PubTalk 4/2011 - Predictable Earthquakes

Video Transcript
Download Video
Right-click and save to download

Detailed Description

Title: Predictable Earthquakes - updating earthquake prediction - fact vs. fiction 

  • Although scientists were optimistic about earthquake prediction in the 1970s, reliable short-term prediction has remained an elusive goal
  • What have seismologists learned from recent earthquakes in Haiti, Chile, and Japan? 
  • Great strides have been made in earthquake forecasting, and to a large extent damaging earthquakes are predictable 
  • Active fault zones have been identified where damaging earthquakes are inevitable — possibly within our lifetimes

Details

Image Dimensions: 640 x 480

Date Taken:

Length: 01:24:43

Location Taken: Menlo Park, CA, US

Transcript

Good evening. It looks like we still have a few empty seats, but I’m delighted to see so many faces out here. My name is Leslie Gordon, and I welcome you to the U.S. Geological Survey and another evening in our continuing public lecture series. And, those of you come regularly – I know there are many of you – you know that I always have announcements because I want you to come back next month and the month after. Next month, May 26th, we’re going to have a lecture on rare Earth elements. Those are those obscure elements at the bottom of periodic table in that little line that’s always sort of a footnote to the periodic table. But they’re ubiquitous in a lot of our electronics, and there’s some interesting political implications about where those elements come from and where they’re mined. Exactly. I heard somebody say China. So do join us next month. It’s an interesting talk. Just as a consumer, you should be concerned where your materials are coming from. In June, which is National Oceans Month, Sam Johnson will be talking about mapping the seafloor offshore California. So join us then. July, we’re going to be talking about climate variability very far away in the Pacific Ocean and how it impacts San Francisco Bay close to home. August we’re going to be talking about repeat photography and using that as a scientific tool. In other words, taking pictures, for example, of a landscape. A hundred years later, taking another photograph and understanding changes in the landscape. September we’ll be talking about groundwater resources. And I know I’m getting pretty far out, but I just want to whet your appetite. October is migratory birds. And November will be Grand Canyon monitoring activities. And then December will be about Hawaiian volcanoes. So we’ve got a great lineup for the rest of the year. Do join us. And in case you haven’t noticed, or if you’re not on our mailing list, there’s a box in the back of the room. You can put your – we prefer emails, but sign up on our mailing list to get a notice of our lectures. I also want to point out there’s some handouts, some fact sheets, and information. There’s also some fliers up there about a safety fair. This is very apropos. This weekend, Sunday, May 1st, in the afternoon, there’s a safety fair co-sponsored by the city and Stanford Shopping Center and – the City of Palo Alto, that is, and the USGS and a variety of community organizations. So very timely. Go to an earthquake safety fair. Or, it’s more than just earthquakes – all kinds of emergency safety fare at home. So do go and get yourself educated and prepared. Tonight it is my pleasure to introduce Dr. Susan Hough. Susan has a bachelor’s degree in geophysics with honors from the University of California in Berkeley. She received her Ph.D. in Earth sciences from Scripps Institution of Oceanography in San Diego. From there, she went to Lamont-Doherty Geological Observatory in New York for several years. And came to us at the U.S. Geological Survey in 1992. And Sue works out of our Pasadena office that’s co-located on the Caltech campus. Sue’s current research interests include ground motions, remotely triggered earthquakes, historical earthquakes, and specifically seismic hazards in South Asia and the Middle East. Through most of the last year, through 2010, she led the USGS effort in Haiti to install and deploy seismometers in the wake of their earthquake, working in partnership with the Haitian Bureau of Mines. She’s published over 100 scientific articles and journals. She’s a fellow of the American Geophysical Union. And she is a sought-after commentator and writer, publishing in mainstream newspapers and magazines such as Natural History, the L.A. Times, CNN, and New Times. It’s my great pleasure to introduce Sue Hough speaking about predictable earthquakes – updating earthquake prediction – fact versus fiction. Thank you.

[Applause]

- Good evening. And the first question is whether people can hear me in the back and if I’m wired for sound. No? Okay, can – I don’t have a booming  voice, so I’m relying on this guy to make me audible. Okay. Is that better? Can you hear me? Okay. Now I just have to not move. So, yeah, I had a few very brief introductory remarks of my own. The first is that as – if you’re regulars, your probably know there’s a small group of people who work very hard to make these lectures happen. Starting with Leslie herself, but some other people you might not see – Amelia Barrales, Bill Rambo, who might be in the wings somewhere, and a few others. So I’d like to acknowledge their hard work. Thank you. [Applause] The second is that it’s – I’ve done this once before, and it’s a pleasure to be up here and be part of this series. Tonight, I’m especially glad to be giving a public lecture here as opposed to my home base in Pasadena. Some of you may know that the Lakers are playing game six, and in southern California, it is not a winning game to compete against the Lakers. [laughter] Here nobody cares. If you do, the Lakers are up by nearly 20 at the start of the fourth, so … [laughter]

- Boo! [laughter]

- And then I put the – this photograph – that’s me. About five days ago, I had – I was in Nepal for a workshop on seismic resilience. It was a very good professional experience. I had a chance to go visit elephants. I got back on Monday from the other side of the planet and have been a bit of a zombie since then. So this is not the way to really plan things optimally. I’m feeling better now. I had a nap this afternoon, so we'll see how this goes. And then the last introductory comment is that I know earthquake prediction is a fairly controversial subject. It tends to arouse some passions on both sides. Maybe not quite as much as like, evolution versus creationism, but it comes close. So I know there are some diverging views out there in the world. I’m not sure about in the room. You’re going to get what I think is a fairly balanced view from one seismologist. One of the points that I want to make through the talk is I’m talking about science [feedback], as we understand it now, what we know, what we can say, what we can’t say. And there is, without doubt, things that are going on in the Earth that we still don’t understand. And there may be things related to predictability of earthquakes that we don’t understand. So when you say – when I say earthquakes aren’t predictable, I’m talking about right now, as opposed to …

- Louder?

- Louder. Okay. Hm. I disappeared. Maybe with that brrr. Okay, am I better? No. Okay. I mean, I can hold it, or I can use a mic. But, okay, better now?

- [multiple responses] Yes.

- Okay. So – yeah, so let me just go ahead and get started before I take too much time. Okay, earthquake prediction. I wanted to start with a couple of quotes to set the stage. The first is by Stephen Hawking. In effect, we have redefined the task of science to be the discovery of laws that will allow us to predict events. A second quote that you may have heard from Ernest Rutherford is that all science is either physics or stamp collecting. [laughs] And both of those essentially are getting at the same point, although they don’t – they sound different. But the idea is that science, at one point, was about describing the natural world. You know, you went out and you collected butterflies and you described them. Or you described rocks. And modern science – the litmus test for modern science is to go beyond classification and to really understand how a system works. And if you can understand it, you can predict its future behavior, right? And so that’s what we’ve kind of defined as the business of modern science, which then leads to a question for seismologists. If we can’t predict … [laughter] Sorry. So, and in fact, there’s answers to this question that I’m sure you know there’s a lot of work in seismology that – a lot of things we can predict, like how hard the ground is going to shake if an earthquake happens, for example. But the specific question of earthquake prediction has been a tough nut to crack, so to speak. So let me talk about that for the next 45 minutes or so. The outline of where I’m going. I just want to introduce the subject – you know, some ABCs. Fairly basic. And then do a bit of a historical retrospective of some of the past predictions – sort of more notorious instances. Talk a little about recent research – the good, the bad, and the ugly. And there is recent research going on about prediction. And then getting to predictable earthquakes. There is a lot of predictability about earthquakes – what we – what we refer to as forecasting. So forecast versus prediction is one – sort of one bit of the ABCs to start with. I think people have a good sense of this in general, that when we talk about forecasting, we make statements in terms of probabilities. What are the odds of a damaging quake in the Bay Area over the next 30 years? That’s a typical way that we cast – forecasting. [feedback] I’m not even sure what I’m doing that’s causing that.

