NH23C-1908
The Role of Real-Time GNSS in Tsunami Early Warning and Hazard Mitigation
Tuesday, 15 December 2015
Poster Hall (Moscone South)
Dara Goldberg1, Yehuda Bock2, Jianghui Geng2, Diego Melgar3, Jessie K Saunders1 and Jennifer Susan Haase4, (1)Scripps Institution of Oceanography, La Jolla, CA, United States, (2)University of California San Diego, La Jolla, CA, United States, (3)University of California Berkeley, Berkeley, CA, United States, (4)UCSD, La Jolla, CA, United States
Abstract:
Evacuation start time and geographical extent are critical factors in reducing casualties in tsunami forecasting, as was evident during the 2011 Mw 9.0 Great East Japan event. The Japanese early warning system is highly dependent on initial estimates of earthquake magnitude and whether or not the earthquake is tsunamigenic. The extent of the evacuation area is predetermined according to magnitude, which for this event was underestimated primarily due to saturation effects associated with seismic instrumentation. GNSS (GPS and other navigation satellites) does not have this limitation and also provides the permanent coseismic motions, but although available was not yet integrated into the Japanese warning system in 2011. We discuss a prototype system for earthquake and tsunami early warning developed at SIO for the western U.S. The system is based on real-time GNSS and accelerometer data, if available at or near a GNSS station. We optimally combine the two types of observations in real time using precise point positioning with ambiguity resolution and accelerometer (PPP-ARA) software to estimate the broadband coseismic displacement and velocity waveforms with a latency of 1-2 seconds. These data allow for accurate detection of P-wave arrivals for near-source medium size earthquakes and greater. Using data from the 2010 Mw 7.2 El Mayor-Cucapah and 2011 Mw 9.0 Great East Japan earthquakes, we demonstrate rapid magnitude estimation, finite-source centroid moment tensor (CMT) solutions for earthquake characterization (is it tsunamigenic?) and static and kinematic fault slip model estimation. In the case of the 2011 event, we show how an accurate magnitude estimate can be obtained in seconds, followed by a finite source model in 2-3 minutes as input to a tsunami prediction model. The first tsunami waves arrived within 30 minutes of earthquake onset and every minute was important in determining an accurate assessment of the geographical extent of evacuation.
