University Place

cc >>

Tom Zinnen

Welcome, everyone, to Wednesday Nite at the Lab. I'm Tom Zinnen. I work here at the University of Wisconsin Madison Biotechnology Center. I also work for UW Extension Cooperative Extension, and on behalf of those folks and our others sponsors, Wisconsin Public Television, Wisconsin Alumni Association, and the Science Alliance, thanks for coming to Wednesday Nite at the Lab. We do this every Wednesday night, 50 times a year. Tonight's a rather special night because about 50 weeks ago we had the same speaker to talk to us about the Japan earthquake and tsunami within about four weeks after it happened. And due to requests from many WNATLers, I'm delighted to have Harold Tobin back to talk about the great Japan earthquake and tsunami of 2011 one year on. Harold is a professor of geoscience at the UW Madison. His research specialties are in marine geophysics and the processes of earthquake faulting both on land and beneath the ocean. Tobin is a veteran of 10 ocean-going research expeditions and is currently serving as chief project scientist for the Integrated Ocean Drilling Program's NanTroSEIZE project. You'll probably have to pronounce that for me. >> I'll take care of that. >> Okay. An international collaboration to study faults beneath the ocean that cause massive earthquakes and tsunamis off the coast of Japan. It is the largest scientific ocean drilling project in history. He holds a BS in geology from Yale University in 1987, a PhD in Earth sciences from the University of California Santa Cruz in 1995, and after a postdoctoral research stint at Stanford University, he held a faculty position at New Mexico Tech for nine years before coming here to Madison in 2006. Please join me in welcoming Harold back to Wednesday Nite at the Lab.

APPLAUSE

Tom Zinnen

>>

Harold Tobin

Well, thanks very much, Tom, for that kind introduction, and this kind of unseasonably warm weather, back when I was in Santa Cruz people talked about as earthquake weather. It was a very hot day in 1989 when the Loma Prieta earthquake happened. It turns out there's no correlation between temperature and weather and occurrence of big earthquakes, but people still talk about it as earthquake weather so let's hope we don't have one here in Wisconsin.

LAUGHTER

Harold Tobin

Fortunately, the chances are actually pretty small. In fact, I was excited a few weeks ago when there was an earthquake that made the news. Maybe you heard about it. About a magnitude 2.4, which is, essentially, barely perceptible. And it was right on the Illinois/Wisconsin border. And I was kind of hoping that it would turn out to be in Wisconsin because we've never had a cataloged earthquake in Wisconsin since the USGS has been keeping records. It was actually in Illinois, not in Wisconsin so we still haven't had one. But parts of the world do, and so what I'm going to do is talk to you today about the Japan earthquake of a year ago with the perspective, essentially, of kind of what we've learned about it over the span of that year since it occurred. The anniversary was just a few days ago. There was lots and lots news coverage that maybe some of you saw. And the news coverage tends to focus on some fairly specific things. They tend to focus on the devastation in part. So statistically this we've been hearing a lot essentially nearly 20,000 people either died and are confirmed dead or have gone missing since the earthquake, very sadly and unfortunately. In fact, more than 300,000 were left homeless at that time, many of whom are still to this day living in temporary housing or temporary shelters. Just massive amount of damage to infrastructure. We've seen a lot of news about how rapidly Japan has done a very good job of clean up and reconstruction in many cases, but that is just still a road that they're really just beginning on. And of course the biggest news in some ways over the long haul post-earthquake is the ongoing crisis at the Fukushima nuclear power plant, Fukushima Daiichi in particular. And I'm actually not going to talk about that or most of these things very much at all today. I'm going to focus on the science, the seismology and geophysics or geology of what happened in the earthquake, and to the extent that it was a surprising event that's caused us to really rethink a lot of things that we know about it. I will say that the nuclear power plant crisis is an ongoing one. They've got it partially under control in terms of reducing the temperature, but they're still not able to go in there. Soil is contaminated around the region. There's still, I think, a 20-kilometer radius total evacuation zone, and it's not really clear that anyone will be able to live there indefinitely. Nobody has said when people might go back in, and it's really not clear that even for generations it will be possible to do that. So certainly that was one of the biggest outcomes in devastation. Another way to look at it though is that 20,000 people were killed by this earthquake, and as far as we know so far not a single person has died because of nuclear contamination from this earthquake. So it depends on kind of how you look at things. The event, though, the earthquake and the tsunami that it triggered were really truly remarkable events to us and have essentially revolutionized our science in at least certain ways. So I'll try to give you a little bit of a flavor of that. A year ago I talked about what we knew right away after the earthquake, and I'll have to at least sort of re-go over some of that, but I want this to be new. So I'll talk a lot about stuff we've learned and stuff we're doing since then. It's just some of the superlatives. It was, in fact, the fourth largest earthquake we've ever recorded. It was the largest earthquake Japan has ever seen which is one of the most earthquake prone regions on the planet. It was a very large tsunami. It was probably not the largest loss of life in a tsunami for Japan and certainly not for the world. The 2004 Indian Ocean tsunami took something like a factor of 10 more lives, 200,000 or more people. And actually that's largely a function of the difference in preparedness between Japan and Indonesia and Thailand and the regions that are around the Indian Ocean. Also some of the lessons learned from that 2004 event came into play here. From the point of view of science, it's just by far the earthquake and tsunami that we have the most data about. And so for a really major earthquake, we've learned more from this because of simply the kind of data gathered. That's actually mainly because the nation of Japan has just done a phenomenal job of essentially instrumenting the heck out of everything that they can. It's a relatively small country with a very large seismic and tsunami hazard, and so they've done a phenomenal job of putting all kinds of instruments in place. And we'll see some of the records that come out of that. One thing we'll see that I'm going to highlight is that this, for us, was a superlative earthquake in the sense that we knew that earthquakes could cause slip on the fault, and I'll explain that a little more later. We had reasonable documentation or inference from the 1960 Chile magnitude 9.5 earthquake that could be as much 25, maybe even 30, meters of slip between the two plates, essentially instantaneously, at least to a geologist, meaning over a few minutes, slip of the rocks on the two sides of something like 25 to 30 meters. We know that in this event, slip was at least 50 or 60 meters and maybe as much as 80 or 90 meters, and I'll show you that. 80 or 90 meters, over something like, well if it's 90, over 300 feet, a football field length of slip in a single event, a single earthquake. It's just unprecedented. And we've learned that in ways that I'll show you in a few minutes. So what we'll do is we'll explore the earthquake itself, and I'll go a little bit back over some things that if you were here last year your heard about, and the tsunami. And we'll talk about how these great earthquakes happen, and by great earthquakes I mean something specific. To a seismologist, a