- [inaudible]

- Okay. It’s not me. It’s the system. Okay. Whereas, predictions – you know, I think – I think that’s generally [feedback] – should I go to the hand mic and just try to – would that help?

[Silence]

Okay. [laughs] [creaking sounds] [laughter] Oh, we’re just having all kinds of fun. Okay. I will do this. There are some days when, you know, you just aren’t going to win. So predictions – I think, again, people, I think, do have a good sense of this. If you’re going to make a prediction, you have to give a usefully narrow window. You know, what’s the magnitude that we’re talking about? What’s the location, and what’s the time window? So an earthquake is going to happen in this place, this magnitude, in this time. And you have to be more accurate than a prediction based on what I call educated guessing. And these things sound pretty simple. In fact, they’re more complicated then they sound. For example, you know, I can make useless predictions that are obviously useless. There will be an earthquake in California tomorrow, you know, of some magnitude. It’s certain. There will be small earthquakes. There will be a magnitude 4 somewhere in the world tomorrow. That’s – we know that that’s going to happen just based on the rates that earthquakes happen. There will be a 6 or somewhere greater in the next six months. I mean, these are clearly useless predictions. Because they’re only telling us what we already know about the rates that earthquakes happen. But there’s a lot of statements that can be made that might sound like they’re more narrow, and in fact, they’re really not, given what we know. And for example, if I say there’s going to be a magnitude greater than 5 earthquake in southern California in the next year, that might sound like a fairly narrow prediction. In fact, there’s a pretty good likelihood that that’s going to happen. So I can make that prediction. And so some of the prediction schemes that come up, and they tout high success ratios, in fact, all they’re doing is just shooting fish in a barrel. And this is – I should say prediction is a huge topic. And so I’m just going to sort of pick a few of the notes to hit. Okay, so the challenge in earthquake prediction sort of fundamentally is, I think, that we really don’t know exactly why and how an earthquake happens – why it starts. We understand the basics. We know there’s plate tectonics. The plates are moving. The plates lock up. Eventually, the stress is released, and you have an earthquake. But why 5:00 on Tuesday and not 8:00 on Thursday? What happens on the fault that causes it to let loose at that particular time? Is there a process that builds up before the earthquake? There’s some ideas, but we’ve never – we don’t really have a good understanding. So if you want to predict earthquakes, it comes down to identifying a precursor. You know, something from the Earth. Some signal that happens that says, heads up, an earthquake is on the way. So – and a precursor – whoops – it has to be something that happens before most big earthquakes and something that only happens before big earthquakes. And so far, a lot of possible precursors have been explored, and nothing has been identified as reliable. Okay. So you’re looking for something in the Earth that happens before every or most big earthquakes, that only happens before big earthquakes. And there’s this distinction that sometimes I think scientists aren’t clear about that there’s precursors versus useful precursors. Sorry. And what do I mean by that? You may have heard stories about anomalies that were observed before certain earthquakes. Right? There was a story in Japan in 1995 that the geochemistry of groundwater at a beer bottling – at a beer distillery changed a few hours before the earthquake. L’Aquila, Italy, there was a supposed radon anomaly. It’s possible that those signals did have something to do with the later earthquake. The problem is, that’s not necessarily a useful precursor. When people looked at radon anomalies back in the ’70s, and they looked at it systematically. You know, you can’t just find an earthquake and look back and say, oh, here’s a radon blip. That’s the game people tend to play. If you measure radon systematically, you find it blips all over the place. And then you find earthquakes all over the place. And it’s just if – maybe radon has something to do with some earthquakes, but it’s not useful. It’s not reliable. And so there is that distinction. Okay. I just wanted to touch on this because it’s a common bit of lore – enduring lore. Can animals predict earthquakes. So this is Charlie. He’s my cat. And he was named for Charlie Richter. [laughter] And it’s a little-known fact that Charlie Richter was a real cat freak, after my own heart. And he had some issues, personality-wise. And my Charlie has issues. But, okay, Charlie, like a lot of cats, sometimes acts bizarre for no apparent reason. And I’m guessing that, if you have cats or dogs, right, they sometimes act bizarre for no reason because that’s what cats and dogs do. If there’s an earthquake tomorrow, some number of cats and dogs is going to have been acting weird today. And people are going to look back and say, my cat was, you know, off the wall. And he knew the earthquake was going to happen. Well, no. He was just being a cat. And so that bit of – you can’t – if you look back for things that happened after the earthquake happens, it’s very easy to find things that look like precursors or anomalies. And that is where this – as far as we know, where this bit of lore comes from. And again, you know, there may be more going on on the planet than we understand. Maybe there are precursors that we don’t know about that animals perceive. That’s possible, but as far as we’ve been able to demonstrate, if you try to look at animal behavior rigorously, they’re not able to predict earthquakes. So Richter was, in fact, an ardent skeptic and critic of earthquake prediction. He has this famous quote. Only fools and charlatans predict earthquakes. And, at the time he said this, back in the ’70s, was sort of a heyday for earthquake prediction. A lot of top scientists really felt like reliable prediction was just around the corner. There was an apparent successful prediction in China. There were promising results coming out of the Soviet Union that a method that had been developed that seemed to be promising. And there were some politics that came into play. You know, this was ’75. Mid-’70s was the heart of the Cold War. And, you know, the Chinese were ahead of us. The Russkies were ahead of us. We needed to catch up. But there was also a sense among scientists and some of the people that work here that great strides were being made to understand how earthquakes work. And people really did think that prediction was on the horizon. And so, in the midst of that, Richter, towards the end of his career, was a notable skeptic. And I think, in time, his position has been kind of supported – vindicated. He also had another quote that I find interesting. He was talking about some of the people who make earthquake predictions that – it’s sort of a long quote, but he’s saying that some of them are mentally unbalanced. But most are sane. What ails them is exaggerated ego plus imperfect or ineffective education. So they have not absorbed one of the fundamental rules of science – self-criticism. Their wish for attention distorts their perception of facts and sometimes leads them on to actual lying. You know, those are pretty strong words. One of the things that I’m going to – points that I’m going to make later in my talk is that scientists will sometimes say, well, that’s the fringe element. We’re the mainstream community. The difference between the two isn’t as great as we like to believe. And I’m going to show some results from a few colleagues that this quote comes to mind. I assume they’re not listening to this. [laughter] But if they are … Okay, so let me go through a couple of notorious earthquake predictions that have been made. And the first one was really kind of a fringe-type thing. Does any remember Iben Browning? In the late ’80s, he had – yeah, a few hands. He had claimed to predict the Loma Prieta earthquake based on the same theory. And right after Loma Prieta happened, he started getting – talking about a big earthquake hitting the central U.S. – the New Madrid seismic zone. And it was going to be December 5th, I think, of 1990. This ended up getting a lot of media attention and a whole lot of anxiety was garnered in the region that people really were worried that this earthquake was going to happen. This prediction was based on the idea of syzygy – that tides have something to do with the occurrence of earthquakes, which is another old idea. Tides may have something to do with earthquakes, as the tides cause the – the moon and the sun cause the tides in ocean, but they also cause stresses in the solid Earth. And it’s not outlandish that those stresses have something to do with the triggering of earthquakes. But here, again, when people try to look at this rigorously, it doesn’t pan out of a strong correlation. Also, if you think about it, you can’t make a prediction every time there’s a full moon. You know, it’s just – it doesn’t – hm?

