Real-Time Implementation of ElarmS for Earthquake Early Warning in Northern California

G. Wurman, R. M. Allen, P. Lombard, D. Neuhauser, A. Kireev, and M. Hellweg
University of California Berkeley

Seismological Society of America, Hawaii, April 2007

Earthquake early warning systems seek to detect and characterize earthquakes within the first few seconds of shaking, and disseminate an alarm ahead of the strongest ground motions. The implementation of Earthquake Alarm Systems (ElarmS) in Northern California incorporates several improvements to the ElarmS methodology. These include the incorporation of p-wave peak amplitude data as well as predominant period in the rapid magnitude determination, which has helped to improve magnitude estimates in smaller events; and the incorporation of attenuation relationships from ShakeMap for prediction of shaking hazard during earthquakes. Efforts are currently underway to validate and test Earthquake Early Warning (EEW) algorithms in Northern California. Validation of the ElarmS methodology was completed over one year from February 2006 through January 2007, using data from every M ~ 3 or larger earthquake in Northern California during that period, processed automatically 10 minutes after the event. The validation demonstrates that ElarmS is capable of estimating the magnitude of an event within 0.5 units, and predicting the ground motion to within a factor of 4. Real-time testing of the methodology has been underway since February of 2007 at UC Berkeley, as part of the statewide EEW effort jointly underway with the Southern California Earthquake Center, Caltech and the US Geological Survey. We present the performance of the ElarmS methodology in validation between February 2006 and January 2007, and the performance under real-time testing since February of 2007.