Urban Subsurface Imaging with DAS: From Seismic Wavefields to Quasi-Static Deformation

Distributed Acoustic Sensing (DAS) is an emerging technology that transforms fiber-optic cables into dense arrays of vibration sensors, offering significant potential for subsurface exploration in urban environments. DAS enables the recording of broadband ground vibrations generated by human activity, including both high-frequency seismic wavefields (>1 Hz) and low-frequency quasi-static deformations (<1 Hz). However, effectively exploiting these signals and leveraging the dense spatial sampling of DAS in complex and highly heterogeneous urban subsurface environments, still remains a major challenge. In this study, we propose two novel approaches for local structural imaging based on seismic surface waves and quasi-static deformation. The first method uses high-frequency surface waves and a gradient-based amplitude estimation technique to achieve local structural imaging using only two DAS channels. Under the assumption of laterally heterogeneous JWKB theory, the ratio of the first-order temporal derivative to the spatial derivative of the surface-wave strain rate is used to estimate the local phase velocity. This approach allows adjacent DAS channels to resolve local one-dimensional velocity structures. The performance of this method is validated through numerical simulations and field experiments. The second method focuses on low-frequency quasi-static strain-rate signals induced by vehicle loading, enabling local structural imaging using a single DAS channel. A Markov Chain Monte Carlo (MCMC) inversion framework is used to investigate the depth sensitivity of quasi-static strain signals. Synthetic results indicate that the quasi-static strain field generated by a typical passenger vehicle can resolve subsurface structures at depths from 0 to10 m. Furthermore, field experiments conducted near a highway show that the derived two-dimensional velocity model is consistent with results obtained from conventional surface-wave inversion methods, confirming the robustness and applicability of the proposed approach. Looking ahead, the widespread deployment of urban fiber-optic communication networks provides an unprecedented opportunity to record broadband vibration signals from diverse sources, enabling large-scale urban subsurface imaging. These methods have promising applications in urban infrastructure design and hazard assessment.