great earthquake is a magnitude 8 or above on what we used to call the Richter scale and now we call the moment magnitude scale, but the difference is not important for us now. So how those earthquakes happen and how they trigger tsunamis and then what was surprising, as I've said, to seismologists and geologists about this one. And one area I'll try to highlight toward the end is some of the developing research, and that includes two areas that I'll especially highlight looking at the history of tsunamis before we had historical records. So sort of the archeology of tsunamis if you will. We call it paleotsunami or paleoseismology. And then I'll talk about the projects to actually try and go into the fault zone and understand what happened there through the use of this off-shore drilling technology. Okay, well, first of all, we know the region that the earthquake happened on. It was this region of the Earth's plate tectonics that we'll call the Japan trench where the crust of the sea floor beneath the northern Pacific Ocean slides beneath the plate that Japan is part of which, maybe confoundedly for this particular part of Japan, is actually the North American plate. Okay, that's confusing. But this plate is actually continuous through Alaska with the rest of continental North America, at least in terms of how the plates on the surface of the Earth move. So we have to face the prospect that while San Francisco is not on the North American plate, Tokyo is. Well, that's strange but that's dealing with plate tectonics, I guess. At any rate, this is a map of that region and this shows, actually, these circles each show an earthquake scaled by size. The main shock, the main earthquake's epicentral location, I'll explain that in a minute, was here. And then all of these are aftershocks of magnitude, I think, 4.5 or larger, and that was within one week of that earthquake. And they actually fill in the region that slipped in the earthquake itself, which obviously extends hundreds of kilometers both north and south and also east and west from the epicenter or the point where the main earthquake actually began. And even stressed areas outside that slip zone, so off on the other side of the deep sea trench which is actually on the other plate, caused earthquakes as aftershocks as well. That's all within the first week of the earthquake. Aftershocks continue potentially for a long time after and earthquake. That was obviously a map taken from Google Earth. Here is Google Earth running live on my computer right now, and this is the same map, essentially, but now this is showing us the current week, today. And in fact there have been a number of aftershocks. All of these yellow dots are aftershocks. You can see there's more of them there where the 2011 happened than, say, in other regions of some of these other trenches down here that I'll talk about later on. And, in fact, just yesterday there was a magnitude 6.9, which I can zoom in on a little bit. Right up here in the kind of corner of these two trenches. Yeah, right there. It was 6.9 on March 14th. That was last night to us or during the day today in Japan. So these are ongoing. These are all still responses to the stress changes that that earthquake caused as one plate lurched out over the other plate. And they will continue to go on for years, in fact, post-event. Let's see, just a little bit of geography since we've got it up here. This whole region is called Tohoku, northern Honshu Island of Japan, and that's why we call it the great Tohoku earthquake or sometimes the off-Tohoku, off-shore Tohoku earthquake. The entire fault slip area was, essentially, all off-shore or underneath the off-shore portion, and the part of the plates that slipped didn't even extend underneath the land. Of course the seismic waves and the shaking did radiate away from there and caused strong shaking at the coastline. The nuclear reactor is located here on the coast just east of the city of Fukushima, and the worst of the tsunami inundation was basically between that point and up here about where Iwate is located. And we'll see some things about that tsunami inundation a little bit later on. Alright, so let me stop there and go back to PowerPoint here. Alright, well, I talked about this a little bit last time, so I'll just be brief. Just to give you the plate tectonic setting and feeling for what these great earthquakes are, earthquakes can happen anywhere. We had a small earthquake on the Illinois/Wisconsin border last week, but really the vast, vast majority, in fact, more than 90% of all earthquakes and more than 96% of all earthquake energy release in terms of how much just physical slip and energy release happens in earthquakes happens at very specific places on Earth, and that is at the point where two of the Earth's tectonic plates contact each other. That includes places like the mid-ocean ridges. Like here down the center of the Atlantic maybe you've heard about the mid-Atlantic ridge, and this map is a 10-year map about all the earthquakes above magnitude 5 on the whole plant. First of all, there's a lot. There's thousands of earthquakes plotted on that map. And secondly, they mostly don't happen out in the middle of plates, except for a few scattered ones, either the ocean plates or the continental plates. They mostly happen along the boundaries. In fact, that's how we define the boundaries of the plates is where the earthquakes are concentrated. And even within the plate boundaries, not all are created equal. The mid-ocean ridges tend to have relatively small earthquakes that are quite shallow in the Earth's crust, just the upper few miles. Whereas, places that are places where two plates are not spreading apart but are converging on one another, which we call subduction zones, where one plate has to slide beneath the other and the two are essentially colliding are the places where virtually all the big earthquakes on the planet happen. In fact, as far as we know all the earthquakes bigger than about magnitude 8.5, somewhere between 8.5 and 9, only occur on these subduction zone plate boundaries and nowhere else. Not the San Andreas fault, not other continental faults. Possibly the Himalayan front may be able to have a magnitude 9 scale earthquake, but largely they're just at these subduction zones. That means the earthquakes happen underneath the ocean. And so in order to study them, that's the reason why I'm a marine geophysicist or marine seismologist is I'm interesting in going to where the earthquakes are, and that requires going to sea. And it gives us special challenges because it's hard to put instruments off-shore compared to on land, for example. The ocean is deep. It's challenging to get equipment to the high-pressure environment, and to get data back and forth you can't send electrical signals through the water, for example. So on land we can do all kinds of things with satellites, in the ocean not so much. Well, the earthquake itself then, if we zoom in on this boundary between the Pacific Ocean plate and the Japanese Islands as part of the North American plate, we can kind of see what's going on in a subduction zone. The down-going plate, we call it, is sliding beneath this sort of wedge-like area that's just kind of the leading edge, you might say, of Japan such that Japan is moving this way relative to the Pacific plate, the Pacific plate is moving that way, and the Pacific plate is actually denser, and so, therefore, it tends to sink in the Earth's relatively kind of fluid, viscous mantle. So it will sink back down underneath the Japanese islands. You may have seen stuff about subduction zones where you see that as it gets deep enough down to a hundred kilometers or so, it actually fuels the melting that fuels the volcanoes of Japan or of Cascadia and the Pacific northwest in the US and Canada. But shallower than that in the portion where the crust, which you can see here is 22 miles thick, of the Japanese islands and the crust of the Pacific plate meet each other then it's relatively cold, and instead of sort of flowing viscously like the deep mantle