Comparing Earthquake Strain Records from Borehole Strainmeters and Fiber-Optic Array

Abstract:

Borehole strainmeter data has been recently shown to have the potential to complement seismometer and accelerometer data used in earthquake early warning applications, such as real-time magnitude and location estimation. However, due to the expense and relative difficulty associated with installing a borehole strainmeter, obtaining strain data on an extensive or dense scale for the purposes of earthquake seismology might be challenging. Fiber-optic arrays have been successfully used to provide strain measurements at a small fraction of the cost of installing a strainmeter or seismometer, and thousands of sensing points can be provided in a relatively small length of fiber. Additionally, optical fibers can be deployed offshore and in urban areas where installing conventional ground-motion sensors is inconvenient. In this work, I compare strain records of ten Northern California earthquakes (M3.3-4.7) from borehole strainmeters (BSM) that are part of the Network of the Americas and United States Geological Survey networks, and the Stanford Distributed Acoustic Sensing Array (SDASA). The SDASA is a figure-eight shaped array consisting of an optical fiber deployed in telecom pipes without any special installations or affixing to the ground. Due to the significant variability in channel outputs for ungrouted fibers (caused by poor mechanical coupling), I present an effective method that helps decide which channels in the DAS are representative, in terms of amplitudes and frequency content, of most of the channels in the array. I compare amplitude and frequency responses of the BSM and SDASA strain shown in the strain records of the ten California earthquakes and report their differences and similarities. An obvious challenge encountered with DAS data is the higher noise floor due to its location at the surface. However, I conclude that, despite the fact that DAS and BSM instruments in question are not co-located and despite of elevated DAS noise levels, the performances of DAS and BSM are generally comparable when capturing low and moderate-frequency signals, which are crucial in estimating magnitudes of larger earthquakes.