- Plus or minus two weeks.

- Yeah. Well, there’s – I think Richter had a quote about that, that he observed that earthquakes always happened within two weeks of a – plus or minus two weeks. [laughter] And you might remember, also, just in March, we had this historically big full moon. And there were some worries about megaquakes being triggered. Well, we just had a megaquake, but it was, like, 10 days before the full moon. And the moon was sort of unremarkable at that time. So, in this case, the story got out of hand, as this cartoon shows. Some people said it’s the media’s fault. If you look back at it, arguably, the scientific community was a little bit complicit in feeding the frenzy. You know, scientists and emergency managers, you know, we’re concerned about earthquake hazard. We want people to take it seriously. And it can be easy to – you know, you try to say things, and it’s hard to get people’s attention. So there is this tendency to ramp up, ratchet up, the rhetoric. And sometimes that can have consequences. Okay, another prediction – I want to talk about a couple of predictions in the past that actually came out of the mainstream scientific community. The first one is before living memory, but this is the first prediction scare that I’m aware of in California. This is Bailey Willis, who was an accomplished geologist. He spent most of his career at Stanford. Accomplished and very interesting character. He got interested in earthquakes, and in the early to mid-’20s, he came across some early surveying data that seemed to show that a lot of stress had built up across the San Andreas Fault in a very short amount of time. So, based on that, it was sort of – it looked like an alarming result, and he came out and said that Los Angeles will likely experience an earthquake larger than 1906 in three to 10 years. And then the media turned that into somewhat more alarming statements. The bottom one, in Time Magazine, Willis declared that Los Angeles will be wrenched by a tremor worse than San Francisco in the next 10 years. And so I’m going through these in part just because they’re interesting and in part because it introduces some of the science behind earthquake prediction. So, in this case, the prediction was based on the idea that the ground – you can see strain in the Earth build up before an earthquake. So this cartoon shows how you have a fault in the Earth, and the details aren’t important, but as we – we understand that, as plates are moving, you build up stress around the fault. We know this happens. And you can actually measure this now with GPS. And then eventually, the earthquake happens. And the idea is that that’s a slow process, but maybe right before a big earthquake, the buildup of stress ramps up in some way that you can see. And that’s what Willis thought he was seeing with this result that he had – essentially precursory strain. It’s an idea – this is a really crummy figure from Japan that I just show in part to make the point that it is an old figure, but this panel – it’s old leveling data. And they said, okay, strain is building slowly, and then it ramps up with the observation, and then there was the earthquake. So this idea has been out there. There have been results that seemed to show it. What’s happened is, all of the data that seemed to show a signal like this was very early data with a lot of uncertainty. And, as the data have gotten better with GPS instruments, with strain meters, the closer we’re able to look, the less we’re able to see. And so it just doesn’t look like faults ramp up in the warping that you can measure before an earthquake happens. A more notorious case of that you may have heard about. It played out, again, in southern California in the mid-’70s. And it was, again, based on early surveying data, and, again, another early figure. But people were doing surveying and measuring, you know, changes in the landscape. And – okay, raise your hand if you’ve heard of the Palmdale Bulge. Yeah. Okay. So that is within recent – within living memory. So there was leveling done across the San Gabriel Mountains. And there were actually scientists here who analyzed the data. And so this is – if you go over Highway 14, you can see Lake Palmdale. And the San Andreas runs along there. There’s the famous road cut where the sediments are all mangled. But the fault basically – it makes that low scarp that you can see. The town of – city of Palmdale is just over the ridge. And so this observation seemed to show that a part of the desert floor had risen by about a foot. And that looked like a really ominous signal like strain was building up. And there’s more scientific details that I don’t have time to get into. But this got a lot of attention. And, again, Time Magazine – the Palmdale Bulge could be an early warning signal of a major and potentially disastrous earthquake. There was a lot of concern. And, again, what happened was that scientists looked carefully at the observation and the analysis, including some scientists here and elsewhere. And what they showed was that there were huge uncertainties in the early leveling data. If you do leveling over a big mountain range, there’s refraction errors and other problems. And when you take that into account, the bulge turned into a souffle. It basically – the whole scary signal kind of disappeared. So, you know, that one went away. The one that’s more interesting and that you can’t really dismiss is with the Haicheng earthquake in ’75. And there was a good technical article written about this in 2006. This was at the end of the Mao era. China was big on citizen science. There was sort of a charge to people to predict earthquakes. So there were a lot of people who were, you know – who were making amateur observations. China was looking – the official – at an official level at a couple of different regions in the country. And north of Beijing in the Haicheng area, there had been seismic activity that started in the late ’60s with moderate earthquakes. And it’s a fairly quiet part of China, so the scientists were looking at that, and there was the citizen science observations. And there was an escalating sequence of foreshock activity. And some earthquakes do have foreshocks. Some don’t. In this case, there were moderate earthquakes, and then a few months before, there were a couple of moderate earthquakes. And then, in the end – in the last 24 hours, there was a burst – a very energetic burst of foreshocks. And it was never really a high-level prediction that scientists evaluated, but it was more grassroots. There was a – essentially a prediction made at the grassroots level. And the reasons it were made – was made aren’t really defensible. But it was really just concern that people were feeling a whole lot of earthquakes in a short amount of time. And this one local official, you know, was telling people, this means and earthquake is going to happen. And so there were actually evacuations. There were accounts of strange animal behavior – of snakes coming out of hibernation – it was February in northern China. It’s hard to know what to make of this, but it is – this was an earthquake that had a pronounced precursory sequence. It did have all these foreshocks. So something was brewing, you know, that – prior to this earthquake. And we really don’t understand fully what was going on. But one thing is clear is that this was an unusual earthquake in this protracted sequence that led up to it. And whatever was going on here isn’t typical. Just a year later, another earthquake hit not far from this – the Tangshan quake. It’s one of the deadliest quakes in history. It was not predicted and killed at least 250,000 people. So, through the ’70s, there was this what I call a heyday of prediction when people really were optimistic, in part based on Haicheng, in part based on these Soviet results and other things. And that – as people started to do more rigorous work and to look at these precursors, time and time again, they didn’t hold up. When you looked at the data rigorously, the signal went away. Or when you looked to see if things were repeatable, they weren’t. And then another thing happened was, big earthquakes were happening in places where there was no reported precursor. So the ’70s gave way pretty quickly to sort of a hangover where prediction got sort of a bad name, and the mainstream community backed away from it. And so it was at the point where the mainstream scientific community was pessimistic about prediction. And what happened was, the people who were pursuing it tended to be outside the mainstream. And so prediction got an even worse name. But prediction-related research has continued – does continue. One example of, I think, you know, good research that was done involves what’s called accelerating moment release, which is a lot of – well, it’s a mouthful. But moment is basically energy. So the idea is – and you may have seen this. This is what has been called the tombstone diagram. It’s – time come across this axis. And then each tombstone is an earthquake of a given magnitude in the San Francisco Bay Area. And the observation was that this area – there were a lot more moderate earthquakes leading up to 1906. And then the big earthquake happened, and then things got quiet. And then there was a build-up before Loma Prieta, and things got quiet. And this observation helped lead to this theory that, before a big earthquake happens, you start to see more moderate earthquakes in the region. And it sort of makes sense that, if stress is building on a main fault, you know, it seems conceptually reasonable that you might get moderate – more moderate earthquakes in the same