Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Detecting Hidden Low-Frequency Earthquakes in Southern Vancouver Island with Deep-Learning

Jiun-Ting Lin, Ananda M. Thomas, Diego Melgar, Loïc Bachelot, & Douglas R. Toomey

Submitted September 10, 2023, SCEC Contribution #12980, 2023 SCEC Annual Meeting Poster #009

Low-frequency earthquakes (LFEs) are earthquakes depleted in high-frequency content, and have been found in various seismogenic zones, including Parkfield, Cascadia, and Japan. Despite debates about their mechanism, LFEs are thought to be a useful tool to understand deep fault zone processes. Specifically, because they often accompany slow slip events (SSE), which release energy aseismically and can be associated with earthquake hazards. The template-matching method is typically applied to detect LFEs; however, this type of technique is computationally expensive, and can be challenging when dealing with varying mechanisms and locations of LFEs. Here, we develop a deep learning model to identify P- and S- arrivals of LFEs in Southern Vancouver Island. The model can determine LFE arrivals within a 15-second time window with an accuracy of ~90%, evaluated by unseen testing data. We apply our model to 14 years long of continuous observations and successfully detect more than 1 million LFEs. We locate the LFEs using a grid-searching approach on a total of 1,024,800 3D grid nodes with 1 km spacing in the horizontal and vertical directions. This new catalog greatly improves the temporal and spatial resolution of the existing LFE catalog in Southern Vancouver Island. Our initial analysis has revealed a previously undetected SSE, highlighted by over 3000 LFE detections spanning a 15-day period. This demonstrates new opportunities to understand the subduction zone processes in this region.

Part of the work was prepared under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-ABS-853093

Key Words
Low Frequency Earthquake, Machine Learning, Earthquake Detection, Earthquake Location

Lin, J., Thomas, A. M., Melgar, D., Bachelot, L., & Toomey, D. R. (2023, 09). Detecting Hidden Low-Frequency Earthquakes in Southern Vancouver Island with Deep-Learning. Poster Presentation at 2023 SCEC Annual Meeting.

Related Projects & Working Groups