Continuous Near-surface Monitoring With Ambient Noise Collected by Distributed Acoustic Sensing
Eileen Martin, Stanford University
Distributed acoustic sensing (DAS) is a technology that repurposes a standard fiber optic cable as a low-cost seismic monitoring device. Permafrost thaw is tied to changes in seismic wave velocities, so we are developing a network to continuously monitor permafrost thaw under infrastructure. DAS can easily collect huge amounts of data continuously at low cost over a long distance, but that data is challenging to utilize due to a low SNR and insensitivity to waves at certain frequencies and incoming angles. Following a successful small-scale pilot test in 2014, a 640-meter trenched DAS array was installed along a road in a patchy permafrost zone in Fairbanks, Alaska. We used this array to monitor the speed of seismic waves in the near subsurface using only ambient seismic noise generated primarily by traffic. We show how traffic noise can generate coherent artifacts in the signals we extract from this kind of ambient noise and demonstrate strategies we developed to mitigate the effect of these artifacts on wave speed estimates. Ambient noise processing is notoriously slow and difficult to scale (typically quadratic with the number of sensors). To speed up that analysis, we have developed an embarrassingly parallel algorithm to calculate dispersion images that scales linearly with the number of sensors. These dispersion images are then used in dispersion-domain inversion for velocity structure in the near surface.
Abstract Author(s): E. Martin