← Back to Sessions

Shakealert Version 3, Current Status and Future Possibilities

Description: 

The primary goal of the ShakeAlert Earthquake Early Warning System is to provide useable warning times before potentially damaging shaking. Version 3 of the underlying algorithm base, live since March 2024, includes numerous improvements aimed at ensuring success in large earthquakes including: multiple ways to initiate events, independent data streams for seismic and geodetic data, and multiple algorithms that characterize large ruptures. It prioritizes speed over accuracy to some extent within 100 km of the epicenter by limiting alerts to that distance for the first five seconds after a detection. Significantly more data become available in that interval to improve ground motion estimates at greater distances. I will give an overview of expected performance in large earthquakes from offline testing and highlight a few recent examples from the production system including successes in the 2021 Petrolia and 2022 Ferndale earthquakes [Lux et al., 2024]. In the 2024 M7 Off-Mendocino earthquake, the MMI 3 and 4 contour products produced maximum warning times of about 10-40 s, 10-20s, and 0-15s at locations with peak shaking of MMI 6, 7 and 8, respectively.

ShakeAlert has clear remaining challenges and several promising areas for system refinement. ShakeAlert allows our partner organizations a range of choices in how they alert their users which requires accurate estimates across a range of magnitude and ground motion levels. While our primary goal is clear, a key need is to clarify what types of alerts we seek to avoid. The answer varies widely among delivery mechanisms and can be tied to legal constraints. Research to-date into the negative outcomes of alerting is limited and represents a key area for improved clarity. Also, ShakeAlert has a well-documented weakness for earthquakes that initiate outside the sensor networks, particularly offshore, and several possibilities exist to improve this. Lastly, new data types may improve our ability to track rupture growth. Improvements in these areas could lead to additional alert products that would better achieve our goals.

Session: Performance and Progress of Earthquake Early Warning Systems Around the World - I

Type: Oral

Date: 4/16/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Jeffrey

Student Presenter: No

Invited Presentation: Yes

Poster Number:

Authors