area. And so people started to look for similar in different regions. This is a figure from an earthquake in Australia. And it’s basically a measure of – you know, you can think of energy release in earthquakes. The details aren’t important. But you saw a lot of figures that looked just like this. You know, over a time span of decades, things build up in the region, and then you have the big earthquake. And there were a lot of results, and they really did look pretty compelling. And then this is the worst slide ever, but there was a study done. The lead author, Andy Michael is here. And then Jeanne Hardebeck is here. Karen Felzer is in my office. They were looking at this statistically and saying, could you see this signal as an artifact of the way you’re choosing the data? And they concluded that you could and that – they cooked up synthetic catalogs. The earthquake catalogs that were sort of realistic but had – and then they said, can you find the signal even if it isn’t there? And they show in this paper that you could. And let me try to explain what happens. This is a map of earthquake activity in the last week or so. I just pulled it off the web. And earthquakes – you know, they tend to cluster in time, but then, to a large extent, they’re random. So just recently, we’ve had earthquakes down towards Mexico. There was a quake in the Gulf. This one just today – an event here. You know, are those related? Maybe. We’re not sure. But let’s say an earthquake happens tomorrow. It happens there. Well, I could draw a circle and say, look, activity was building up before that earthquake. Right? But say an earthquake happened there. I could draw a bigger circle and say, look, activity was building up. And what you start to see is that, you know, how did I choose the circle? Well, I sort of chose it to make my argument. So what this complicated paper showed is that, when you’re – if you’re looking back to find the signal after an earthquake’s happened, you have latitude in how big the circle is and how long you take it back in time. And with that much latitude, you can find a signal, even if it doesn’t exist. So that one – you know, again, the tombstone diagram – something like this could be real. But as far as we can demonstrate rigorously, it doesn’t, again, hold up. So, okay, that’s that. But it was a case – I mean, this was good science that was done. And the biases that come in are very subtle, and unraveling the issues took a lot of work. Another example that is – sort of gets into the bad and ugly is called pattern informatics, which is another mouthful. And this has generated bona fide media releases from none other than NASA in recent years. And this was October 2004, which was – that was after the Parkfield earthquake, or was since – it was after – which one. One of the moderate earthquakes. And if you – if you can read this, it talks about how this forecast method has accurately predicted the locations of 15 of California’s 16 largest earthquakes this decade. Wow, that’s impressive. Including last week’s tremors. This method was developed by researchers at University of Colorado, now at Davis, at NASA, da da da. Okay, what this technique is about. This slide gets really weird. When I – can you see what happens? The colors fade out. But, okay, these researchers came up with a way to look at earthquake activity and make these maps that have these measle spots that – they’re disappearing measle spots because of PowerPoint, but … [laughter] It’s really weird. I’ve never seen PowerPoint do this. But so, they have these maps that are supposedly areas of heightened likelihood. And then, if a – if a earthquake would happen, like one of these blue circles, and it falls kind of close to a measle spot, they say, oh, look. We’ve predicted the earthquake. Well, they haven’t predicted anything. They’ve got half of the state of California covered by these blotches. And if an earthquake – you know, this is – and this gets funded by NASA and touted as an earthquake prediction success. So this is, I think, an example – a good example of bad research that’s been done, but fairly recently. One – another example that’s sort of in between good, bad, and ugly – it’s called earthquake chains. And I’m not sure – this got a lot of attention in southern California because a prediction was made in 2004 that actually got a fair amount of media attention. But this is a group at UCLA led by Keilis-Borok. And the idea here is that you get these – what’s called earthquake chains. You get a series of little quakes that sort of stretches out over a certain distance. It sort of looks like a chain. And there’s – exactly what a chain is is defined. And so, if you see a pattern like this, then the method looks around the area to see if earthquake activity has been building up. So there’s two parts of it. And if both of those things are true, then you get a alert. And they’re fairly big regions. And this was the one – there was a prediction made in 2004 that covered much of southern California –  magnitude 6.4 or greater. And people actually calculated there was a decent chance that this earthquake might happen just by chance. In fact, it didn’t. And, as this group has made predictions, they’re getting a lot more misses than hits. And if you look at the statistics, it’s not significant. And it’s the same sort of issue as the accelerating moment release. But if you – the way you develop a prediction method is you find the big earthquakes that have happened, and then you look back at the data. And you try to find patterns that are significant. And you find patterns that look significant, but then you try to predict the earthquakes that haven’t happened yet, right, which is what we’re trying to do. And the methods don’t pan out. And what’s happening is that you’re identifying just these patterns. You know, it’s like these correlations that you sometimes see that, when the NFL wins the Super Bowl, a Republican candidate wins the presidency nine times out of 10. They’re just completely spurious, but they’ve just happened to work out that way. And you can always find that sort of pattern, you know, out there in the world. And you find it. You think it’s significant. But when you start to play the game forward, it doesn’t work. So will prediction ever be possible? There are some seismologists who are pretty adamant that earthquakes are fundamentally unpredictable. In this quote – let me check the time. This quote is actually, nothing is less predictable than the development of an active scientific field was an answer to this question by none other than Charles Richter, who was the biggest skeptic. And I side with this. If we don’t understand what’s going on in the Earth before earthquakes happen, we can’t say earthquakes will never be predicted. There may be things that come up a year from now that we never thought about. So what about this idea that, well, earthquakes are overdue. You know, the southern San Andreas is waiting to explode. What about that idea? And that’s something that scientists tend to say – if you look at the San Andreas – we had 1906, northern California, 1857 in central California. The southern San Andreas hasn’t had a big quake in at least 300 years. And people like to point to this and say, overdue. The Hayward Fault you may have heard about. [laughter] You may live on top of it. I used to live – this used to be my alma mater. It is my alma mater. [light applause] Go Bears. [chuckles] Yeah. So, you know, the 1868 earthquake, you’ve – I’m sure you’ve heard about. And eventually – and there’s been really good work done led by Jim Lienkaemper, who is here, looking at the Hayward Fault. You can – if you look back at the earthquakes sequentially that have happened, and they have dates attached, you see a fairly regular-looking sequence. If you look at the last five that are the best-established, and you – they happen every 140 years on average. The last one was 1868. You do the math, and you get a little worried. But if you look at the full chronology that you know about, and you draw the line, and you, you know, allow for the uncertainties, what you find is that the next quake could be tomorrow. It could be tonight. It could be 50 years from now. And that’s – even for a case like this where an earthquake is fairly regular, there’s a lot of variability in the system. And it’s … [laughter] It’s like clockwork in geologic times, but in the terms that human care about, it’s very irregular clockwork. And, in geologic terms, if the next quake is tomorrow or 50 years from now, it’ll still fit on that plot. It’ll still look like a very regular sequence. But it makes all the difference to us in the room – personally, at least, between now and 50 years from now. So why aren’t earthquakes more regular? You know, stress builds up on a fault, stress is released. Why doesn’t that happen like clockwork? One possibility is that earthquakes tend to cluster. For reasons that we don’t fully understand, earthquakes tend to be clumped in time. And a useful analogy – completely different science – is traffic. So – and it’s one that we can all relate to as Californians. But suppose you’re in the median strip along this highway. And traffic, like earthquakes, tends to cluster for other reasons. But suppose you’re told – you’re given a stopwatch, and you’re told that cars pass by, on average, once a second. And suppose you’re standing right back here, and you’ve just seen two cars whiz by in less than a second. You may think, well, I’m safe. You know, I’ve had two earthquakes. Nothing’s going to happen. In fact, you’re in the middle of a cluster, and the next car is going to come by real soon. Whereas, if you’re somewhere like here, it’s been two seconds since a car’s gone by, and you might be thinking, it’s overdue. It’s going to be any nanosecond. And in fact, you’re in the middle of a gap. So you may have even longer to wait. And so that may be what’s going on with earthquakes and why we really can’t say overdue even though it seems like it’s been a long time. And the problem – okay, before I get to that. You know, so there has – people have been looking at the southern San Andreas Fault for a long time and itching to say, overdue, overdue. This is a quote from Bailey Willis in 1925. Within 10 years, there will be an earthquake bigger than San Francisco – 1925. 