An Earthquake Early Warning System Validation Framework for Western Canada
Description:
We present a framework for validating synthetic seismograms for use in earthquake early warning (EEW) systems, that are at the forefront of seismological and seismic hazard research. The Cascadia Subduction Zone (CSZ) is generally seismically-quiet yet has historically hosted large (>M8) events, and seismologists estimate a 7-15% chance of a margin-wide rupture in the next 50 years. Developing an EEW algorithm is crucial in minimizing fatalities and loss resulting from an imminent event. The algorithms used in most EEW systems require seismic data with which to train them. Although the paucity of seismicity along the CSZ is a limiting factor in training algorithms, synthetic seismic and geodetic data can be used as a supplement. However, a framework has yet to be published validating synthetic data for such applications.
We use a set of semi-stochastic forward modeling codes to generate 112 CSZ, as well as two crustal Leech River Valley Fault Zone (LRVFZ) rupture scenarios and associated waveforms. The CSZ scenario events range from 6.8 < M < 9.5, and the LRVFZ scenario events are an M6 and M7. The waveforms are generated for 29 collocated GNSS and accelerometer stations located on Vancouver Island, 157 accelerometer stations located along Washington and Oregon, and 5 offshore accelerometers. To ensure the synthetic data are representative of true earthquake scenarios, we develop a validation framework consisting of three main steps. (1) We compare high-frequency peak ground acceleration, peak ground velocity, and Modified Mercalli Intensity; (2) low-frequency peak ground displacement; and (3) P-wave displacement amplitude (Pd) with well-validated models. We also evaluate whether the scenario P-waves would trigger an STA/LTA algorithm, which is key in determining the onset of an earthquake before the more destructive waves arrive. Through the validation process, we find the synthetic seismic data to be representative of true earthquake scenarios and a valuable component in the training of EEW algorithms in western Canada.
Session: Earthquake Early Warning Optimization and Efficacy
Type: Oral
Date: 4/20/2023
Presentation Time: 05:15 PM (local time)
Presenting Author: Tara Nye
Student Presenter: Yes
Invited Presentation:
Authors
Tara Nye Presenting Author Corresponding Author tnye@uoregon.edu University of Oregon |
Valerie Sahakian vjs@uoregon.edu University of Oregon |
Angela Schlesinger schlesin@uvic.ca Ocean Networks Canada |
Diego Melgar dmelgarm@uoregon.edu University of Oregon |
Alireza Babaeimahani alirezab@uvic.ca Ocean Networks Canada |
Benoit Pirenne bpirenne@uvic.ca Ocean Networks Canada |
|
|
|
An Earthquake Early Warning System Validation Framework for Western Canada
Category
Earthquake Early Warning Optimization and Efficacy