4-D permafrost thaw observations from ambient road traffic noise and a very dense distributed fiber optic sensing array

How does frozen soil thaw? The answer to this question affects hydrology, ecology, climate, and human infrastructure. We are using the local ambient noise field from a road recorded on a distributed fiber optic acoustic sensing (DAS) array to monitor the evolution in seismic parameters related to the top-down permafrost thaw process in the upper 10 m. Our field experiment demonstrates the advantages of "Large N" ambient noise studies using DAS, particularly to probe near surface critical zone dynamics. Over 60 days beginning in August 2016, we made continuous seismic recordings with a >4000 channel trenched fiber optic DAS dataset above a controlled permafrost warming demonstration experiment in Fairbanks, AK. The warming experiment accelerated the state of permafrost degradation by approximately two decades in a small 15 m x 20 m area, deepening the permafrost table from 4 m to 5.5 m. Continuous seismic DAS recording of high frequency surface waves (5-30 Hz) generated by vehicles traveling along a nearby road enables our investigation of hypothesized shear wave speed and attenuation changes, which lab measurements suggest may result from decreasing shear modulus and increasing saturation. We develop daily auto- and crosscorrelation function estimates using combinations of horizontal inline, collinear, and crossline DAS sensor orientations and vertical component geophone data, and then invert for maps of Love and Rayleigh wave speed that are sensitive to the upper 30 m. Many issues related to the accuracy, stability, and repeatability of the recovered empirical Green's tensor, as well as the sensitivity of the DAS sensor network will be considered.