1969 – we are predicting another massive earthquake certainly within the next 30 years, most likely within the next decade. Talking about California, so ’69. ’92 – a colleague and a – you know, and a first-rate scientist, Al Lindh. I don’t mean to pick on him. But he was quoted in the paper saying most of us have an awful feeling that 30 years is wishful thinking. He’s talking about the next big one on the southern San Andreas. Quote from the L.A. Times, 2009, the latest earthquake looks to be overdue, according to recent research. Overdue, overdue, overdue, for going on 80 years now. And [chuckles] – and the problem with overdue is that it has the wrong connotations. It’s – and it’s more than we can say. If a baby is overdue, it’s going to show up within a few weeks. If your – if your mortgage payment’s overdue, you’re going to deal with it in a few months or there will consequences. An earthquake could be technically overdue and still 50 years away. So there’s sort of a problem, I think, with, you know, this drumbeat of overdue, is people start to conclude that we don’t know beans about when earthquakes are going to happen. And it has some consequences. You may have seen Simon Winchester’s op-ed that the scariest earthquake is yet to come. And he was saying, well, we’ve had activity in three corners of the – of the Pacific Plate. You know, we’ve had Chile. We’ve had Japan. We’ve had New Zealand. Clearly California is next. Well, it was scientifically out to lunch for about 10 different reasons. And – but it gets fueled by the scientific community, I think, is the point. That we have to be – I think it’s up to scientists – the onus is on us to not overstate the case or the rhetoric. Because it fuels this kind of thing. And there was one point I wanted to make because people wonder, if you’ve had an earthquake in Japan, it’s all the Pacific Plate. Why doesn’t that trigger earthquakes here? And the point that it’s helpful to make with my hands – but the plates are moving. The Pacific Plate and the North American Plate are moving all the time, steadily. They’re only locked up right in a narrow zone around the fault. So when an earthquake happens, it’s just that narrow zone catching up with the rest of the plates. It’s not like the whole Pacific Plate lurched. So it’s – there’s only that narrow zone that it’s confined to. And so we don’t expect to see an earthquake on one side of the Pacific affecting the other. Okay. So earthquakes are not, as far as we understand right this second, predictable. On the other hand, they are inevitable. And we actually know quite a bit about where earthquakes happen. We know that the crust is split into plates. We know that the earthquakes – the lion’s share of earthquakes on the planet follow the plate boundaries. In just a few examples, the Caribbean wasn’t widely recognized as an active plate boundary before last year, but in fact, it is its own little miniature ring of fire. It has volcanoes. It has earthquakes. Tsunamis. Historical earthquakes – there have been some biggies, but not so much over the last 100 years. So people weren’t really thinking about it. There were scientists who were looking at the Caribbean. This was a paper that came out in 2008 saying that the Enriquillo Fault in Haiti is capable of a 7.2 earthquake if you look at how much strain has accumulated. So this is a – and they weren’t making a prediction. But it’s – this is a case where you said an earthquake is possible, and it happened sooner rather than later. We know that the 2010 Haiti quake happened, and we know the consequences. Japan, we just saw, is an active plate boundary. Lots of big earthquakes. Lots of small earthquakes. The part of the coast – most of the concern in Japan has been for the plate boundaries closest to Tokyo, in part because that’s the big city. But in part because there’s expected big earthquakes there that are considered somewhat overdue. And then the earthquake that we have – this huge magnitude 9 – was along this part of the subduction zone that had been judged relatively less hazardous. And this was the hazard map for Japan showing how hazard had been assessed to be lower here than closer to Tokyo. And we all saw the aftermath of that earthquake. China was one more example. China – the plate boundary is more complicated because India is moving  northward. It’s hitting Eurasia. You’re getting a diffuse zone of earthquakes. And it seems like ancient history now, but they were hit by a devastating earthquake just a couple years ago. In that case, again, if you look in detail at the hazard map, where the earthquake happened was a zone where the hazard was judged to be relatively lower. And, wham, you have a magnitude 8. So we know where the active earthquake zones are. We know where the plate boundaries are. I think sometimes – I think the lessons of the last few years of the Wenchuan earthquake and of Japan is that scientists – we sometimes look in fine detail at plate boundaries, and we say, well, we think this part of the plate boundary is a little less hazardous than this – than this part, for complicated reasons. And I think the Earth is telling us that we shouldn’t be picking nits quite that finely. I just reviewed a paper looking at the subduction zone north of Puerto Rico, where there hasn’t been a great earthquake in the historic record. And the conclusion was that this subduction zone may be weakly coupled. It may not be capable of producing large earthquakes. And my review of this paper was, do you really want to be saying that after Japan? You know, I think the strong likelihood is that we just haven’t seen earthquakes in a fairly short historic record. And if you have an active plate boundary, the assumption needs to be that big earthquakes are fair game anywhere along it. Now, that said, there is a distinction – you know, along the San Andreas, we have the faults moving – the plates are moving laterally. And so, as opposed to Japan or the Caribbean, South America, where you have the subduction zones. And the subduction zones are where we expect to see the really great earthquakes – the magnitude 9s or 9-1/2 – because you have huge faults. They are very long, and they’re very deep. Whereas, the San Andreas is long, but it’s skinny. And that limits the magnitude that you can generate just because there’s only so much fault. So we don’t expect to see a 9 or a 9-1/2 in California. I wanted to throw this in to address the point – people sometimes think that earthquakes are getting worse – we’re having more big earthquakes, you know, than we used to. That’s not happening. The earthquakes happen on their own time and have for millions of years. What’s happened – this figure shows the growth of the world’s cities. It kind of washes out, but the cities have gotten massively bigger, so the earthquakes are happening on their own time, but the targets that they’re hitting are bigger. There’s more to be knocked down and more people. So the hazard isn’t increasing, but the risk to life and limb and structures is. So you are here. You know, we’re all sitting along one of the active plate boundaries. The hazard in California has been assessed pretty carefully. There was a big study that was done, led by Ned Field, the U.S. Geological Survey, that came out a couple years ago. And there’s some bottom-line numbers looking at probabilities of earthquakes anywhere in the state. So probability of a 6.7 or greater anywhere in the state is over 99%. That we can say is nearly certain. 7 or greater, close to certain – 94%. When you get up to 7-1/2, that’s 50/50 odds in 30 years. So the 7-1/2s are relatively infrequent. And the odds of an 8 in the state are estimated at 4%. So we don’t expect to see magnitude 8 earthquakes like 1906 very frequently. But those are sort of the best numbers that we can estimate given all the science that’s known. So, you know, we sort of – the scientific community has moved away from prediction and towards emphasizing preparedness. And you may have seen this excellent publication, Putting Down Roots is a pamphlet that is put out, and I’m sure it’s around here, that’s basic earthquake information and preparedness information. There’s a website that was developed initially for southern California, but it’s pretty generally applicable – daretoprepare.org. This is just nuts and bolts information of, how do you prepare for an earthquake? How do you secure your home, your office – you know, what steps can you take? Because the steps that you take personally within your own, you know, sphere of influence can make a huge difference in what happens after an earthquake. And that’s sort of the bottom line at the end of this, that, as far as we know, right this second, we can’t predict earthquakes. But we can tell you that they’re inevitable. And so, if you’re going to live here, you need to be prepared for something to happen pretty much any time. So that’s the end. And I’ll happily answer questions if you have them. Thank you.
[Applause]