does, the rocks are what we call brittle. They are in frictional contact with each other, meaning that the rocks sit on each other and there's a lot of force, of course, with all the weight of this overlying slab or wedge of material sitting on top of that boundary between the two, so they tend to stick together just like this podium is sort of resting on this carpet and the weight of it makes it hard to push. It might be bolted down, but maybe if I pushed hard enough I could slide it across the carpet, but that would require friction and a lot of stress to get it moving. So the boundary between the two plates then is the fault that generates the earthquake. And, in fact, that generated this earthquake. The Pacific plate is moving only about nine centimeters per year in the direction of the Japanese islands. So for decades, maybe hundreds of years, maybe even thousands of years, the two plates can just sort of converge on each other and the part that's between them is stuck, and instead of sliding freely, the plates just deform. They bend and they kind of, just the way an archer pulls a bow back and bends the bow, they elastically deform more and more and more until eventually that deformation has stored up so much strain energy that it overcomes the frictional resistance on that fault and, boom, the fault slips. And so that nine centimeters per year accumulating over long periods of time suddenly turns into 90 meters or 9,000 centimeters of slip. That's a thousand years of accumulated plate tectonic stress that slides in one big fault slip of that. 90 meters is the outer limit of what's estimated right now, but at least 60 meters. So hundreds of hundreds of years of accumulated, maybe a millennium of accumulated fault slip popped in one go when that slip happened in the earthquake. Well, so it starts in a particular location where the fault begins to break, and we call that the nucleation point for the earthquake, and that also becomes the point that defines where the seismic wave started to radiate away from. So that's how we define the epicenter. Both the epicenter and what we call the hypocenter, or is marked here as the locus, which is really wrong, it should say focus, down in the subsurface, but that's not really the whole story of the fault slip because the slip starts in a patch and then it radiates away from that patch on the fault. And it can move down the plate boundary, it could also move up towards the surface, and it can move laterally along the plate boundary. And so the fault break grows from a small break to a bigger and bigger and bigger break. That's actually fundamentally the difference between earthquakes of a different magnitude. If that fault had started breaking, maybe expanded its slip to an area the size of this room and then stopped, we would have had a magnitude 1 earthquake. The earthquake, the moment scale, the energy scale, is logarithmic. If it expanded to the size of Madison, we might have had about a magnitude 3 or 4 earthquake or so. And only by expanding to a scale hundreds of kilometers long by about 200 kilometers up and down the plate boundary here did it achieve that magnitude 9 amount of slip. It turns out the way we measure magnitude is how much the fault slipped times the area over which it slipped times how much friction there is, how much the strength you need to overcome to push the fault goes. The other thing that we think we've known for a long time about how these earthquakes happen is that they start somewhere down here in this part of the fault that's locked up, but they don't just extend infinitely down and up along the fault zone. As the slip propagates down the fault, it reaches this hotter place deeper in the Earth where you're into the mantle of the overlying plate, and the fault isn't fictional anymore, meaning the slip kind of goes into a kind of material that's hotter. It's not really molten but it's sort of punky enough that it slows that slip down and it stops. It tends to stop by the time you're at that depth which is why the slip doesn't just keep extending all the way to the center of the Earth or, in fact, doesn't even extend really underneath the land in this case. But also, as we go up towards the deep sea trench up here, so up towards the surface of the Earth beneath the ocean, it's been very canonical to say that that fault slip, as it comes near the surface, starts to die out and the slip energy is absorbed by the softer, weaker rock, colder rock materials as you get near the surface. An analogy I could draw is imagine some bedrock, like granitic bedrock, underneath a layer of soil. Well, if you made the granite, somehow you pushed it from below, and you made it slip, you could imagine the rock would just sort of grind on itself and move along the fault, but the soil layer is soft and it would just sort of deform without actually causing a sharp break potentially. It might absorb that slip energy. I don't know if that analogy helps or not, but that's essentially the way we've thought of the uppermost about, actually about 10 kilometers in depth. So the outer 50 kilometers or so of this very low, gently sloping fault plane would be the area that absorbed the energy and that the break of the fault would never make it all the way to the surface of the Earth out here, that the properties of the rock were wrong for having slip of that fault go all the way up to that trench. Even without slip to the trench the slip of this earthquake fault can cause a tsunami to take place, I'll show you that model in a minute, but with the slip extending all the way to the trench, as I'm going to show you it did, clearly, the tsunami has this potential to be much, much larger. And that's part of what we've learned about this thing over time since then. Well, I said that this earthquake was unique, and it really was. This is a map of all the earthquakes of magnitude 9 or above that we know about that have ever occurred. So we've had seismometers on the planet making measurements for a little over a hundred years now, and these are all the 9s in bold up here. The biggest one ever was this one in Chile in 1960, Alaska in 1964, 9.2, one you probably never heard of off the coast of Kamchatka here in '52, and then the one in Sumatra in 2004, and then the Tohoku earthquake, the one that we just had a year ago. By the way, it says the 12th of March because that's in Greenwich Mean Time; whereas, it was March 11 to the Japanese. The other one that's showed great out here is the earthquake that we infer with a high degree of certainty, well, with a high degree of confidence anyway, happened off the coast of Washington, Oregon, California, and British Columbia a little over 300 years ago. And for a long time it was thought that this particular subduction zone, we knew it was a subduction zone, didn't have big earthquakes because no small earthquakes were being recorded, but we now know that it seems to have caused a tsunami in a time before there were any Europeans around to write anything down, the Native Americans didn't keep written records, because a tsunami arrived in Japan with no earthquake associated with it, and the characteristics of that tsunami and computer model simulations show that it could have only come from the Pacific northwest. There's corroborating evidence in the northwest that this coastline was suddenly drowned and inundated, and this is part of this paleotsunami studies. So it had to extend along this whole length. We're pretty sure to an approximation that was a magnitude 9 earthquake that took place there. Do you have a question? >> How can we have subduction zones in every direction? >> Well, in this case, you have the Pacific plate that's all moving in this sort of northwestward direction, but there is a small, it doesn't look like much, mid-ocean ridge and this small plate here. So there's spreading between that plate in that direction. Whoops, I clicked it again. And a plate called the Juan de Fuca plate moving underneath the Pacific northwest. A larger version of that would be down here, this is the east Pacific rise. The Pacific plate is on this side but this Nazca plate on the other side is subducting underneath South America. So you have spreading of sea floor and then subduction. And you can take Geology 100 and learn lots more plate tectonics if you want to. I'll put it in a little bit more perspective. This is a pie chart, and what this shows is the amount of energy released by all earthquakes on the planet from 1906 through the end of 2011. And so it shows you distributing just the total amount of energy. So all the earthquakes smaller than magnitude 6 that ever happened, there's millions of those in that span of time, or hundreds of thousands of those, take up this little wedge here. Magnitude 7 to 6 is this wedge. 