Repeatability of Surface Wave Velocity Estimates from Distributed Acoustic Sensing (DAS) Data

The repeatability of surface wave velocity estimates from local ambient noise hinges on the stability of the crosscorrelation function for the receiver pair in the presence of a variable noise field, assuming near-surface soil properties are invariant over the duration of the surveys. Distributed acoustic sensing (DAS) data recorded on a linear trenched fiber optic cable sensor can accurately sample surface waves in a near continuous fashion (>1 kHz) with high spatial resolution (>1 receiver/m) and long range (10's of km). DAS recordings of ambient noise represent a unique means to explore the practical reliability of field-scale seismic property estimation from seismic interferometry. We test this hypothesis using continuous DAS field recordings from a shallow trench experiment near a busy road with diurnally-variable traffic patterns. Continuous records are processed using a modified ambient noise workflow consisting of receiver pair crosscorrelation, signal stacking, dispersion analysis, and a Monte Carlo search procedure to determine a best-fitting Vs model. The same processing flow is also applied to campaign data acquired with geophones to determine the repeatability benefit of trenched DAS deployment.