Yes? Oh, it – do you want people to use a mic?

- Yeah. Let’s see. Is this on? Okay. Is this on? There you go. You can hear me. Yes. I think Sue will be happy to answer questions. And some of you already know the drill. I’d like you to please use a microphone so everyone can hear you. We have two microphones set up on either side of the aisles. And if you don’t mind, that gentleman had his hand up first. You. Yes. And then we’ll just go back and forth with the mics. And go ahead and answer – ask your question.

- What is the driving force for these plates to move? Why – because it takes a tremendous amount of energy. Where is it coming from? Why are they moving?

- Okay. Yeah, and I should – I know it’s 8:00. So if people need to leave, that’s fine. Yeah. What’s happening – if you go sort of deep in the Earth, the Earth’s core is radioactive, so it’s generating heat. And then you have – most of the Earth is what we call the mantle. And it’s more plastic, and it’s hot. And so it’s convecting or sort of – it’s like a big soup pot, except it moves slowly. So things are moving, you know, a few inches a year. And the crust is just, you know, like, a brittle layer on top of it. So the mantle is trying to move, and it’s pushing the plates around. And the plates are moving, as I said. They just get stuck at the boundaries. So, yeah.

- So if you had any – if you could have, like, a magic wish, and you could have any measurement that you wanted – unlimited budget, whatever – what measurements do you think you could do that would help you predict earthquakes?

- You know, what I would really love to see is, like, an earthquake catalog from the Earth that’s a million years long. I mean, I would really like to see the patterns that have been played out. Are there any patterns? But things you could measure – I guess one question is …

- I mean, like, drilling down into, like …

- Yeah. No, if – one question is, you know, what’s happening on faults? And is there a build-up of stress or strain? There are – measuring at the surface, you’re really limited. You know, suppose some fault is starting to move slowly. The amount that that’s going to move the surface is so tiny that, with the best instruments, we’re pretty limited. So if you could measure stress or strain very precisely at depth, that would tell you a whole lot. And I use – I’m using stress and strain. I probably should have explained them. Stress is – you know, is – is the – well, not quite a force, but it’s essentially the pressure that’s driving a fault. And strain is the warping that happens in response to stress. But, yeah, that’s one of the big unknowns.

- Let’s see. I think I’m supposed to the …

- Okay. What is the prediction f or the New Madrid Fault? [laughter]

- Oh, boy. You’re going to get me into trouble.

- Yeah.

- No, it’s – and that’s one of my research interests, and I actually just had a paper that came out with revising – with revised estimates for the magnitudes of the earthquakes that happened 1811 and 1812. And I actually estimate them to have been closer to magnitude 7 than magnitude 8. So there’s a big question about how big the earthquakes were. But we had a sequence in 1811, 1812. There was a sequence in around 1450 A.D. and a sequence in 900 A.D. And we know about those from geology. So, over the last couple thousand years, there have been big earthquake sequences. They're about every 400 years or so. But in terms of what’s going to happen next, it’s really not clear in that case. Because it’s not like the San Andreas. There’s no plates that are moving. And we really don’t understand why you have earthquakes at all in the middle of the North American Plate. So there could be more earthquakes – you know, there could be another sequence a hundred years from now. It could be over. We could have seen the last of this little cluster of earthquakes. But that’s – compared to California, it’s just much, much more uncertain.

- And what information have you gotten from Parkfield?

- What …

- Over – like, over the last 20 years. What has it actually told you?

- Yeah. There was the – you may have heard about the Parkfield prediction experiment made, you know, from scientists here in – [chuckles] that was the next question? In the early 1980s, people identified Parkfield and realized that there had been a fairly regular sequence of magnitude 6-ish earthquakes in ’34, in ’66, and they made a prediction of the next earthquake in 1988 plus or minus four years. And 1992 came and went. No earthquake. But then the earthquake did happen. And it’s – I always forget if – 2005, I think. Or 2004. ’05, I think. So the earthquake did happen, and it was on the part of the fault that been identified. It was a magnitude 6. And I think that was a qualified success scientifically that it was identified – you know, and that's sort of the irregular clockwork. If you plot Parkfield earthquakes, you know, it fits the pattern. But it just – it’s the disconnect between clockwork in a geologic sense and clockwork in a human sense. You know, in a human sense, if you say ’84 to ’92, and the earthquakes is in 2005, that’s a pretty big deal. And that’s what happened. But that is a part of the fault that these moderate quakes do happen relatively regularly.

- That was such a refreshing presentation that you gave. It makes one proud to have been associated with the Geological Survey. My question has to do with the North American Plate. It’s really a huge plate.

- Yeah.

- From your pictures and the pictures that are in the papers now. I don’t think many people realize that it just doesn’t include the North American continent. It includes all the way over to Japan and all the way across the top. Is there any chance that, within that plate system, there are smaller plates which have not been identified yet? Like you – like the Caribbean system that you showed?

- Yeah, actually – I mean, that’s – it’s a good point. Because it includes half the Atlantic too. You have the Atlantic coast, and it goes from the continental crust to the oceanic crust. That’s a pretty big transition, but there’s not an active plate boundary there. So that’s why you don’t have as many earthquakes along the East Coast as you do in California. But it is a massive plate. No, people have GPS measurements now, and GPS – you can literally watch the plates moving. And so, if part of North America was moving significantly relative to another part, I think we would – we would have seen it by now with GPS. And, in fact, someone asked about New Madrid. People have looked with GPS trying to find some strain that’s going on. And they can’t find any. The better the measurements get, the closer the strain gets to zero, which really raises the question of why you’re having earthquakes there. But at least right now, the plate seems to be pretty stable, pretty rigid, and there’s no major earthquakes zones. Well, I’ll take that back. Okay. Not – some earthquakes are happening as a result of other forces than plate tectonics. And the one notable example that I know about, at least, involves post-glacial rebound. So much of North America was covered by these massive ice sheets. You took that ice away. The crust rebounds, just like putting your hand on a cushion, and it rebounds up, right? But that goes on over thousands of years. So, up in Canada where it was closer to where the ice sheets are at, the crust is bouncing back. And you can actually see that with GPS. And that’s driving earthquakes that happen along the St. Lawrence in particular. So there is – within the North American Plate, there are stresses that are associated with other things, but it’s not really breaking up the plate.

