GSA Annual Meeting in Phoenix, Arizona, USA - 2019

Paper No.
Presentation Time: 5:00 PM


COCHRAN, Elizabeth S.1, MINSON, Sarah E.2, BALTAY, Annemarie S.2, BUNN, Julian3, PARKER, Grace A.4, KILB, Deborah L.5, HOSHIBA, Mitsuyuki6 and KODERA, Yuki6, (1)Earthquake Science Center, U.S. Geological Survey, Pasadena, CA 91106, (2)Earthquake Science Center, U.S. Geological Survey, Moffett Field, CA 94035, (3)California Institute Of Technology, Pasadena, CA 91125, (4)Earthquake Science Center, U.S. Geological Survey, Moffett Field, CA, (5)Institute of Geophysics and Planetary Physics, Scripps Institution of Oceanography, 9500 Gilman Drive, La Jolla, CA 92093-0225, (6)Meteorological Research Institute, Japan Meteorological Agency, Ibaraki, Japan

The 2019 Ridgecrest earthquake sequence was the first real test of the ShakeAlert earthquake early warning system. At the time of the sequence, public alerting had been recently initiated for Los Angeles County through the ShakeAlertLA smartphone application developed for the City of Los Angeles. Additionally, a host of technical users had access to a more detailed alert stream from ShakeAlert. We provide an overview of the technical performance of the ShakeAlert system, and also describe the initial public response to the system. We also look ahead to new methods that are being tested for possible integration into ShakeAlert. We have been assessing the PLUM (Propagation of Local Undamped Motion) method that is one of two methods used in the Japan Meteorological Agency (JMA) earthquake early warning (EEW) system. Traditional EEW methods attempt to infer the location and evolving moment release of an earthquake in real-time, and then use that information to calculate expected ground motion and thus identify which regions should be warned. PLUM instead uses the observed ground motion field to forward-predict shaking to nearby sites. This allows PLUM to adapt to ground motion variability such as the basin amplification observed in Los Angeles during the July 2019 Ridgecrest, California, earthquake sequence as well as higher (or lower) stress drop events that cause ground motions to deviate from expected median values. Recently, we began retrospective and real-time testing of the potential of using PLUM for EEW in California, Oregon, and Washington. Fortuitously, a real-time version of PLUM was running on one of the ShakeAlert EEW system’s development servers at the time of the Ridgecrest sequence, allowing a side-by-side performance comparison with ShakeAlert’s traditional earthquake-source-based approach. We present a holistic overview of earthquake warning during the Ridgecrest sequence and discuss how the PLUM algorithm can contribute to EEW for the West Coast of the United States.