Earthquake recording at the Stanford DAS Array with fibers in existing telecomm conduits

The Stanford Distributed Acoustic Sensing Array (SDASA-1) has been continuously recording seismic data since September 2016 on 2.5 km of single mode fiber optics in existing telecommunications conduits under Stanford's campus. The array is figure-eight shaped and roughly 600 m along its widest side with a channel spacing of roughly 8 m. This array is easy to maintain and is nonintrusive, making it well suited to urban environments, but it sacrifices some cable-to-ground coupling compared to more traditional seismometers. We have been testing its utility for earthquake recording, active seismic, and ambient noise interferometry. This talk will focus on earthquake observations. We will show comparisons between the strain rates measured throughout the DAS array and the particle velocities measured at the nearby Jasper Ridge Seismic Station (JRSC). In some of these events, we will point out directionality features specific to DAS that can require slight modifications in data processing. We also compare repeatability of DAS and JRSC recordings of blasts from a nearby quarry. Using existing earthquake databases, we have created a small catalog of DAS earthquake observations by pulling records of over 700 Northern California events spanning Sep. 2016 to Jul. 2017 from both the DAS data and JRSC. On these events we have tested common array methods for earthquake detection and location including beamforming and STA/LTA analysis in time and frequency. We have analyzed these events to approximate thresholds on what distances and magnitudes are clearly detectible by the DAS array. Further analysis should be done on detectability with methods tailored to small events (for example, template matching). In creating this catalog, we have developed open source software available for free download that can manage large sets of continuous seismic data files (both existing files, and files as they stream in). This software can both interface with existing earthquake networks, and efficiently extract earthquake recordings from many continuous recordings saved on the users machines.