Item

Abstract

Distributed Acoustic Sensing (DAS) provides a new, non-invasive means for high resolution subsurface imaging in urban environments by repurposing existing telecommunication cables. However, urban fibre optic networks, which are designed around telecommunication needs, introduce unique challenges to passive seismic imaging including complicated array geometry and non-uniform seismic noise. A 25-km long DAS array was deployed along a telecommunication fibre that spans across metropolitan Melbourne, Australia for a duration of 3 months (December 2021 to March 2022). This dataset provides an ideal test case to address the challenges of using urban dark fibre and establish an effective workflow for ambient noise correlation with DAS recordings in urban settings. Traffic noise from vehicles and trains are the dominant signal at a frequency range of 1-30 Hz. Ambient noise correlation is performed using NoisePy, a high‐performance python tool specifically designed to deal with large data volume. Cross-correlation functions show clear surface wave dispersion up to 15 Hz. Acausal move-out times are also observed, indicating strong scatterers off the fibre array and varying ground-coupling conditions. The subsurface velocity model obtained from ambient noise correlation reveals strong structural variations at 10-m scale up to 1 km depth across Melbourne and shows good correspondence with the mapped geological boundaries including an 800 kyr basalt flow and Miocene marine and terrestrial sediments. The exciting results from the Melbourne experiment demonstrate that DAS can be used to build high-resolution subsurface models for metropolitan areas with high seismic hazard risk that may be poorly instrumented with inertial seismometers.