AbstractVehicle‐induced seismic waves, generated as vehicles traverse the ground surface, carry valuable information for imaging the underlying near‐surface structure. These waves propagate differently in the subsurface depending on soil properties at various spatial locations. By leveraging wave propagation characteristics, such as surface‐wave velocity and attenuation, this study presents a novel method for near‐surface monitoring. Our method employs passing vehicles as active, non‐dedicated seismic sources and leverages pre‐existing telecommunication fibers as large‐scale and cost‐effective roadside sensors empowered by Distributed Acoustic Sensing (DAS) technology. A specialized Kalman filter algorithm is integrated for automated DAS‐based traffic monitoring to accurately determine vehicles' location and speed. Then, our approach uniquely leverages vehicle trajectories to isolate space‐time windows containing high‐quality surface waves. With known vehicle (i.e., seismic source) locations, we can effectively mitigate artifacts associated with suboptimal distribution of sources in conventional ambient noise interferometry. Compared to ambient noise interferometry, our approach enables the synthesis of virtual shot gathers with a high signal‐to‐noise ratio and spatiotemporal resolution at reduced computational costs. We validate the effectiveness of our method using the Stanford DAS‐2 array, with a focus on capturing spatial heterogeneity and monitoring temporal variations in soil seismic properties during rainfall events. Specifically, in non‐built‐up areas, we observed an evident decrease in phase velocity and group velocity and an increase in attenuation due to the rainfall. Our findings illustrate our method's sensitivity and resolution in discerning variations across different spatial locations and demonstrate that our method is a promising advancement for high‐resolution near‐surface imaging in urban settings.
Read full abstract