8 is there. Come around to here. Chile in 1960, Alaska, Sumatra, Japan, and Kamchatka, just those earthquakes by themselves are half the seismic energy the Earth's ever released, at least in the last century. So that's the power of the log scale, I guess. Another way to look at it, one more way to look at it is this one, and this one's kind of interesting because now we have a time map from 1900 to 2011, and it's just the magnitude scale, 8, 9, and so on, all the magnitude greater than 8.8 earthquakes that have happened since 1900. Well, you notice that they're not very evenly distributed over time or randomly looking distributed. There were four between 1952 and 1964. I was born in 1964 so my entire life and career until after I had tenure there had never been a great earthquake of this scale until 2004 in Sumatra. So you can see why it started to become a kind of exciting time in my field recently. We waited 40 years, in essence, for an earthquake of this scale to happen again. And, of course, the instrumentation was much more poor back in these times than they were in the last decade. So we've learned so much more from these recent events. Does this clustering, apparent clustering, in time mean anything? We don't know. We honestly don't know. The canonical story, again, is that earthquakes happen randomly in time. There's not association with phase of the moon, weather, all sorts of other events. The Air Force doing experiments, no. That's one you find on the Internet a lot. Sorry, it doesn't correlate. But there is this fact that the largest earthquakes happen four within about a decade, and now we've had three within about a decade. So of course that means there's one more coming, right? Well, no. We don't know that either. The problem is this is a very small number. You can't do statistics on seven earthquakes. So we don't know if that's just random clustering or if there's some meaning in terms of a linkage between different events. Go ahead. Question. >> Could you translate the word tsunami into English? >> Sure. I think I'll get there just now. Let's go on to take a look at what tsunamis are. I can, well, let's see. What I'm going to do, I think I've dealt with that already. So the earthquake off the coast of Japan, just hold the thought for a second about tsunami and I'll get to it, was recorded by many, many seismometers in Japan and of course many, many seismometers around the world. Seismologists are able to use the way the waves look and their amplitude and the signature at different locations around the planet to do a simulation, a numerical inversion for how much slip there was on the fault, how large the area was, and how much it slipped not just on average but on different patches on the fault. And this picture I've put up on the screen, it's pretty technical but it's three different versions by three different research teams of the slip on the fault off-shore in Japan during that event, and they're based on three different data sets. This one in the middle says teleseismic. That means they took the instrumental records from around the world and inverted the fault slip, and so they get these small vectors. You can see the epicentral location is the star there, and then these colors represent amount of slip going up to more than 40, or that says more than 60 meters in the off-shore region. And notice how much of the slip is concentrated right where I said the slip shouldn't be, up at the shallowest part near the deep sea trench. People who use what's called strong ground motion which means the seismometer is very close by in Japan got a sort of similar, not quite identical but similar model of the fault slip. A little bit smaller amount but still 40 meters, still record-breaking amount of slip. And then people using the arrival of the tsunami at the coast and around the world also could invert for the fault slip and got this coarser model, the one that also is consistent with very large amounts of fault slip in the shallow part. This is just all wrong to our sort of canonical thinking again about seismology. We expected the largest slips to be deep and maybe where the epicenter was and then sort of dying away as you went up shallow. So something is funny about that. Alright, well in Japan, the seismometers, as I said, they're instrumented like crazy with GPS sensors. Now, GPS you carry around or you drive around in your car, it tells you where you are. If you fix a GPS sensor to the ground and fix it in a very careful way so it's locked into the bedrock and leave it recording continuously, you can actually get an incredibly precise measurement of the position of that point on Earth in sort of absolute coordinates, latitude and longitude or relative to the GPS stations. So you can actually measure the location of a point on Earth to about a one millimeter plus/minus precision. One millimeter. That much. And so that's actually one of the ways we've verified plate tectonic motions is we see the plates changing their position relative to one another by nine centimeters per year. But close in when a big earthquake happens, we see a big signal. And I know I showed this one last year, but this is too good not to show again. So let's take a look, actually this is a refined version of what I showed last time around. What you're going to see is two different maps of Japan here. This one is showing you horizontal displacements of those GPS stations, and every one of those little dots is a GPS sensor in northern Japan. The one on the right is going to show you the vertical displacement, so how the ground moved up and down as a result of the earthquake, and the horizontal is how the ground moved north, south, east, or west. And what's going to happen is each point is going to show you a little vector, and here's a scale bar for the vectors down at the bottom. One meter is that long, so if a vector is longer, it's longer than a meter. And then a different scale for the vertical, half a meter up and down. So I'll run the little movie, and it will start a little before the earthquake, and then it will just step through time and you'll see after the earthquake takes place. Alright, so now we are before the earthquake. You see a little bit of noise. That's the jitter in the GPS recordings. The star is now the earthquake has happened. 