ShakeAlert: A Unified EEW System for California

Holly Brown, Richard Allen, Maren Böse (Caltech), Georgia Cua (ETH), Egill Hauksson (Caltech), Thomas Heaton (Caltech), Margaret Hellweg, Oleg Khainovski, Doug Neuhauser, Kalpesh Solanki (Caltech), Michael Fischer (ETH)


Earthquake Early Warning (EEW) is a method of rapidly identifying an earthquake in progress and transmitting warning messages to nearby population centers before damaging ground shaking arrives. The first few seconds of the initial P-wave arrival are used to detect the event and predict magnitude and peak shaking from a single or multiple stations, and combined station detections are used to locate the event. Warnings of imminent shaking can be used to activate automatic safety measures, such as slowing down trains, isolating sensitive factory equipment, or opening elevator doors. Warnings can also be sent directly to the public via cell phone, computer, television, or radio.

The Berkeley Seismological Laboratory (BSL) worked for several years to develop the Earthquake Alarm Systems (ElarmS), an EEW system specifically for California. ElarmS was tested in conjunction with two other prototype EEW systems in a three-year proof of concept project by the California Integrated Seismic Network (CISN) to demonstrate the potential for EEW in California. In August 2009, the proof of concept project was completed and declared a success. The CISN EEW partners, the BSL, the California Institute of Technology (Caltech) and the Swiss Institute of Technology (ETH), are now collaborating again in a new three-year project, to build a single, integrated, production-grade system for testing purposes in California. The new system, called ShakeAlert, will be capable of continuous long-term operation, exhibit reliability and redundancy during extreme ground shaking, and rapidly provide warning to large numbers of users across the state. ShakeAlert is already under development and will provide warnings to a small group of test users by August 2012.

Project goals

The new ShakeAlert algorithm utilizes the best aspects of each of the three test systems from the proof-of-concept project. Caltech's OnSite algorithm uses P-wave data from the single station nearest the epicenter to provide an extremely rapid estimate of likely ground shaking. The BSL's ElarmS algorithm and ETH's Virtual Seismologist algorithm use data from multiple stations around an event epicenter to produce a slightly slower but more reliable estimate of magnitude and location. Combining these methods produces an algorithm which has the speed of a single-station method but is then promptly confirmed and adjusted by additional station data to form a more accurate description of the event.

When an event is identified and determined to be above pre-defined thresholds for magnitude, ground shaking intensity and statistical likelihood, event data is broadcast to system users. Currently, event information is only being sent to the developers while system components are being developed and refined. By 2012, event information will be sent to a small group of test users outside the seismological community. The event data will be received by a user-configurable alert program, which will sound an alarm, predict shaking at the user's specific location, and/or display a map of regional shaking. Emergency responders and regionally diverse industries (e.g. trains, utilities) may choose to display a map, showing multiple points of interest, while an individual may only be concerned with the immediate hazard at their specific location.

Figure 2.35: Components of the new ShakeAlert EEW System
\begin{figure*}\centering\epsfig{, width=18cm}\setlength{\abovecaptionskip}{0pt}\end{figure*}

Current Progress

ShakeAlert will be comprised of four primary software components (Figure 2.35). The first, called the Waveform Processing Module, receives seismic waveforms from every early-warning capable seismometer or accelerometer in California (Figure 2.36), identifies P-wave arrivals, and calculates the relevant P-wave parameters necessary for EEW magnitude estimation. The second component, called the Event Monitoring Module, combines the current algorithms from OnSite, Virtual Seismologist, and ElarmS to recognize events in progress and calculate event magnitude and location. The third component, called the Decision Module, can receive event notifications from multiple systems include the existing three test systems and the new integrated system, reviews events, and determines whether to send warnings to users. The final component is the User Display, which will be installed at an EEW user's site. The User Display will receive warnings from the Decision Module and generate an alert message, a map of ground shaking intensities at various locations, or other output, as per the user's preconfigured settings.

During the last year, Caltech programmers developed an initial prototype of the User Display Module, while Berkeley programmers built an initial prototype of the Decision Module to combine output from the three original event monitoring algorithms. The three CISN EEW partners (Caltech, BSL, ETH) are working together to jointly build a new, robust Waveform Processing Module. The BSL, with significant input from the other partners, is also beginning work on the new Event Monitoring Module, combining the three original algorithms into a single, efficient algorithm that takes advantage of the best parts of each original algorithm.

Figure 2.36: Seismic Stations used by ShakeAlert
\begin{figure}\centering\epsfig{, width=8.5cm}\setlength{\abovecaptionskip}{0pt}\end{figure}

Future Steps

During the coming year, the CISN EEW project members will connect all the new modules into a single end-to-end data flow. We will also begin sending warning messages to a few test users outside of the seismology community. Concurrently, we will continue to update and improve our new modules for speed and accuracy. The new event monitoring module in particular presents many opportunities for testing new methods and theories of EEW. While it will use the OnSite algorithm for estimating earthquake characteristics from a single station's P-wave data, and the Virtual Seismologist and ElarmS algorithms for analyzing earthquakes when three or more stations provide P-wave data, none of the three original test algorithms offer insight into how to combine data from exactly two stations. The CISN EEW group has been investigating the best method for analyzing two-station events using Bayesian statistics.


Allen, R.M., H. Brown, M. Hellweg, O. Khainovski, P. Lombard, D. Neuhauser, Real-time earthquake detection and hazard assessment by ElarmS across California, Geophys. Res. Lett., 36, L00B08, doi:10.1029/2008GL036766, 2009.

Allen, R.M., P. Gasparini, O. Kamigaichi, M. Böse, The status of earthquake early warning around the world: An introductory overview, Seism. Res. Lett., 80(5), 682-693, 2009.

Böse, M., E. Hauksson, K. Solanki, H. Kanamori, T.H. Heaton, Real-time testing of the on-site warning algorithm in southern California and its performance during the July 29 2008 $M_{w}$5.4 Chino Hills earthquake, Geophys. Res. Lett., 36, L00B03, doi:10.1029/2008GL036366, 2009.

Cua, G., M. Fischer, T. Heaton, S. Wiemer, Real-time performance of the Virtual Seismologist earthquake early warning algorithm in southern California, Seism. Res. Lett., 80(5), 740-747, 2009.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2007, The Regents of the University of California