Introduction to Geophysics
Professor Hiroo Kanamori 金森 博雄
John E. and Hazel S. Smits Professor of Geophysics, Califonia Institute of Technology
Japan Academy Prize 2004
Well, one thing not too far from the mind of Californians is earthquakes. How does it occur? Can we predict them? And what should we do in case there is one?Joining Frank Ling (FL) is Professor Hiroo Kanamori (HK) to talk about the science behind earthquakes. Dr. Kanamori is professor of geophysics at the California Institute of Technology and he was recently awarded the Japan Academy Prize. Below is an edited transcript:
FL: Professor Kanamori, thanks for joining us today. First of all, could you tell us what you were awarded for…for the Japan Academy Prize?
HK: I have mainly been working on seismological problems. However, I have been working on other problems too. But mainly this time, they focused on my seismological research and I guess if I view my research myself, the important thing is I tried to understand the basic principles behind earthquake physics. What kind of physics is working behind it. In the past, many years ago, the study of earthquakes was empirical. Empirical means you accumulate observation and plot the data on graphs and see what you can get, but I wanted to understand the basic physics behind it and so, that is my general side of research. Recently, I’m trying to apply it towards earthquake hazard mitigation. Earthquake science or seismology has two basic aspects: one is the more basic research where we want to understand how nature produces earthquakes and the second aspect is how to use that information to reduce the impacts of earthquakes. This is called hazard mitigation. One is more fundamental science and the other is application, and I have been interested in working in both, trying to link those two. More recently, I have been trying to use modern seismology for effective hazard mitigation and so that has been one my research areas these days. For example, when people talk about hazard mitigation, they think of earthquake prediction. Unfortunately, that is a very difficult thing. Earthquakes are essentially a failure process and there are so many factors involved so its very difficult to make precise predictions. We can sort of understand long term behavior but it’s very difficult to tell when and where and how large is earthquake is going to be. Prediction may not be possible or at least it’s not easy to make, so then the next question is how can we use seismology for hazard mitigation. One area is what we call real time seismology. We collect data very quickly after an earthquake happens and by looking at the very beginning of the signal, sometimes we can tell how large the earthquake is going to be. So if you come up with an estimate very quickly, you can make some sort of prediction regarding how large the earthquake is going to be and if you send that information to the people at some distance away by radio communication or computer communication, then people there can tell or at least expect something to happen. And of course this has not been completely implemented because we are still looking at years downstream. We are trying to make it happen, but it hasn’t happened yet.
FL: Realistically, how many seconds or minutes of warning can we get?
HK: Usually we are trying to get within a few seconds after an occurrence of the earthquake and then within 3 seconds or so you can make some estimation regarding how strong the ground motion is going to be. You send in the information and then the question is how people can use that information? Human beings can not react very quickly and one potentially important application is control of structure. There are many large structures these days. High rise buildings, water tanks, highways, and all kinds of things. To some extent, they can control the structures very quickly. Suppose you know something is coming, you can change the properties of the building. You can reduce the impact of ground motion on the structure. Like an automobile, some modern cars have this control system to some extent. If you drive on a rough road, it detects how bumpy that road is. Shock absorbers smooth out the impact. It’s the same kind of concept. If they can transmit information very quickly, they can do something and control the structure so that it can minimize the impact of ground motion. We are working from the seismology end and we are making some progress. But until the engineering end is ready, we can not make it work. So at this moment, we are working in that area, basically to link very fundamental seismology to modern engineering practice.
FL: There was a controversial finding from UCLA recently. A team there had predicted that there’s a good chance an earthquake will occur in Southern California by September 5 of this year. Do you have any comments on that?
HK: This is Keilis-Borok at UCLA. That was a traditional method. You look at a lot of seismic data and you look at the past examples to see whether there is any sort of pattern that are common to the occurrence of large earthquakes in the past. And if you see some pattern, you can make a prediction, but I must say there is tremendous uncertainty in that. I don’t know how often it can be successful. No one knows because they have just started. I am not so enthusiastic about that approach because of the nature of the process. The process is very unpredictable. For example, you can think of some accident on the freeway as rush hour comes. I think the probability of having an accident is higher, that you can tell. But when you single accident somewhere, sometimes because of chain reaction, that accident can grow but if everyone is very careful around it, the accident can be contained to a small event. That is very difficult to predict because you have to really understand the attitude of every driver on the freeway. Small earthquakes are happening all over and whether it grows into a large one or not will depend on lots of factors surrounding it and of course we can not really understand every detail of that. So that’s the reason it is so unpredictable.
FL: So these models, are they based on chaos theory?
HK: Chaos concept is there, yes. You mean the Keilis-Borok theory? In general, behind the chaos theory, there is non-linear dynamics. So in that sense it is, but I would say it’s more empirical by looking at the past examples and to see if there is any common sort of rule to predict an earthquake in the future, but that rule isn’t really a rigid rule. There are all kinds of variations so this is why it is so difficult to make precise predictions. I guess the deadline is September 5th, so something may happen of course. In California there is always a possibility of having a large event so it can happen by chance but if they are right to an extent, something may happen, so we can’t really say that approach is really invalid, but I guess we need to accumulate more experience.
