Probing Further the Cascadia Initiative Data to Detect New Offshore Events
Description:
The Cascadia subduction zone is notorious for being seismically quiet, and its young, hot, and buoyant subducting slab may narrow the seismogenic zone. Permanent seismic networks that monitor the region rely heavily on onshore stations and a few offshore stations on the OOI Regional Cabled Array. The most spatially complete offshore dataset comes from the Cascadia Initiative (CI) experiment, which successfully deployed over 60 ocean bottom seismometers (OBS) across the entire subduction zone at ~10-month increments, providing a total of 4-5 years of temporally fragmented data. Several attempts were made to detect events. Stone et al. (2018) initially found ~271 between the coast and the deformation front events using conventional network detection methods. Morton et al. (2023) used refined techniques to expand this conventional catalog with several thousand more events but found few new offshore events near the subduction front outside of the triple junction.
Recent advancements in Deep Learning have demonstrated notable success in earthquake detection and phase picking from seismograms, particularly in challenging environments in noisy offshore OBS data. Bornstein et al. (2023) used transfer learning to re-train phase pickers on a curated data set of offshore records, and Yuan et al. (2023) employed an ensemble deep-learning approach also to improve phase picking in the same dataset. We use the latter to detect events within continuous data of the CI network. Additionally, we utilize the PyOcto phase associator developed by Munchmeyer (2023) to associate and crudely locate the events. We present the preliminary results of this search with a direct comparison between the ANSS-USGS catalogs, the Stone et al. (2018), and the Morton et al. (2023) catalogs. Preliminary results indicate that ensemble deep learning finds several additional events near the subduction front, suggesting that this method can successfully supplement conventional methods in the ocean-bottom setting.
Session: Marine Seismoacoustics [Poster Session]
Type: Poster
Date: 5/1/2024
Presentation Time: 08:00 AM (local time)
Presenting Author: Hiroto
Student Presenter: Yes
Invited Presentation:
Authors
Hiroto Bito Presenting Author hbito@uw.edu University of Washington |
Marine Denolle Corresponding Author mdenolle@uw.edu University of Washington |
Yiyu Ni niyiyu@uw.edu University of Washington |
Qibin Shi qibins@uw.edu University of Washington |
Zoe Krauss zkrauss@uw.edu University of Washington |
|
|
|
|
Probing Further the Cascadia Initiative Data to Detect New Offshore Events
Category
Marine Seismoacoustics