- Thank you.

- I haven’t read your book, but I did read the flier for your book. And it – and it mentioned – you were discussing in that, earthquake lights. And do you feel that there’s any possibility for the electromagnetic effects that might be generating earthquake lights can become a precursor?

- Okay. I’m smiling because I wrote a book about earthquake prediction, and I did that on my own time. And then I’m giving this talk in my capacity as a USGS scientist. And the ethics rules say that I can’t benefit from – use my official position for personal gain. So I’m allowed to talk about earthquake prediction, but I was very careful not to mention the book because I can’t. But I’m allowed to answer a question. [laughter]

- It’s a fine line.

- No. It is this – you know, this is the position I am in, and it’s – you know, okay, so yeah, there’s a lot more to the story. I did write this book. And I didn’t talk about electromagnetic precursors very much, but that’s a whole realm of earthquake prediction. And one of the interesting observations is, as you mentioned, there are these anecdotal accounts that go way back hundreds of years about earthquake lights that are seen in the sky before earthquakes. And that sort of, you know, maybe suggests there's some sort of electromagnetic precursor that’s going on. And so, yeah, there could be electromagnetic precursors that are going on before earthquakes. And you’re getting earthquake lights. You know, you may be getting other phenomenon – earthquake fog. You know, so those things are possible, and it’s worth – it’s worthy of investigation. As far as people have seen, if you get really good data, there hasn’t been any documented electromagnetic precursor that’s really been shown to be significant. But, yeah, it’s possible that it’s happening.

- Not even Tony Frazier-Smith’s work?

- Okay. That was – so that was the case of the ultra-low-frequency anomaly that was supposedly observed before Loma Prieta. And some of you may have heard about this. There was an article in Science because, by chance, there was a magnetometer in Corralitos. And scientists have looked at that record very carefully, and there’s pretty good evidence that it was some sort of instrumentation glitch that caused the signal. And this was back in ’89. You know, computers are not what they – what they are now. The way the data were – scientific instruments – the instrument wasn’t even able to store the raw data because it didn’t – there wasn’t enough computer memory. So there was sort of processing that was done, and the data were stored. So it’s pretty limited data, and there’s some really strong indications that something went a little haywire with the instrument a few weeks before the earthquake. But it’s never – people have been looking. That sort of launched a cottage industry to look for similar signals. And they haven’t been documented, seen, before other earthquakes.

- Thank you.

- It seems everybody wants earthquake prediction for lots of reasons, and we’re working hard on it, but what happens when we do get earthquake prediction? What’s – would you care to comment on the economic impact?

- Well, it’s – you know, people ask that question. Because it’s a good point. And especially – you know, if I could say – if science could say there’s going to be an earthquake tomorrow, 100% probability, you could sort of imagine dealing with that. Although, you know, you quickly get into these nightmares. You know, how do you evacuate San Francisco? But what happens if you can say that an earthquake is very, very likely to happen in the next six months? You know, what do you really do about that? I think if scientists could predict earthquakes, they’re going to predict earthquakes. And then, you know, just the consequences will play out as they – as they may. But, yeah, I mean, it raises some interesting questions. We think we want prediction. You know, we want a heads-up that an earthquake is going to happen. But, you know, if we get to that point, it’s going to be a challenge to deal with.

- Okay, thank you. I’m surprised you didn’t mention Billy Meier’s prediction about a large earthquake in the land of cherry blossoms, about 35 years ago.

- Okay. That’s …

- Anyways, it’s another one of these …

- That’s one I don’t know about.

- Okay. You should, you know, Google that.

- This is for D.C.? Or Japan?

- The land of the cherry blossoms. So it’s pretty wide. [laughter]

- Nostradamus.

- So, but anyway, my real question is, I do understand, in Japan, they have some sort of earthquake sensors or detectors that maybe sense some P mode oscillations that give you at least a 60-second – perhaps you can talk about the geology of the P modes versus the S modes.

- Yeah, there’s …

- And the time difference to – I guess depending on how deep it is.

- Yeah, Leslie and I were talking about this. There may be, in the future, a public lecture would be on earthquake early warning, what we call it. Early warning is different from prediction. And it sounds like you’re familiar with this, and maybe a lot of you are. If an earthquake happens, the P wave – well, okay. An earthquake happens. It generates waves. So there’s the P wave. The S wave travels more slowly. But the thing is, if an earthquake happens, and I had a seismometer sitting here, and I could tell that there was a magnitude 8 earthquake, I could send a signal at the speed of light to you and say, an earthquake just happened here, before either the P wave or the S wave get here. So it’s the same delay between thunder and lightning. And you can get – so if you have to observe an earthquake, figure out that it’s going to be big, get the warning sent out, and you can get yourself, you know, seconds, tens of seconds, of warning, potentially. Japan does have a system. They stop the bullet trains. Mexico had a rudimentary system. I heard that it actually managed to kill somebody with a false alarm, and somebody jumped out of a window. [laughter] But it’s – people are working on it – people at Berkeley, at Caltech, USGS. It’s – the science is sound, but it’s a huge technological challenge. A big earthquake is going to last for two minutes, maybe. The 1906 earthquake – the fault was actually in motion for two minutes. Well, you can’t wait two minutes to determine the magnitude because you’re – the warning would be over. So you have to look at the very first few seconds of a seismogram and know it’s going to be a big earthquake, and that’s tricky. Because the beginning of a 7 may not look that different from the beginning of an 8 or a 5. So that’s one challenge. And, you know, pulling the data together quickly to look at data from different seismometers – it’s a big – it’s a challenge, but people are working on some fairly clever approaches. And, you know, it may be – it may come to pass in California in the next 10 years or so. But you’re never going to be able to count on a lot of warning because it really works – you need to have distance between the fault and where people are. So it works pretty well in Japan. The big faults of offshore. The people are some distance away. Here, you have people and faults and faults and people. And, you know, the San Andreas just isn’t that far from here. The Hayward isn’t that far. So if you’re too close to where the earthquake happens, you’re not going to – there’s not going to be time for warning.

- Thank you.

- Didn’t the Japanese earthquake recently last more than two minutes? I think it’s – at least some reports seemed that it lasted an immensely long time, considering.

- There are – there’s two different things. One question is, how long was the fault actually moving? And when a big earthquake happens, it’s not like the whole fault just lurches at once, right? It’s like, if you had to pull a long carpet, and you put a ripple into it, and the ripple travels down, and you can move the carpet on the floor. So that’s how big earthquakes happen – like a ripple. And the ripple goes down the fault about 3 kilometers per second. So if an earthquake is – how long the earthquake is sort of gives you a pretty good idea of how long it takes in time. Sorry. That was a little garbled. But the Japan quake was about 300 kilometers long, so the fault was probably moving for about two minutes. The earthquake in Chile in 1960, it may have been closer to five minutes. That’s the duration the fault is moving. But then, what you feel is much longer because, as the earthquake moves, it sends out waves. The whole time the earthquake is going on, you get the P wave, the S wave, other waves. They bounce around. So you can easily feel five minutes of strong shaking from an earthquake that lasted two minutes. And that – you think about it – you know, if you felt Loma Prieta, and it felt like it lasted forever, that was a 6.9. And, you know, a magnitude 8 would last many, many times longer than that.

