John Rundle on the Statistical Mechanics of Earthquakes
Special Topic of Earthquakes Interview, December 2010
Page: < 1 | 2 >
Basically we now have these classes of models, and they're leading to a lot of interesting new ways, which had not really been available in the past, to look at earthquakes and to think about them. One of the interesting things that we have discovered is related to the fact that earthquakes are known to be highly clustered in space and in time. That clustering is one of the first-order features of earthquakes. There is a school of thought that says that this is something you want to eliminate and you want to focus only on the background rate of earthquakes. We think the signal is in the clustering and not in the background.
That has motivated us to move to the natural time domain. What you do there is look at sequences of events, like the first, second, third, fourth, fifth earthquake, rather than trying to look at them in linear or calendar time. And this idea of looking at earthquakes in natural time, that is to say the stress-release time, is fairly new and it's not hard to do.
Why it has not caught on more I don't really understand. But it leads to ways of thinking about the forecasting problem which are new and very different. For example, it's well known that a magnitude six earthquake corresponds to roughly 1,000 magnitude three earthquakes, but that kind of thinking is not used in more conventional approaches to earthquake forecasting.
In the next month or two, we're going to fairly dramatically improve the forecasts we have online. We're doing it based on the following ideas: you construct a model on some basis, and then you use some of standard testing, verification and validation procedures to select the best models. This is what the weather community calls data assimilation, but it's typically not been done in the earthquake community.
We've actually found that by doing this we can make significantly improved forecasts compared to would otherwise be the case. I'm doing this along with two former students—James Holiday, who is now a postdoc, and Kristy Tiampo. I work closely with these two, and much of what I've done has been done with them.
What would you like to accomplish in the next five years?
"...this idea of looking at earthquakes in natural time, that is to say the stress-release time, is fairly new and it's not hard to do."
I would like to get a serious global forecast done and published. We have a good one now but we can improve it substantially, which we will. I'd like to really make significant progress on understanding the statistical mechanics of how earthquakes work, and that depends to some extent on getting better data, which I think we're doing now. The global dataset is improving dramatically.
I've been appointed to be on an earthquake forecasting commission, a US-China commission, established by NASA on the one hand and the Chinese space agency on the other. I think we're going to see more and more of this. As Dave Wald at the USGS has said, if you look at what happened to Haiti, there's probably on the order of 60 to 70 cities around the world with populations of five million or more, of which maybe 40 or 50 are in earthquake zones.
The kind of fault that ruptured and destroyed Port-au-Prince has a typical recurrence interval of once every 250 years. So if you do the math, you realize that probably once every five years we're going to see a city like Port-au-Prince destroyed by a major earthquake. The importance of this really can't be overstated. We're talking about the dire need to anticipate these earthquakes so we can do something about it on a global scale. That is something I would really like to contribute to, because I think it's obviously important from the standpoint of future sustainability of people on the planet.
That implies that we've been having a Port-au-Prince-type even every five years. Is this so?
One thing that's changing is people are moving more and more into at-risk areas and the populations of the cities in these areas have been increasing fairly dramatically. There has been a fairly large movement of people in the Pacific Rim from rural areas into cities.
If you go and look at megacities, the location and size of the largest cities on earth, as Dave Wald has said, many of these cities are in seismically active regions where you can expect these earthquakes. This is a fairly new phenomenon in the 20th century and 21st century. The world's population is now something on the order of 6.7 billion, and perhaps 2 to 3 billion live in these seismically active zones.
Now just in the last 10 years, we've had the tsunami in Sumatra, which killed 230,000 people, we've seen the Sichuan earthquake, May 12th 2008, which killed 80,000 people. The Pakistan earthquake in the fall of 2004 killed 80,000 people. We've seen the Port-au-Prince earthquake in Haiti. There was the Penang earthquake, which killed 1,500 people last year. And 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.
What do you find to be the most challenging aspect of the kind of research you're doing?
A couple of things. One is maintaining continuity in the research group, which is probably a common complaint regardless of the field. When you have to go out and search for grant money every two or three years to keep your group going, it's difficult to maintain continuity.
I think if you're trying to bring new ideas into a field like this one, that's also always a struggle, particularly if the idea comes from outside the field—in this case, from physics. People are naturally resistant to new ideas from other fields and this is always a challenge when you're trying to overturn established dogma and trying to bring new ideas into an area.
John B. Rundle, Ph.D.
Department of Geology
University of California,
Davis
Davis, California, USA
Page: < 1 | 2 >
KEYWORDS: EARTHQUAKES, NONLINEAR DYNAMICS, CHAOS THEORY, COMPLEXITY, PHASE TRANSITION, FIELD THEORY, ELASTIC REBOUND, SLIP, FAULT, ENERGY BARRIER, SCALING RELATIONS, POWER LAWS, MODEL, PARKFIELD, PREDICTIONS, FORECAST WEBSITE, REAL-TIME EARTHQUAKE FORECASTING, CLUSTERING, BACKGROUND, NATURAL TIME DOMAIN, DATA ASSIMILATION, GLOBAL FORECAST, STATISTICAL MECHANICS, CITIES, SEISMICALLY ACTIVE REGIONS.