SUMMARY Ambient seismic interferometry of distributed acoustic sensing (DAS) data acquired on optical fibre arrays is an increasingly common approach for subsurface investigation. The fixed infrastructure and low maintenance costs of commodity telecommunications fibre also supports cost-effective DAS-based seismic monitoring solutions over extended periods of time—especially when using repurposed telecommunication fibre infrastructure in urban settings. To investigate whether ambient waveform data acquired on such an urban DAS array are sensitive to seasonal subsurface variations, we present a case study using ‘semi-continuous’ DAS time-series data with hourly 150 s sampling windows that were acquired over a 10-month interval in the central business district of Perth, Australia. We apply a cross-coherence analysis to transform pre-processed ambient waveform data into sliding-window weekly interferometric virtual shot gathers (VSGs). We then use these data volumes to compute time-lapse velocity–dispersion panels, which we input to a multichannel analysis of surface waves (MASWs) to generate depth-averaged S-wave velocity estimates of the top 30 m ($V_{S_{30}}$ ). Our time-lapse analyses show that weekly stacked interferometric VSGs exhibit up to 5.8 per cent variations in observed surface wave traveltimes whereas the MASW inversion results capture up to 9.4 per cent variations in $V_{S_{30}}$ estimates between the winter and spring months. We note that these observations are inversely correlated with time-averaged rainfall patterns in the Perth Metro region and are likely attributable to the associated seasonal variations in near-surface groundwater content. Overall, our analysis suggests that semi-continuous ambient seismic monitoring on urban DAS fibre arrays is a computational tractable acquisition strategy that records data volumes useful for monitoring the seasonal variability of groundwater resources beneath urban centres as well as potentially other time-lapse subsurface behaviour occurring over calendar time.
Read full abstract