S046-0011
ShakeAlert Earthquake Early Warning Time-to-Alert Improvements in California with Seismic Network Build-Out

Monday, 14 December 2020
Poster
Glenn P Biasi, USGS Geological Survey, Earthquake Science Center, Pasadena, CA, United States and Igor Stubailo, Caltech Seismo Lab, Pasadena, CA, United States
Abstract:
Time-to-alert and coverage area are two critical quality-of-service metrics to consider in deciding whether an earthquake early warning (EEW) system is ready for operations. We describe the method and time-to-alert and area results that were developed as technical support ahead of the decision to offer ShakeAlert statewide in California. Time-to-alert includes times for wave travel to seismic sensors, transmission of seismic data from the field to a waveform processing center, detection of alert-level ground motions, and communication of alerts to technical users and public applications. Time for P-waves to reach a selectable minimum number of stations (n>=4) determines the minimum time to alert. This time is reduced by increasing station density.

A statewide snapshot of time to alert and coverage area was developed ahead of the ShakeAlert public rollout for October 2019. For a grid of potential earthquake sources, the P-wave travel time from the origin is calculated to all currently contributing stations. Assuming 1.0 s, 0.7 s, and 1.0 s for transmission, event detection and alert decision-making, and alert communication, respectively, urban areas of California with the greatest fault hazards would be alerted in 8 seconds or less. The area with useful time to alert increased significantly between May and September 2019, reflecting strategic efforts by northern California networks to add stations in the Sierra Nevada foothills and the northern Coast Ranges. Improvement in the area covered supported the decision to proceed with statewide public distribution of alerts via two methods; the MyShake app (UC Berkeley development with Cal OES support) and by FEMA’s Wireless Emergency Alert system (USGS developed). Coverage has improved even further since then. The time-to-alert estimation method outlined here could be of help to Oregon and Washington as their state officials consider where and how to extend ShakeAlert to their regions.