Testing Cascadia and San Andreas Fault-specific Finder Templates with the 5 December 2024 M7 Offshore Cape Mendocino Earthquake and Scenario Earthquakes
Description:
FinDer is a finite-source characterization algorithm that estimates the rupture-area, magnitude, and centroid location of an ongoing earthquake by matching interpolated images of observed peak ground acceleration (PGA) to precalculated PGA distribution templates. We test FinDer’s performance with Cascadia Subduction Zone (CSZ) and San Andreas Fault (SAF) fault-specific templates and apply code changes that optimize the fault-specific template matching process. The CSZ template set contains more than 18,000 templates. To streamline the template matching process, FinDer limits its search to templates within a specified magnitude range and distance range from the epicenter. Our code has been modified to check templates within a specified distance of the centroid, which can change significantly during a large rupture, instead of the epicenter. The logic for selecting the starting PGA threshold for template matching has also been modified to default to the overall best-fitting PGA threshold from the previous iteration of template matching instead of a per-set threshold. Our suite of test events includes real waveforms from the December 5, 2024 M7 Cape Mendocino earthquake, synthetic seismograms from six simulated M9 earthquake scenarios on the Cascadia fault with varying hypocenters, down-dip rupture extents, slip distributions, and locations of high-stress drop subevents, and three M7.8 SAF earthquake scenarios with different hypocenters. The Cape Mendocino event, located west of the Mendocino Triple Junction, is ideally located for testing performance for a large, but not great, offshore earthquake with onshore ground motions that FinDer could falsely match with one of the CSZ or SAF fault-specific templates. Preliminary results show that constraining the matching process to templates corresponding to ruptures with centroids within a specified distance of the centroid estimate rather than the estimated epicenter and the modifications to the PGA threshold selection logic improve the agreement between matched templates and observed ground motions in the M9 earthquake scenarios.
Session: Performance and Progress of Earthquake Early Warning Systems Around the World [Poster]
Type: Poster
Date: 4/16/2025
Presentation Time: 08:00 AM (local time)
Presenting Author: Mika
Student Presenter: No
Invited Presentation:
Poster Number: 125
Authors
Mika Thompson Presenting Author Corresponding Author usherm42@uw.edu University of Washington |
J. Renate Hartog jrhartog@uw.edu University of Washington |
Maren Böse mboese@sed.ethz.ch ETH Zurich |
Jessie Saunders jsaunder@caltech.edu California Institute of Technology |
Claude Felizardo claude@caltech.edu California Institute of Technology |
Jennifer Andrews jen.andrews@gns.cri.nz GNS Science |
|
|
|
Testing Cascadia and San Andreas Fault-specific Finder Templates with the 5 December 2024 M7 Offshore Cape Mendocino Earthquake and Scenario Earthquakes
Category
Performance and Progress of Earthquake Early Warning Systems Around the World