HR: 08:00h
AN: S21B-01 INVITED     [PDF]
TI: Earthquake Early Warning
AU: * Kanamori, H
EM: hiroo@gps.caltech.edu
AF: California Insitute of Technology, Seismological Laboratory, Pasadena, CA 91125 United States
AU: Allen, R M
EM: rallen@geology.wisc.edu
AF: University of Wisconsin, Dept of Geology and Geophysics, Madison, WI 53706 United States
AB: The complexity of earthquake physics, and the resulting unpredictability of exact rupture behavior makes accurate prediction of earthquakes difficult. However, once the rupture occurs and seismic waves are generated, the behavior is controlled by elastic properties of the earth, and becomes more predictable. If the ground motion information detected near the source is transmitted immediately as an early warning for ground motion to sites at some distance away before the ground motion begins there, the information could be utilized for various damage mitigation measures. An early warning system has been used in practice in Japan, Mexico, and Taiwan. Two approaches are possible: (1) Regional warning, and (2) Site-specific warning. In (1), the traditional seismological method is used to locate an earthquake, determine the magnitude, and estimate the ground motion at other sites. In (2), the beginning of the ground motion (mainly $P$ wave) observed at a site is used to predict the ensuing ground motion (mainly by $S$ and surface waves) at the same site; no attempt is made to locate the event and estimate the magnitude. The first approach is more reliable, but takes a longer time and cannot be used for the sites at short distances. In contrast, the second approach is less reliable, but very fast, and could provide useful early warning to sites even at very short distances where an early warning is most needed. The first approach has been already used in Japan, Mexico, and Taiwan. Here, we investigate the second approach in some detail from the point of view of rupture physics, and practical applications. The basic principle is that $P$ wave carries the information of the source, but not much energy. $S$ wave and surface waves carry most of the energy and cause the damage. Thus, if we can extract critical information regarding the size of an earthquake from the first few seconds of $P$ waves, the information can be used to predict the severity of ground motion at the site. In general, if $P$ wave is small, the event is either small or large but at large distances, and no warning is warranted. However, a large $P$ wave does not necessarily warrant a warning, because the event can be a nearby small earthquake with short duration of slip motion. Thus, it is important to determine whether the event's slip motion has stopped or keeps growing. To determine whether the event is growing or not, the parameter $\tau$ used by Nakamura provides a good diagnostics. $\tau$ can be interpreted as a spectrally- weighted period during the first few sec (3 sec in this study). Using simulated records computed for an earthquake rupture model, we have verified that $\tau$ is a good measure of the lower-bound of the size of an earthquake. We have determined $\tau$ for events with $M_W=3$ to 7.6 (1999 Chi-Chi earthquake), and the results are consistent with the simulation results. If $\tau <1$ sec, the event has already ended or is not likely to grow. If $\tau > 1$ sec, it is likely to grow, but how large it will eventually become cannot be determined. In this sense, this method provides a threshold warning. Thus, a combination of the amplitude and $\tau$ from the first 3 sec can provide a useful site-specific early warning. A use of multiple sites is desirable to increase the reliability. This approach can provide a very rapid warning that strong ground motions are imminent, but does not provide the ground-motion time history. For the next step of early warning applications, it is desirable to develop a method to estimate the time history which can be incorporated in predictive structural control in engineering practice.
DE: 7200 SEISMOLOGY
DE: 7209 Earthquake dynamics and mechanics
DE: 7212 Earthquake ground motions and engineering
DE: 7215 Earthquake parameters
DE: 7223 Seismic hazard assessment and prediction
SC: Seismology [S]
MN: 2003 Fall Meeting