Insights Into the 2019-2022 Southwest Puerto Rico Seismic Swarm and Broader Caribbean Seismo-tectonics With an Automatic Workflow Aided by Machine-learning Pickers
Description:
In December 2019, the southwestern portion of Puerto Rico began experiencing several large, widely felt earthquakes, including a Mw 6.4, which has thus far been the largest of the sequence, on January 7, 2020. The sequence has produced 15 M5.0 or greater and 121 M4.0 or greater earthquakes from late 2019 through mid-2021. Some of the largest mainshocks have caused damage to structures in southwest Puerto Rico (Lopez et al., 2010) and, as of early 2022, the seismicity rate in southwestern Puerto Rico remains elevated above background levels. Seeking to better understand the evolution of this complex sequence, we applied machine-learning (ML) detection, as implemented through the easyQuake (https://github.com/jakewalter/easyQuake). The easyQuake Python package consists of a flexible set of tools for detecting and locating earthquakes from FDSN-collected or field-collected seismograms. We will present a comparison of several different deep learning pickers (GPD, EQTransformer, and PhaseNet) that are user-selected within easyQuake and show relative tradeoffs between different detection capabilities, relative to the routine network catalog generated by the Puerto Rico Seismic Network (PRSN). We show that systematic bias is evident in the timing of phase pick arrivals relative to human-picked phase arrivals and highlight how these biases could lead to cascading errors in ML automatic earthquake catalogs. Aside from the very active portion of southwest Puerto Rico, the results suggest interesting new discoveries of seismicity along the Enriquillo Fault in southern Hispaniola near the Dominican Republic extent of that fault zone. There is ample seismicity during the study time period near the Puerto Rico trench, adjacent to the proposed location (Doser et al., 2005) of the 1943 Mona Passage earthquake, a finding qualitatively similar to routine results from PRSN. We plan to extend the catalog into the end of 2022 and identify the degree of dynamic triggering across Hispaniola and Puerto Rico, if any, that might be occurring due to the large amount of seismic energy release from the Southwest Puerto Rico Seismic Sequence.
Session: The 2020-2021 Southwest Puerto Rico Seismic Sequence: Current State of Knowledge and Implications
Type: Oral
Date: 4/18/2023
Presentation Time: 10:30 AM (local time)
Presenting Author: Jacob Walter
Student Presenter: No
Invited Presentation:
Authors
Jacob Walter Presenting Author Corresponding Author jwalter@ou.edu University of Oklahoma |
|
|
|
|
|
|
|
|
Insights Into the 2019-2022 Southwest Puerto Rico Seismic Swarm and Broader Caribbean Seismo-tectonics With an Automatic Workflow Aided by Machine-learning Pickers
Category
The 2020-2021 Southwest Puerto Rico Seismic Sequence: Current State of Knowledge and Implications