Machine-Learning Detection and Waveform Correlation to Probe New Madrid Seismogenesis
Description:
Decades of research has failed to determine the physical mechanism(s) leading to reactivation of the New Madrid Seismic Zone nor resolve the long-term deformation and earthquake clustering history expected for the broader Central US. As in areas associated with induced earthquakes, where some but not all intraplate faults have become active under very small changes in subsurface stress, in New Madrid, seismic reflection data has revealed faults with Quaternary deformation that do not currently host earthquakes. Here, we focus on providing fundamental new earthquake datasets through application of state-of-the-art improvements in detection, association and location, including waveform correlation to improve relative location. We show preliminary results from the application of a machine-learning (ML) approach (easyQuake - https://github.com/jakewalter/easyQuake) to build out a detection catalog extending from the mid-1990s to present applied to all permanent (Cooperative New Madrid Seismic Network) and high-quality temporary seismic networks (ie., Transportable Array, NELE, OINK). EasyQuake includes an associator and earthquake location algorithm and we explore how to best associate earthquakes with multiple ML algorithms (PhaseNet, EQTransformer, GDP) and training datasets applied to the same continuous waveform data. In parallel, we update the event-to-event waveform correlation datasets for the New Madrid area from 2008-2022 using the New Madrid Cooperative Network phase and waveform catalog to understand if highly similar (swarms) or repeating earthquakes have continued since initially reported (Bisrat et al., 2012). Ultimately the ML and correlation data will be combined to update and expand a high-resolution New Madrid seismic zone earthquake catalog, which in turn will allow for updated earthquake statistics, identification of repeating and clustered earthquakes, and improved earthquake source characteristics and focal mechanisms.
Session: Tectonics and Seismicity of Stable Continental Interiors
Type: Oral
Date: 4/19/2023
Presentation Time: 04:45 PM (local time)
Presenting Author: Heather R. DeShon
Student Presenter: No
Invited Presentation:
Authors
Heather DeShon Presenting Author Corresponding Author hdeshon@smu.edu Southern Methodist University |
Jacob Walter jwalter@ou.edu Oklahoma Geological Survey |
Prajwal Nuepane pnuepane@smu.edu Southern Methodist University |
Raymond Ng raymond.ng@ou.edu University of Oklahoma |
|
|
|
|
|
Machine-Learning Detection and Waveform Correlation to Probe New Madrid Seismogenesis
Category
Tectonics and Seismicity of Stable Continental Interiors