Implementing Machine-Learning Earthquake Detection to Augment the Public Regional Seismic Network in Oklahoma
Session: Network Seismology: Keeping the Network Running While Integrating New Technologies I
Type: Oral
Date: 4/22/2021
Presentation Time: 10:45 AM Pacific
Description:
Last year we developed and publicly released a Python package, easyQuake (https://github.com/jakewalter/easyQuake), that consists of a flexible set of tools for detecting and locating earthquakes from FDSN-collected or field-collected seismograms. The package leverages a machine-learning driven phase picker, coupled with an associator, to produce a Quake Markup Language (QuakeML) style catalog complete with magnitudes and P-wave polarity determinations. The program outputs catalog QuakeML-formatted files that can be split into event QuakeML files. After the nightly runs on day-long seismograms we add the event QuakeML files into the real-time SeisComP system, which are then reviewed by OGS analysts. Since implementation, easyQuake has doubled the number of events that we are able to detect and enter into the ANSS catalog in Oklahoma. Because the fundamentals of the package are scale invariant, it has wide application to seismological earthquake analysis from regional to local arrays, and has great potential for identifying early aftershocks that are otherwise missed. While easyQuake is flexible enough that it can perform earthquake detection for regional networks and microseismicity studies in arbitrary user-defined regions, we discuss the public network application. We envision scenarios whereby networks can run easyQuake in parallel with other systems as a “check” on the performance of a real-time system, as a failsafe, or to supplement primary systems during the early aftershock period. In our presentation, we plan to showcase further progress on developing the next stages of easyQuake, including quasi real-time detection, the ability to choose from several pickers, and further refinements for plotting and seismicity visualization.
Presenting Author: Jacob Walter
Student Presenter: No
Authors
Jacob Walter Presenting Author Corresponding Author jwalter@ou.edu University of Oklahoma |
Paul Ogwari pogwari@ou.edu University of Oklahoma |
Andrew Thiel athiel@ou.edu University of Oklahoma |
Fernando Ferrer fernando.ferrer@ou.edu University of Oklahoma |
Isaac Woelfel iewoelfel@ou.edu University of Oklahoma |
|
|
|
|
Implementing Machine-Learning Earthquake Detection to Augment the Public Regional Seismic Network in Oklahoma
Category
Network Seismology: Keeping the Network Running While Integrating New Technologies