printable banner

U.S. Department of State - Great Seal

U.S. Department of State

Diplomacy in Action

Haiti Earthquake: What We Knew, What We Know Now, and What We Can Do To Improve

Dr. Wayne Pennington
Washington, DC
February 23, 2010



Wayne Pennington:

Let me introduce myself. I’m Wayne Pennington. I’m a Jefferson Science Fellow here at USAID. Jefferson Science Fellows are tenured faculty members at universities. They come in, usually they end up working at the State Department for one year. This year, there are three of us at AID. One of them, my cubicle partner, is back there, Cynthia Baldwin. And we go back to our universities after that year here as full time -- but we go back and are available part time, for at least the following five years, after we go back. With any luck, we learn the tricks and we know how to work things here and contribute even on a part time basis. So while I am here though, I am in EGAT and the infrastructure and engineering group. And I am by training a seismologist. I spent 20-25 years of my career in earthquake seismology. I have experienced large earthquakes, devastating ones. I have usually been in the wrong place at the wrong time. So if you ever wonder, can predict earthquakes for timing, no we cannot. I am living proof of that, because I would not have been there for a magnitude 7.7 in southern Mexico or another large one in northern Pakistan, not the famous one, but a previous one that killed a few thousand people there, too. But what I am going to talk about now is the science, the seismological aspects of the Haiti earthquake. We’ll ignore the fact that people died and there was a lot of crisis and things. We’ll just talk about what the earth is doing; what it has been doing, and what it’s still doing in the Haiti area, and how it is that scientists know about that. What we know now since the earthquake and what we are trying to do to improve communications and understanding of earthquake risk in the area, in general.

So, start out with what we knew. Well, in 2008, multiple authors had demonstrated the possibility of a magnitude 7.2, in the area where the earthquake occurred. The forecast was based on geologic information and GPS measurements made over time and knowing what strain was building up. So, seismologists were not at all surprised at this earthquake, at least not much. Some of them were surprised that it occurred here, rather than on a fault on the northern side of the island, but nobody was really surprised that this occurred here, because earthquakes are known to occur here. But, in general, the earthquake was a big surprise to most everybody who wasn’t a seismologist. And I found that to be the biggest surprise I had -- was that, even though the scientists involved in it knew full well that Haiti experiences a lot of earthquakes over a period of time, it seemed that a lot of other people did not. So we’re going to take a look at that.

We can look at the seismicity of the Caribbean; the different colors represent different depths. These blue ones are very deep, and this color is the shallowest -- and this time period includes after the earthquake near Port Au Prince, there. And you can see right away, of course, that most of the Caribbean does not have earthquake activity. But all around its edges, it does. So, even if you didn’t know about plate tectonics already, you could start inventing it or discovering it, just from a map like this. So, we will look at this in some more detail – that’s the area affected by the earthquake.

A little dark here. I might turn off the light, just for a moment, so that you can see this better, a Google Earth image of the Caribbean area. And even just from the morphology of the sea floor, you can see that there must be something going on around the edges, and in fact there is. And what we do, we define this place as the Caribbean plate. So if we were to define one single Caribbean plate, the red lines there are doing a good job of describing that.

Pretty quickly, we realize that there are some complexities and we have to add in some smaller plates, here, and give them different names. And in fact, my PhD dissertation was done in this area -- 100 years ago, the geology hasn’t changed much, but the science has improved since then. And if we get a little bit more complicated, we realize that the relative motion between the North American plate and the Caribbean plate, is funny. It’s at some crazy, oblique angle relative to the boundary, the plate boundary up here. Well, that turns out to cause some complexities. If it were completely parallel to that boundary, there would be a simple fault there. If it were completely perpendicular to that boundary, there would be a simple fault there. But since it’s oblique, we end up with a series of other faults there, sort of slicing the plate up into some slivers. Well, let’s take a look at that in a little bit more detail, here.