- Before I get to the question that I wanted to, what you just said made me think of something. When you think about the origin of the – what was actually starting – what actually started that earthquake that lasted, let’s say, one minute, was – that one minute, is that all the reverberations that you heard? And the actual earthquake was only five seconds? Or is it really moving, you know, in one direction for a whole one minute?

- Yeah, so you’re talking about Loma Prieta?

- Or just an arbitrary earthquake.

- Yeah, so Loma Prieta was pretty small. That earthquake was probably over within 10 seconds – the motion on the fault. But, I mean, how many people here felt it? And … [laughter] Wow. I was in New York. Do you remember the study that – about the Nimitz Freeway and that it collapsed where it was built over mud? Okay. That was my study. I was a postdoc at Columbia. And I came out to study the earthquake, but I didn’t feel it because I was a long way away. But, no, I felt the Northridge earthquake in southern California – similar magnitude, similar duration. And, yeah, I mean, so how long was the duration that [audio cutting out] of shaking that people felt, for example. [inaudible]? [very quiet audio] I mean, it’s going to vary depending on where you were, so …

- I was on the second floor of a building in South San Jose, and I – after it was going a long time, I figured, I got to time this. I timed it for about 15 seconds after that, so it was probably 30 seconds total.

- Yeah. So it’s going to vary. And depending on the geology – if you’re in the middle of a basement, it’s going to slosh more or [inaudible]. So, yeah, it’s – what you feel is going to be a lot longer than the actual earthquake on the fault.

- The question I really wanted to ask was, do you see any great scientific value to the amateur seismographs that are increasingly being deployed and connected?

- Yeah. I think, you know, making observations is great. And, you know, I talked about how the gulf between the amateur community – you know, you can call it fringe – and the mainstream community, that gulf – and I concluded this working on the book – isn’t as wide as we like to think. There’s some interesting work being done in the amateur community on any number of [normal volume] aspects of seismology. Collecting data is almost always a good thing. There’s a new project that’s being led by a woman who’s going to be joining the Pasadena office to develop very low-cost seismometers. I think this is really exciting. Okay. If you – if you have an iPhone, do people know about – or a Mac laptop – do you know about SeisMac? Or let me just make sure I have the name. For an iPhone, you can look for iSeismo. So there are – there’s an accelerometer – there’s basically a seismometer in this phone. And there’s – there are little, tiny seismometers that have been developed for commercial applications, for airbags, for video games, for iPhone. The way the phone knows what direction it is is by measuring the acceleration. So there’s actually an app. I can turn my iPhone into a seismometer. It’ll measure – if you can see – so just look for iSeismo. Download it. It’s free. It’s really cool. [laughter] So if you’re a geek, this is, like, the best thing ever. [laughter] So there’s – there’s – okay, for a Mac laptop, there’s an app called SeisMac – M-A-C. If you have an iPhone, it’s, like, iSeismo. And they’re both free. And I actually – I was riding an elephant, and I was holding my seismometer to see how the – and then you can actually do some basic analysis of this. So – okay, people realized that, well, I could put this down on the ground, and poof, I have a seismometer. You know, so you could have a seismic network that’s made up of everybody’s iPhone. Well, that doesn’t work so well because most of the time, my iPhone is in my pocket. You know, but then people had the – the next idea was to take these itty-bitty instruments and build them into a scientific – a better housing to use for monitoring earthquakes. They’re dirt cheap. You can get a little instrument for 50, 80 bucks. And they’ve designed the housing. You plug it into a computer through a USB cable. You install some free software. You have a seismometer in your house for 80 bucks. They’ve developed the software that’ll gather the data kind of in the background. If an earthquake happens, it gets beamed to home base, so …

- This is iSeismo. If anybody can see the screen here. It’s jumping up and down.

- Yeah.

- [inaudible] you can see the lines moving. You can see the lines. [inaudible] seismographs out there.

- Yeah, so …

- [inaudible] you can steady it on the chair and then stamp your foot on the floor, you can see it. [laughter]

- Yeah. Oh, no, there are – you can measure the vibrations of an elephant. I mean, it’s just the coolest thing ever. But I think there’s some really exciting possibilities. And I just put together a proposal. You could deploy 100 instruments in Kathmandu for, you know, a relatively cheap amount of money and monitor the earthquake shaking. You could have thousands of instruments in the Bay Area. So I think, as this technology is exploited, there’s going to be some really exciting opportunities for scientific studies, I think.

- One of my friends in Los Altos Hills has a seismometer in his basement that’s bolted to the floor. He says it’s hooked up on the net to a master system.

- Okay.

- Do you know many of those are – are there around here?

- Yeah, a conventional seismometer may cost on the order of $5,000 to $10,000. So there are – there are instruments – there’s probably, in total, a thousand of them in California. There was an instrument developed recently called the NetQuakes. It’s a little bit cheaper. It’s more like 4,000, and there was a program to deploy those in people’s houses. I think the number of those is a few hundred. But it still – I think the – you know, using this kind of technology is going to be a game-changer. Because that’s what’s going to make this affordable to really, you know, get into monitoring on a much bigger scale.

- My next question, though, is, my biggest fear, really, is shutting off natural gas. And does PG&E have automatic shut-off in a major quake?

- Yeah. Well, I don’t – I’m sure they have systems for their main lines. But if you – I have a little box on my gas line at home. I bought it through the gas company, and they were – they billed it in a few payments spread out over my gas bill.

- It’s hooked up to the valve? Is that it?

- Yeah. And it’s basically a little strong motion seismometer. And if a big earthquake happens, it’s designed to shut the gas valve off. Because that is a concern. If your house is well-built in its wood frame, bolted, it should be in pretty good shape. But if you’re – if the gas line ruptures, fires are a big concern, so …

- Of course, my insurance friends say, if you have a major quake damage to your house, you really want it to burn down because the fire insurance will cover it. [laughter]

- Yeah, but do you really want your house to burn down? So I would urge you, if you have natural gas, look for these little devices. They’re not that expensive, and it’s a really good investment.

- Quick question. What is the method for figuring out the dates of pre-historic earthquakes? Is there an easy answer for that one?

- Ah. It mostly – if you find geologic evidence – so you look along a fault, and you can see breaks that happened in the past. So, like, along a creek bed, if a fault cuts a creek, and sediments are coming down, an earthquake happens, it moves the sediments, then more sediments come down. And if you date – you can find the cracks, and you date with carbon-14 the sediment layers, you can sort of pinpoint the earthquake. That’s basically how the game is played. Sometimes there’s, you know, historic accounts if it’s within the historic record. That gives you the precise date.

- I want to thank all of you for your great questions and for coming tonight. [Applause] And, of course, I want to thank Sue Hough. It was a wonderful talk. I think everybody enjoyed it. It’s 8:30, and you’re still all here. That’s a good sign. So thank you very much. Thank you, Sue. And a lot of people asked about things like the NetQuakes system and about early warning, and I’m pretty sure, within the next year or so, or sometime next year, we’ve got some lectures scheduled on those subjects. So stay tuned, and I’ll see you next month.

[inaudible background conversations]

- Thank you, Leslie.