along the DAS array, we established the correlation between shear-wave velocity information and shallow structure. We obtained the unbiased cross-correlation function in the presence of continuous mechanical noise along the DAS array direction, demonstrating the importance of site noise source distribution analysis in ambient noise tomography. Our research is informative for research related to the construction of accurate, high-resolution near-surface shear wave velocity models.