30 seconds since the earthquake and you start to see the GPS sensors move. And you can actually see the surface waves, the seismic waves, moving through the Japanese islands, and then here in a second as it moves by you'll see the first big aftershock will occur. Where is my mouse? There we go. Right in over here. There's a blank out, actually that's just a glitch in the file, and then there's the aftershock. That's actually a magnitude 8 aftershock that happened while the main earthquake was still sort of shaking the ground and going on. Ground shook in Tokyo for between six and eight minutes strongly. In fact, it went on so long people complained about becoming seasick from it, essentially, especially in tall buildings that were swaying back and forth. I'm going to go back in that video for just part of it again because I want you to watch how the earthquake slip seems to sort of take two steps. We're going to see the event, you're going to see the arrows grow and pause and then grow again. Okay, there they go. Pause. And then take off. Okay? That actually is a signal of the process that was going on just as the earthquake got going. Now, these places in Japan moved as much as four meters to the southeast, permanently in the earthquake. Think of what that would do to the survey markers on the boundary of your property. If the property moved four meters underneath the latitude and longitude, I don't actually know how they're dealing with that. Here's just the final sort of results of all of that. And it shows all these arrows with a one-meter scale there showing that all of northeast Japan was almost like it was being sucked out towards a point somewhere off-shore. It turns out that point off-shore is where that very large slip occurred shallow on the fault in just the upper few miles of the fault near the Japan trench. And this is really a remarkable change for us. So that's an amazing result right there, but actually Japan went one step further and did something unique that we've never ever had for one of these subduction earthquakes, and that is a number of groups from universities and government in Japan had actually placed some stations on the sea floor even in thousands of meters of water depth that were with acoustic transponders beaming information about their position to satellites to tie into the GPS network above them. So much coarser, lower resolution, but these stations on the sea floor, they knew where they were. So they could go back in after the earthquake and resurvey the location of a series of markers that had been placed on the sea floor, monuments, essentially, that had been places on the sea floor. Well, at the coast we saw up to, whoops, this is a problem with playing with the mouse. At the coast we saw up to four meters displacement. That's that little arrow right there. Off-shore we saw displacements of 15, 20, or 15 here, 20-some meters, and in fact, this station going out towards the trench between 3,000 and 4,000 meters water depth actually moved horizontally 31 meters. So this is the first time this has ever been documented. We've never had data like this before. And it just shows that the slip was even greater off-shore than near the coastline. This goes into a much better understanding of what the fault did in the earthquake. Actually, even stop better than that, another group had some instruments on the sea floor that are called ocean bottom seismometers, and they had those surveyed in pretty well too, not as well as these GPS stations, but they went back and resurveyed them. They lost some of them after the earthquake. Some just never talked to their ships at all, but they went back to the ships and surveyed the locations. And even with very large error bars, very large uncertainties, like 20-meter uncertainty here or here, if a station moves 74 meters, a 20-meter uncertainty doesn't matter very much. This station moved 58 meters towards the trench, this moved 74, so there's just this enormous motion that is modeled, and this diagram is pretty ugly but it was published, as much as 80 or more meters along the plate boundary fault, and this is a cross-section through that outer portion near the Japan trench. So the whole wedge of that part of the Japan plate just slid outward, and that slip actually instead of dying down as it got near the surface, got bigger and bigger and bigger as it got near the surface. It took off kind of just like some kind of runaway train. Or actually the analogy I like to use is that it's something like as the fault slip got going, this whole part of the Earth's crust, this outermost part of the Earth's crust essentially was moving fast enough that it starts to ride on a pillow of high fluid pressure that's in the sediments or rocks in the fault. It's like it was your car tires hydroplaning on a wet road when you're going too fast because the water can't get out from underneath the tires, and it looks like something similar to that took place on the fault off-shore in that region. There's corroborating evidence for this amazing amount of fault slip in another just really cool study. I guess I won't give you too many of the details except to say that basically ships can do bathymetric surveys. So what they did was they had a place where there had been a bathymetric survey before the earthquake, actually several, and they went down the precisely same line with the ship three days after the earthquake and resurveyed the sea floor topography, that's what bathymetry is, and found that it had changed. It had moved. And so this is a map that shows the difference in that surface elevation along this cross-section that goes from the diagram down here. This is just the very point where the two plates meet at the Japan trench, and this whole region had slid outward. So it's consistent with that other evidence. So it kind of leads to something that I just recently kind of was provocative enough to propose at this meeting I just came back from in Japan, this hydroplaning idea. That, essentially, as it moved 60 to 80 meters, it slid on not exactly a layer of just water but of pressurized pore water within the fault zone. So there's water inside the cracks and pores in the rock, and that if the sliding was happening that fast, the water can't escape and so it would literally have sort of hydroplaned. Well, all of that displacement of 60 to 80 meters is like if you're in the bathtub and you put your hand underwater and suddenly go like that. What's going to happen? The water is going to create a wave forward. And of course that's what the tsunami is. So, a tsunami, the word tsunami, you asked me about the word tsunami, it's actually a Japanese word, and it's two words. "Tsu" means harbor and "nami" means wave. It's called a harbor wave. Well, why a harbor wave? Because normal storm waves are short wavelength, and so protected bays and harbors, the waves wouldn't come in to. They break outside the breakwater or natural defenses of a harbor. But occasionally these water surges would be so overwhelming and so long in their wavelength that they would come right into the harbors and cause devastation of the sort that we saw in the event here. And so in Japan, historically, they've been given that name of a harbor wave. Well, how does a tsunami occur? It's really pretty straightforward. If this is, sorry I turned the subduction zone around the other way for you here, but if this is the plate going down, the two plates are stuck along what I called that locked or coupled portion. It bends the plate back. It actually lifts the plate up on the on-shore or back region while pulling back on the outer part, and then eventually that breaks. When that ruptures, the land actually goes down, and that's exactly what we saw in the GPS data. The land at the coast of Japan subsided as much as a meter, but the off-shore portion was launched outwards and it started the water moving because it, essentially, just uplifts, of course this diagram is very much of a cartoon, it's exaggerated, but it uplifts the sea surface by as much as several meters. Not 50 meters or anything, but just several meters and then that starts a wave propagating away from there, both back towards the land and out in the off-shore direction. So in a kind of a little animation version of that, a tsunami is generated by the displacement of the sea floor due to that elastic sticking of the two plates together with each other. And, of course, as that propagates, it sent a tsunami across the whole Pacific Ocean. Last year I talked about that. We saw effects of it even in California, but the devastating tsunami was what reached the shoreline close by. Now, a meter or two doesn't sound like much, and in the open ocean if you were in a ship at sea, you wouldn't even notice it because it has a very long rise time and then decline. It has a long wavelength. But a hundred kilometers wide zone one meter high, when you translate that into shallow water, the wave slows down, the water piles up and piles up, and we saw very large tsunami inundation at the coastline. And that's, of course, what caused the devastation in Japan and the tsunami running in over the coastal plains and into all those small coastal towns and harbors and fishing villages and things along those lines. I want to just run another little video because it kind of illustrates that we've all seen, I think, a lot of footage of the tsunami, and maybe you've even seen this particular footage, it's not so much that a huge breaking wave comes and breaks over your head like the Hokusai woodcut drawing. It's more like a rapidly rising tide. Just over a few minutes, just the water surges up and up and up and up. And you'll see this is the surge in one harbor in Japan. There's a tsunami seawall that was designed to withstand a three meter high tsunami, but it was, of course, overtopped and inundated. In this particular region, the tsunami was as much as 20 meters in height. 