Research articles predict next earthquake

November 11, 2005

Two research articles published this week illustrate continuing efforts to predict earthquakes. The first article builds upon the idea that analysis of the first seismic waves from an earthquake can provide a way to generate a several second warning that an earthquake will hit. However, for such warnings to be useful, the early seismic waves should also allow prediction of the strength of the earthquake. An article published in this week's edition of Nature concludes that scientists might be able to distinguish small earthquakes from big ones by using the very first second of information contained in seismic waves. Researchers from the University of California, Berkeley, say the measurements of seismic waves soon after a trembler can signal whether it will be a minor or monster trembler. "Basically, a high-pitched squeal means that you'll get a smaller quake. A low-groan means something bigger," says the University of California's Richard Allen.

In the study, Allen and colleagues analyzed records of 71 major Pacific Rim quakes in the past decades including 24 events that were greater than a magnitude 6. Using a mathematical model, they were able to estimate a quake's size to within one magnitude unit from as little as four seconds of data of the frequency of the energy in the primary wave. These low-energy waves typically cause a jolt, signaling the occurrence of a quake.

According to a traditional theory, called the cascade model, we cannot know anything about an earthquake's ultimate magnitude until it is finished. This is because spread along a fault line depends on the stress in each individual part of the fault. Given that this information is not available to the initial rupture point, it should be impossible to tell from that first slip how far it will go or how long it will last.

The study by Berkeley's Allen and Erik L. Olson of the University of Wisconsin-Madison builds upon earlier work by Yutaka Nakamura in Japan. Nakamura's research indicated that the frequency of the P wave can allow prediction of which earthquakes will have a magnitude greater than 6 and distinguish them from small tremors that do not require warnings. Some fire stations use P wave detectors to automatically open doors in an effort to make sure that emergency vehicles are not trapped inside by doors that can be jammed shut by earthquake damage.

The goal of this research is to make the best possible use of the information from the earliest seismic waves emitted by an earthquake as part of an alert system to give seconds to tens of seconds of advance notice of an impending quake -- enough time for schoolchildren to take cover, power generators to trip off and valves to shut on pipelines. A major limitation of this work is that the greatest earthquake damage is near the epicenter where there is little time difference between the early P waves and the strongest damaging earthquake waves.

The second research paper, published this week in the scientific magazine Physical Review Letters, builds upon earlier research into patterns of earthquake activity at specific faults. The new research by scientists from Israel and Germany led by Prof. Shlomo Havlin, of Bar-Ilan University's Department of Physics in Israel, concerns prediction of the timing between earthquakes.

Prof. Havlin's research, in collaboration with Prof. Armin Bunde, of the Justus-Liebig University in Giessen, Germany, and Bar-Ilan University graduate student Valerie Livina, involved a large number of past earthquakes at many different faults. The data used in this study were for earthquakes ranging from magnitude 2 to magnitude 5.5 on the Richter Scale. Data were from several areas of the world including California, Japan, and New Zealand. The new results are consistent with previous studies which have indicated that some faults have more frequent small earthquakes while other faults have less frequent and larger earthquakes.