John Rundle on the Statistical Mechanics of Earthquakes

Special Topic of Earthquakes Interview, December 2010

John RundleAccording to our Special Topics analysis on Earthquakes research over the past decade, the work of Dr. John Rundle ranks at #9 by number of papers, based on 64 papers cited a total of 579 times. In the Web of Science®, his record from 2000-2010 includes 102 articles, reviews, and proceedings papers cited a total of 967 times.

Rundle is Interdisciplinary Professor of Physics and at the University of California, Davis. His group focuses on "statistical physics of complex earth systems."

Below, correspondent Gary Taubes talks with Rundle about his work as it relates to earthquakes.

SW: The idea of applying nonlinear dynamics—i.e., chaos theory—to earthquakes is relatively new and you deserve a lot of the credit. How did this come about?

I was doing classical geophysics until the mid-1980s when I became aware of this area called complexity and chaos theory, which sounded like statistical physics, a subject I had always enjoyed. In 1989, I met Bill Kline, who was at Boston University, and we had the idea that earthquakes were an example of a phase transition. Once you think of them like that, you can describe them with a field theory, which is pretty much the same way they describe phase transitions in high-energy physics—the decay of the false vacuum in the early universe, for instance.

How were conventional geophysicists thinking about earthquakes back then?

Conventional geophysics people were thinking of it in terms of elastic rebound, a theory that goes back to Harry Fielding Reed in 1910. He was one of the group investigating the 1906 San Francisco earthquake with the Carnegie Commission. In the report they published, he described the build-up of forces and then the occurrence of a crack and the slip on the fault, and then the rebound to a new position, and then the repeat of that.

Earthquake cycle.
© 2001 Brooks/Cole - Thomson Learning.

Download additional figures and descriptions of John Rundle's work on earthquake research.

He didn't know anything about plate tectonics, of course, but that was the basic idea. That's kind of where things were, and people were describing earthquakes as elastic dislocations, where you make a cut in an elastic material and move the two sides and glue it back together, and that's how you describe it. The notion of using statistical mechanics was not there in 1989.

Your most-cited paper in the last decade is "Precursory seismic activation and critical-point phenomena," which was published in Pure and Applied Geophysics back in 2000 (Rundle J, et al. 157[11-12]: 2165-82, December 2000). What were you doing in that paper and what did you conclude?

That was a paper where we were trying to understand what leads up to a major earthquake in the context of this phase transition idea. The phase transition is between the unslipped and the slipped state. An earthquake represents a slip on a fault. So the unslipped state is the initial state and the slipped state is the final state. And then there's an energy barrier that the fault has to get over to slip. That's the traditional way of describing a phase transition.

What we were trying to do is come up with a model and its associated scaling relations—power laws, for instance, that are associated with it—and to actually predict the exponents that you would see in association with that. That was the general idea. We were trying to really map the idea of phase transitions into this model of earthquakes and actually compute the scaling laws, from first principles, that you would see in nature.

Why do you think that paper is so highly cited? What made it so influential?

I think one reason is that it's relatively simple. It has a nice physical picture in it. It makes the case better than some of the other papers we did and it makes it better to the particular community that we were trying to reach, the earth science community. It's pretty easy to get lost in a lot of high-level math in this field, and the physics community speaks a very different language than the earth science community.

We were trying to appeal in that particular paper more to the specialists in earth science than to the physicists, and at that time, they probably had more interest in it than the physics community. Although that may have changed since.

Is that why you published in Pure and Applied Geophysics instead of the Physical Review journals or Physical Review Letters?

Yes, we publish in both places but that was the idea there, to try to reach the people in the geophysics community that hadn't been reading our physics papers.

How has that model held up over past decade?

"...the reason these earthquakes are so deadly is not because they're occurring more rapidly, but because more people are living in the at-risk areas."

It's really not a question of how that particular model has held up. That's one of a number of papers, all of which are describing aspects of the same model. So I'd say that particular model has held up really pretty well, and I think other people like Daniel Fisher, who's now at Stanford, worked in this area and were using similar ideas. It's all part and parcel of the same basic model or concept.

Has the research community looking at earthquakes changed their view of your work in the decade since that paper came out? What happened to make it more mainstream and less radical?

Well, a number of things have happened, of course, since we published that paper. One of them is that the USGS and others had predicted that the Parkfield earthquake would happen in 1993 and it didn't end up happening until 2004—sixteen years later. And, as it turned out, the Parkfield earthquake was in many ways quite different from what had been expected, just in terms of the distribution of slip and that sort of thing.

So at that time people started to look around for alternative models. They were dissatisfied with the more traditional points of view and were looking for new things. Around the same time we published a paper in PNAS saying that we thought we could predict where large earthquakes would occur in the future, based upon where the most intense areas of activity had been in the past (Rundle JB, et al., "Self-organization in leaky threshold systems: The influence of near mean field dynamics and its implications for earthquakes, neurobiology and forecasting," Proc. Nat. Acad. Sci. USA 99, Supplement 1: 2514-21, 2002).

Our ideas struck a chord in some way. Also our work provided a point of entrance for some other people in the physics community to come in and start looking at these models as well.

What are you working on now? Has your focus shifted?

We're doing a number of different things. We started a forecast website called, mainly as a form of outreach, but incorporating some of these ideas we've come up with over the years. The idea is to actually do real-time earthquake forecasting that is freely available to the public. Our website is meant to help people understand what their risk is, in terms of probability.

Page: < 1 | 2 >


   |   BACK TO TOP