20 meters, 70 feet, 75 feet in height or so, ultimately. So this is the first big tsunami wave arriving and topping that seawall in the town of Kamaishi. You can see how muddy the water is because it's picked up a whole lot of just shallow water sea floor sediments out of the harbor. And many people survived by being on high elevations of buildings that had actually been designed to withstand it, but many, many other people of course didn't survive. In Japan, researchers had gone out after the tsunami and have documented the height of the tsunami with thousands and thousands of observations of essentially where the bathtub ring is, where the water line is that shows how far the water came inland, both in terms of its lateral extent and its height, and they've mapped that out. I don't know how well you can see it here, but all of these, this is northern Japan. Now we're looking sort of from the northeast down on Japan. This is that same coastline and these bar graphs, essentially, are tsunami height at various locations. And we can see all the concentration of where it was the largest along the north coast there. It actually arrived even down as far south as Tokyo and Yokohama but not enough height to be really devastating in that region. The nuclear plant location is just about there. If we take that same data and put it on just kind of a plot which maybe makes it a little easier to look at, then here is, again, that same coastline. And here is a plot of data points that show the height up to 40 meters, 150 feet almost, 140 feet, of the tsunami in various locations. All the circles are the 2011 event. And so you can see that it reached that kind of height in this central region in here, and even significant heights of 10 meters of more for hundreds of kilometers along the length of the coastline. But what you can also see from this plot is that in 1933 and in 1886, it should actually say 1896, there were previous earthquakes or previous tsunamis in that same region that in some cases caused just as large a tsunami. At least locally, just as large a run in of the water. However, none of them had as large an aerial extent, and in particular, none of them caused as large run ups down here in sort of the central and southern portion of where the slip happened. By the way, there's kind of a gap in there. That's the region right around the nuclear plant where they didn't go in and collect the data in those early months, and it's still part of the exclusion zone. So that's where the nuclear plant is right there. Tsunamis of this size were not unprecedented in that region of Japan. That's just a true statement, full stop. They had a 39 meters run up in 1896. So people knew about large tsunamis. Many people saved themselves. Hundreds of thousands of people were in the way and they got out of there. They felt the shaking from the earthquake. They didn't even need a government warning to tell them to get to high ground. That's well understood. There was also the warning, though. The warning went out and based on the extremely early, the first few minutes of calculations about the earthquake size, it was estimated that about a 7.9 earthquake had happened and could cause a 3- to 4-meter tsunami height. So some people, probably, that caused a little false complacency. They thought three or four meters, we have seawalls, we have tsunami defenses, it won't overtop them. And, of course, the tsunami was much, much larger than that. And a big part of why it was much larger than that was that very large slip of the fault in the shallow part that really moved the sea floor happened. People didn't understand that at first and didn't recognize it. I think I will skip that one. I went to the Sendai region around the city of Sendai. In fact, if any of you remember the videos of the airport that was flooded by the earthquake, I took off from that same runway in August just a few months later, four or five months later. This is some of that same region. This is several kilometers inland from the coastline, and the top of this gas station had been hit by debris. So that's how deep the water was even several kilometers inland. This is the foundations of a village that had been there and the houses had been completely destroyed. They've cleaned everything up. There are massive piles of debris now that they have to deal with, and the solid waste problem in the tsunami region is enormous, actually, in Japan. But I want to talk is on this same trip to see this stuff, we weren't just sort of geo sightseeing. It's a bit ghoulish probably. What we were looking at is actually the work that's been done over a number of years by some researchers in Japan who have studied the occurrence of tsunamis before the last few hundred years, the sort of record that is really well-known historically speaking. And so the question was, why didn't the nuclear planners basically take into account the possibility that a tsunami of this size could come in and inundate that region, not just the nuclear power plant but also the people of the whole region? And the answer, in part, is that their estimation was that the maximum size tsunami they could get was up to three to five meters, I'm not talking about the estimate from the early part of the earthquake, but that was their estimate for the maximum tsunami hazard at the location of the power plant. And so they had some defenses for an inundation of water that high, but, of course, those defenses were overtopped. The backup emergency generators were flooded with seawater, so when the power plant shut down the emergency generators couldn't keep the cooling going, and you probably know most of the rest of that story. The meltdown occurred in multiple reactors at that plant. Well, it turns out that that was the official risk assessment by TEPCO, the power company, but there were geologist who knew that there was strong evidence that in the past there had been a tsunami of at least this size, if not larger, but more than a thousand years ago and that perhaps they just hadn't looked at long enough time. And so in this region, this is the Sendai plain similar to the landscape a little further to the south where the nuclear plant is, a group of researchers from Japan have been studying the soil deposits in just the upper few meters. And the way they do that is they cut basically slices or dig pits or cut a core sample down through the soil, just maybe two meters long or something like that, and they extract those. And here I'll jump ahead and then go back. They extract those soil samples. And you're actually seeing Daisuke Sugawara, a guy who led a field trip that I was on in August, showing you one of these slices down through the soil. And what they find is that in the soil there are occasionally these buried layers of coarse marine sand with all kinds of shells and marine organisms buried in them, and they're clearly the deposit of a rapid inundation of ocean water over the land. That could happen from big storm but it turns out probably not to this scale and this far inland. That can really only happen from tsunami. And with radiocarbon dating and some other methods, they can date the occurrence of those, and so they actually have tied this back to a known earthquake and tsunami in 869 AD. It's called the Jogan tsunami in Japan. There was historical record but it wasn't all that