Influence of Strong Directional Sources in Ambient Seismic Imaging using Traffic-induced Noise

Abstract

Continuously moving seismic sources, such as vehicles and train cars, play a crucial role as passive sources for non-destructive exploration in urban areas. Seismic interferometry is commonly employed in field data acquisition and processing, using linear arrays. However, the simultaneous recording of high energy noise, excited by buildings and factories, cannot be overlooked. We simulate moving-source seismic records with strong noises using orthogonal arrays. When strong noise originates from specific angles, the energy-phase velocity curves of arrays in different directions significantly diverge. We demonstrate that an L-shaped array, formed by orthogonal arrays, can effectively mitigate this effect, yielding more consistent results. Nonetheless, spatial constraints often preclude the deployment of L-shaped arrays. To mitigate this issue, we propose a new parallel observation system in this paper. Field tests conducted on a main road validate that the new array is comparable to the L-shaped array in terms of dispersion extraction. Similar to synthetic data, the phase velocity extracted from the linear array field data is found to be unreliable. Drilling data aligns well with the inversion results of the parallel observation system. Given the challenges of urban traffic-induced signal acquisition, deploying multi-azimuth arrays to minimize noise impact is essential. Considering spatial limitations, the convenient parallel observation system emerges as a good choice.