FL: So let’s talk about the geophysics of an earthquake. Can you describe what a typical California earthquake looks like from a physical point of view.
HK: In a way earthquakes are similar wherever you go, but of course in California, there is a big geological structure called the San Andreas fault which runs from the Gulf of California all the way to Cape Mendicino so it’s a long fault. It’s a major structure but the last event that happened in Southern California was in 1857 and that fault has been very quiet because by common sense, if something big happened, it takes some time to accumulate enough strength to produce another earthquake. In a way it has been very quiet; however, because of the structure, there has been a lot of stresses in the crust in Southern California associated with this structure. In 1994 the Northridge earthquake or in 1992 the Landers earthquake, all of these happened off the San Andreas, not on the San Andreas itself. In California, you need to worry about every place, not just the San Andreas.
FL: Which brings me to the next question. Some people have speculated that smaller magnitude earthquakes relieves the stress to prevent a big one. To what extent is that relevant?
HK: Well, theoretically it is true, but it is a very small amount. The energy released in these small earthquakes is so small that in order to have sufficient influence, you really need to have lots of them. In the ordinary sense, some small earthquakes don’t help at all. It’s just a small amount compared to the energy released in big earthquakes. So to the first order, it doesn’t help.
FL: So I understand that in an earthquake, most of the energy is absorbed as heat rather than as seismic waves. What percentage of it goes into heat?
HK: It depends on the earthquake. This has been a very controversial subject and people have very diverse opinions, but it’s really from 0 to 100. For most events, we think the amount of heat generated is fairly small, maybe less than 10 percent, for shallow events like those on the San Andreas. But some people think it is almost 90 percent, so you see how diverse it is. We wrote a paper some 10 years ago on the heat, thermal energy budget, on deep Bolivian earthquake. There was an earthquake on the depths of 600 kilometers, very deep. In this case, we are certain that a large fraction of the energy went to heat, more than 90 percent. So small amount of energy was released as seismic wave and this was for a very large deep focused earthquake. But for shallow events, there is a big debate. Depending on who you ask, the answer can be very different, but we think that a fairly large fraction of energy is released as seismic waves.
FL: I understand you have also done some work in understanding the correlation between the atmosphere and the ground, the acoustic waves. Could you tell us a little about that?
HK: This is called Morning Glory. That’s an interesting wave actually. This is a wave in the atmosphere and unlike seismic waves, it’s a very small wave and propagates at about 10 meters per second. So if you can run fast enough, you can catch it. It propagates only in LA basin and this is produced by a temperature inversion. If you lived in Southern California, you’d know the smog and the smog is mainly caused by the inversion layer. Sometimes temperature increases as you go up. Very often, the temperature on top of Mt. Wilson is about 10 degrees higher than the temperature in Pasadena here. So with temperature inversion, we build some sort of layer, atmospheric layer below one kilometer or so. It’s almost like having an ocean in the atmosphere. If you have an ocean, you can have a wave like a tsunami and this particular wave is very similar to a tsunami but it happens in the atmosphere rather than water. So the propagation speed is slow. It is an interesting wave and it happens maybe five times or so in two years. And exactly what causes this wave is not very well understood. You asked me about the correlation and this correlation is kind of mysterious at the moment. We have only two years of data and we have to get more to see if that correlation is real or not. We don’t know yet actually.
FL: Are there any exciting trends occurring in geophysics or seismology these days?
HK: Well, lots of things are exciting, but to be excited you actually need to know something about it. Because if you don’t know anything about it, nothing looks interesting. Let’s see, one thing exciting in seismology that happened in the last ten years or so is that we can now measure things very quickly and very rapidly. Twenty years ago, for example, it took a long time for us to really come up with the model of an earthquake that has just happened and so these days it’s almost all real time, mainly because of the advancement in computers and computational methodology and high quality seismic instruments, so you can really measure everything very rapidly. Most recently is the development of numerical calculations. You can take very complicated structures of the crust and compute wave field in it and compare that with observed waves so that you can determine the structure and the type of earthquake that has happened very quickly, so in a way if you can understand something rapidly, we can make more rapid progress both in basic science and also for hazard mitigation.
FL: And one last question, what is the difference between seismology and geophysics.
HK: Geophysics covers a lot of other things. The atmospheric problem is not seismology and geophysics also covers ocean science. Like geomagnetism, magnetic problems and magnetic fields falls under geophysics, not seismology. This definition is somewhat arbitrary but there are lots of things that are not covered under standard seismology, like geology to measure the deformation of the crust. Sometimes it is included in seismology but probably traditionally it’s a separate discipline.
FL: Professor Kanamori, thanks for joining us today.
HK: All right!