Noise Characteristics of One Year of DAS Monitoring Data From Penn State Foresee Array

Session: Advances in Seismic Interferometry: Theory, Computation and Applications
Type: Oral
Date: 4/30/2020
Time: 11:15 AM
Room: 230 + 235
Description: 

Near surface time-lapse seismic monitoring is potentially useful for tracking and interpreting changes due to temperature, hydrology or natural hazards. Massive seismic arrays are hard to maintain over months to years, but Distributed Acoustic Sensing (DAS), enables affordable long-term monitoring with high resolution via a new technique converting unused telecommunication optical fiber to sensor arrays. DAS data has been continuously recorded since April 2019 along 4.2 km of underground telecommunication fibers under the Penn State campus. This experiment is the Fiber Optic foR Environmental SEnsEing (FORESEE) project. We will present results of spatiotemporal noise characterization of a year of recordings and noise interferometry. We observed an S-wave polarity-reversal of Peru M8.0 earthquake at corners of the FORESEE array. By comparing with synthetic modeling, we investigate directional sensitivity of the array. We investigate the noise source distribution in this urban area by spectral analysis and beamforming throughout the year. Noise sources vary across the array, primarily due to nearby anthropogenic activities. At most channels the noise field is dominated by high frequency traffic noise 10-20 Hz, particularly during work hours. Strong energy in narrow frequency bands is recorded on the edge of campus, most likely from nearby energy facilities and agricultural activities. Construction site noise can be sensed up to 40 Hz. Multiple of noise distribution patterns are illuminated by beamforming during rush hours, daytime and nighttime. Lower frequency (1-9 Hz) noises are more evenly distributed than higher frequency traffic noise. Due to the wide variety of noise sources, we focus on selecting proper noise sources and subsets of the array to estimate the Green’s function between stations. We test the validity of the stationary cross-correlation assumption over frequency ranges and retrieve long-term subsurface velocity changes with minimal effect due to the ambient noise field.

Presenting Author: Junzhu Shen

Authors