well-known how extensive it had been. They went in, there's a lot on this diagram you don't have to worry about too much, but they've basically dug these soil slices all up and down the coastline here and come up with a map of how far inland the tsunami deposits go. And, in fact, they go as far inland as something like five or six kilometers in the same region that was inundated by the tsunami now. And they back-calculated that it had to be at least something like I think it's a magnitude 8.6 or so earthquake and maybe larger that happened 1100 years ago during that event. So it's pretty clear that that's happened in the past. And what we saw was an earthquake and tsunami that didn't repeat the most recent events in that area but repeated the much larger one that slipped this shallow portion of the fault further back in time. Actually, that group had advised the power plant people that this had happened and they'd been parts of various risk assessment panels over the years, several years before this tsunami, and here's actually a quote from a paper they published in 2001, published in English, they had also published a lot in Japanese. But basically it says given the recurrence interval, the possibility of a large tsunami striking the Sendai plain is high and our numerical findings indicate this similar tsunami would inundate the coastal plain two to three kilometers inland. That was basically a smoking gun that this has happened in the past and could happen in the future. Now, the question, and this is the question for New Orleans and hurricanes and a whole lot of other natural disasters is, can you plan for the once in a thousand years occurrence? Should you even try? Can you plan for the once in 10,000 years occurrence? I don't have an answer to that. This is not my area. Very high risk but very infrequent events are challenging for people to deal with. Okay, well, I want to talk about what's going on in terms of research into this event and how that fault slip, the distinction of that fault slip is something we're trying to understand better. Why did the shallow part of that fault slip so much in the earthquake when we didn't think it could? And is that theory that it essentially rode out on some kind of low stress pillow of wet sediments viable or not? And it gets to this question, this is a little bit of an aside, but it gets to this question that I always get asked, can big earthquakes like this be predicted? And in a sense it's sort of like talking about the climate versus talking about the weather. I can tell you that in Wisconsin we typically have cold winters with a fair amount of snow, and in early March it's usually still pretty wintery, but back in December if you had said on March 14 will it be 75 degrees out, somebody would have said not very likely but we can't really tell you. That's short-term prediction versus long-term. Well, in earthquakes we think we have some understanding about the long-term forecasting. Which places in the world can have big earthquakes. Well, subduction zone for one. We'd like to know more about which ones might have the big earthquake and which ones might be due versus which ones we probably don't have to worry about for a thousand years. And we thought we understood that a lot better maybe before this earthquake than we do now honestly because this region, as I'll show you in a second, was not considered at a high risk for magnitude 9 earthquake. It didn't fit our sort of conceptual model for which subduction zones make magnitude 9 earthquakes. I'll get to that in a minute. Short-term warnings, though, no. The answer is just no. Whatever you hear about earthquake prediction is we do not have a mechanism that we can understand, we can see data, and make a forecast in advance that says this place is about to have an earthquake and show better than random chance kind of confidence that that's really happening. Some earthquakes have foreshocks but other ones don't. Some earthquakes the animals all leave the forest but other ones they don't, honestly speaking. Maybe you've heard of some of those things. And, really, the question is whether it's even possible or not. Some seismologists think that as an earthquake begins there's characteristics of how the fault zone is stressed and sort of prepared that from the very beginning of the earthquake it's kind of predestined to become a magnitude 9 whereas another one is predestined to only become a magnitude 2 or 3 or 4. But other people don't really think that's true that it's sort of a very subtly balanced, cascading series of events that causes a really big earthquake, and so that as it begins even the Earth doesn't know if it's going to be a big earthquake or not over the next, say, few tens of seconds. And the way that fault slip in this one happened on the deep patch seems to have hesitated and then broke through to the shallow patch and had the much larger slip shallowly that I showed you in the GPS data kind of suggests that maybe that, unfortunately, what's going on is that as it breaks the fault, very subtle differences in the stress and conditions on other parts of the fault mean they could go or not go. We know that this same fault zone that slipped in this earthquake has had magnitude 7.5 to 8 several times over the past hundred years in smaller parts of that fault. And so all the seismologists in Japan thought that that was about as big as the earthquakes could get in that region. This is actually Japan's earthquake hazard forecasting map, published in 2005. It was the map that their government planners used for risk of large events over time. And so the details are not that important. It's a 30-year probability of a certain amount of ground shaking from an earthquake. So red is high risk over the next 30 years. This whole region of southern Japan called the Nankai Trough was high risk. Northern Japan was among the lowest risk places in Japan on that map. Well, of course, that's where the magnitude 9 earthquake that's shown here occurred. So clearly the map is humbling in the sense that basically most of the big events that have happened in Japan since that map was published have not happened in the highest risk locations. Does that mean the risk is low here? We don't think so. We know there's a long history of earthquakes about every hundred years in this region. I'll tell you more about it in just a couple minutes. I don't want to overstay my welcome, but I do want to tell you a little bit about the research that's now getting going to go right out to the fault zone off-shore and try to sample that fault itself. And so in order to do that, it's something that's called the Integrated Ocean Drilling Program, an international project that uses, essentially, oil drilling style technology but to do scientific research underneath the ocean floor. In fact, deep sea drilling was invented by scientists and the oil companies picked it up later, but they've certainly taken it a lot further than science has. But this vessel, which is called the Chikyuu, or Earth in Japanese, is part of an international program that I'm very much a part of to drill into various places off-shore. Chikyuu's biggest project so far in the last few years has been my project at the Nankai Trough, that region with the high probability of risk, but now Chikyuu is about to set out on April 1st, just a few weeks from now, to try to drill into the earthquake fault that slipped in that shallow portion off the north coast of Japan. That's Mount Fuji in the background, and that has nothing to do with my talk. I just like the picture of the ship there. So what we do in scientific ocean drill then is really two things. One is we would like to be able to drill down into some place of geologic interest and collect core samples because the geologists would like to have the material that actually was within that fault zone, let's say, or within some other zone of interest, to study and to do all kinds of laboratory work on, measure the strength, measure the kinds of minerals that are in there, how much pore water is actually locked up in those rocks, how many fractures are in it, what type, all those sorts of things. And we also have ways of running instruments down into the bore hole and making measurements sort of under the actual ambient stress conditions deep beneath the surface of the Earth. So we can drill as much as several kilometers into the subsurface even in deep water, and then bring those cores up, do work on them on the labs on the ship, and then even more work in or labs back home. So this project that has now been developed is called JFAST. I can't even remember what that actually stands for. I'm sure it's a tortured acronym, but the idea is that basically it was put together very quickly, I was part of the proposal team, to do rapid response drilling. And the idea is to reach at a place off-shore, that's the epicenter of the earthquake, this is the region the drilling is going out to do where the fault slipped while not much has happened, and in particular while the temperature of the fault is still similar to what it was during the fault slip. And the reason for this is to test that kind of idea that I talked about before. If we look way off shore and the drill project, I showed you this before, the drilling project is way down here, it looks like almost nothing but it turns out to get to the fault in only a thousand meters of drilling, only one kilometer deep hole, you have to go to where the water is seven kilometers deep, way out there, it will be the deepest hole every drilled to any depth in the ocean, deepest water hole that is, and to drill into the fault zone. And what they want to do is actually place temperature sensors across that zone and collect core samples of the fault, and the goal of the temperature sensors is to measure any residual frictional heating from the fault. And the reason for that I think I can kind of demonstrate. If you take your hands and you place them together gently but don't put any pressure on them and you rub a little bit, what happens? They warm up a little, right? That's friction, frictional heating. Now push them together as hard as you can and rub them back and forth. They heat up a lot more. There's a lot more frictional heating because what we call the normal force, the stress on the fault is much higher when you push harder. If the fault slipped under what we consider ordinary or typical or well-understood earthquake stresses, there should have been actually in a flash during that slip hundreds of degrees of heating of the fault. It should have heated way up as it slipped 80 or 90 meters. Imagine grinding two rocks against each other. You really heat them up. If it slipped under the temporary, very low stress conditions of the kind of hydroplaning idea I was talking about, then we expect a much smaller heating signal. So that heat will dissipate over time so that's the reason to do it rapidly. You want to get out there, drill into the fault zone. If we zoom way down at just this region right in here where it says trench access there, that just means the deep sea trench, then this is an image using our seismic reflection, using our sound waves to image below the sea floor. That's the seabed. This is the top of a block of the oceanic crust going down. And so we think the fault zone lies right there where the blue and the green is and drill a hole down into that fault zone and get those temperature sensors in place and also get some fault rocks back. So that's the goal. Alright, well, I will finish up by talking just a little bit about this NanTroSEIZE project that Tom mentioned at the beginning which is also ongoing. Now, this region was deemed the highest risk for future big earthquakes and tsunamis in Japan, and that risk has not declined at all post-Tohoku earthquake. So about 10 years ago, actually, we started planning for a deep sea drilling project with some fairly similar approaches but trying to get instruments into the fault zone before the next earthquake and sample the rocks from the previous one in this region to the south. So if we look at it in some perspective, here's Japanese islands again. We're kind of looking, north is up that way. Here's the trench that had the 2011 earthquake. This region is called the Nankai Trough. That's just the name of that deep sea feature. This is the Philippine sea plate, a different tectonic plate, sliding beneath that part of Japan. Japan is so lucky it has two subduction zones, not just one. And better yet, Tokyo, the world's largest city by population, sits on top of the point where all three plates meet, and there are literally two subducted slabs beneath the city of Tokyo not just one. It is probably about the most earthquake prone place in the whole world, and it's the world's largest city. That's the way it goes, I guess. At any rate, this region in the southern part of Japan had a magnitude 8-plus earthquake in 1944, another one in 1946. Both of these produced 10-meter tsunamis along the coastline, and we know that the off-shore is structured very similar to the Japan trench. So if we look off-shore in kind of a cross-sectional mode, there's the seabed and here's the subsurface. Again, we have a fault zone that comes up. We expect that earthquake slip came up from deep and maybe made it all the way to the trench like in Tohoku or maybe didn't. Until Tohoku we thought it probably stopped here. So we've drilled a series of shallower holes into the shallow part of these faults, and I'll show you an example in just a second. And our goal is to drill, actually, seven kilometers below the sea floor down into the main plate boundary where it's really at a depth where we know the earthquake slip occurred in the past and will again and place long-term monitoring systems right inside the fault zone, temperature, pressure, seismometers, what are tilt and strain meters, to get a whole picture of the whole thing and basically something like place a stethoscope on that fault zone and then watch it as it builds up strain in the future. Well, in that drilling, actually, this is a cross-section of the faults. It's too complicated probably to explain, but we've drilled into the very tip, and just a few hundred meters below the sea floor we've already sampled that fault zone. And it turns out we got core samples back that look horrible. They're all broken up rock. This is a core sample. It's about this wide, six centimeters wide. You're just looking at about 20 centimeters or so of core right there, but it turns out this little black band right there, only two millimeters wide, shows evidence for having been the location in plate boundary fault slip in earthquakes in the past and even for having been heated up by significant amount by that frictional heating that I'm talking about. In fact, we can take x-ray CT scans of those fault cores, and this is a scan of one of those cores using an x-ray CT just like a medical x-ray CT scanner. And in the interior of the core, well, I can't do this with the mouse easily. In the interior of the core, this white patch here is actually the fault zone. We have the real live material that participated in the earthquake fault. It's so thin, just a few millimeters, and yet it had probably tens of meters of slip on multiple earthquakes in the past. So the properties of that, and now I've made it transparent like the Andromeda galaxy there, but it's still just the core sample from that fault zone is telling us a lot about the faults. We take all of that and then we don't just get samples out of the faults but the bore holes we leave behind long-term monitoring instruments. So if this is an off-shore bore hole with steel pipe left permanently in place to keep the hole open, then, as I said before, we can lower those instruments down, lock them in place, and watch the development of stress, strain, and small events over time to see how that fault is actually behaving, looking at a live fault as opposed to just what geologists get to do on land which is usually look at dead faults, essentially. Okay, well, I think what I'll do is I'll stop there. I'm happy to answer any questions, and I really appreciate your attention today.

APPLAUSE