We call that oblique convergence, of the two plates. This would be the North American plate, subducting beneath the Caribbean plate, at the Puerto Rico trench here. This is my feeble attempt at making a three dimensional diagram, but I think you can see what is happening. And we’ve got convergence that’s at a funny angle. This plate is trying to slide that direction, underneath. Well, it ends up resolving itself into two senses of motion. First is convergence that’s perpendicular at the plate boundary, at the trench itself. But then, to accommodate the rest of that motion, there has to be some slip back here. And I realize the arrows on this are pointing the wrong way. The next slide will have them the right way. But this part, then, is sliding relative to that side there. That’s how this convergent motion, in an oblique sense, gets resolved. So, this would be the Puerto Rico trench and this would be say, the Enriquillo Fault in Haiti, or another fault, the Septentrional Fault in northern Haiti. And Haiti, because it is complicated enough, we do get these two faults there. Dim the lights again just for a moment. We do get these two faults, the Enriquillo Fault system down here merging into the Muertos trough and some other things there. And this is called the Septentrional Fault, up here. We’re concerned with these zones, here. We see that we get perpendicular convergence here. We can tell that from earthquake motion and other things at the Puerto Rico trench. And we get strike slip motion here and the arrows are pointing in the right direction in that figure, on the Enriquillo Fault there. Well, this is the part that ruptured; of course the entire fault is active there. We’re very concerned about Jamaica, parts of the Dominican Republic and about this fault up here, but this is the one that we are concerned with right now, part of the earthquake that took place recently. We’re going to take a closer look in the direction of that arrow there and see the fault, in a way that you, most of you are not geologists -- I know at least one person here is -- but if you take a look at this three dimensional figure, based on a digital earth model from the shuttle topography mission, Port Au Prince is laying here. This is the southern peninsula of Haiti, coming out here; we’re looking towards the east. Can you find the fault there?

Male speaker:

I know where it is, but on the two dimensional map it was easier to picture, than here.

Wayne Pennington:

Well, let me turn off the light again, just for an instant, and I’m going to go through a little animation on the slide here, and we’ll see if you can spot it.


Male speaker:


Wayne Pennington:

Okay, some of you can.

Male speaker:

Right above the A in first-aid radar, is that where it starts now. Yeah, right there.

Wayne Pennington:

There it is, there.

Now I’m going to take that red strip off and it will become glaringly obvious to you now. There. I haven’t doctored the figure in the meantime. That’s where it was all the time.

So you don’t really have to be an expert geologist to see that fault there. It’s obvious, it’s there. The big question, of course, then is, “Is it active? Does it slip smoothly with time? Or does it wait, store up the energy, and suddenly release it in a big earthquake?” Well, it turns out this one does that. So, the best cartographers in the world for making things easy to see come from things like The New York Times and The Washington Post. This is a New York Times figure and we’re looking at a cross section there, that goes essentially this way, through the North American plate, which we see here, and the Caribbean plate, is there. And we’ve got the Septentrional Fault up here. In fact, many seismologists had expected this to be the next fault to go in a big earthquake. So, we’re all a little bit worried about that, yet. And if there was a surprise to seismologists at all, it was that this one went first, before that one. But these things have a mind of their own to some extent, so there is the Enriquillo Fault there and a little micro plate in-between, that’s given different names by different authors. So, that’s the general scenario we’ve got there. By the way, while I’ve got this up, and I should have shown you on an earlier figure, the Puerto Rico trench, which runs all along here, and every now and then you’ll see on the news, they’ll call it the Haiti trench or something, whatever they’re talking about. But usually it’s called the Puerto Rico trench, extending from there to here, is a thrust boundary. And that type of boundary is the type that generates or can generate, and typically often does generate, tsunamis, when the earthquakes occur.

And so, if there is a tsunami hazard to the east coast of North America, the east coast of the United States, Florida and so on, it would be from this. That is, of course, a tsunami hazard to the Caribbean and because of the awareness of tsunami risk from the Indonesia-Sumatra earthquake in 2004, the awareness was increased in this area and a tsunami warning system is now in place. It’s run out of the university there in Puerto Rico and coordinated throughout the area.

Now, take a quick look at the earthquake history. A lot of the history of the area is known not from seismic recordings, for the larger earthquakes, they have been too infrequent, but from historic records. Now, you can see that we have decent historic records for most of the places where there are islands or where there are ports, because people kept records there. We’ll just zoom in a bit here, on Hispaniola, and if we look at the historic record we see 1751, 1770, 1842, 1887, and so on. And so I have a question for you; If you were a Haitian living the start of this year, would you remember any of these earthquakes? Would earthquakes have been on your radar screen, for the construction of your home, for example? And the answer, in general, is no, because the earthquakes occurred generations ago. And so, no living person has a memory of those earthquakes and that is a big problem. Sometimes we’ll make a comparison with, say, the recent earthquake in Chile. Chile had, since 1973, 17 earthquakes the size of the one in Haiti, or bigger, since 1973. So, the Chileans are very aware of this. When somebody says that you have to build it to earthquake code, they are likely to say; “Oh yes, I understand that.” And here, because the earthquakes occurred more infrequently, it’s less obvious that earthquake code needs to be followed. So, this little animation is going to go through a sequence of earthquakes from 1751, 1761, 1770. I have changed the color here, because these two may not be on the Enriquillo Fault system. This one actually may be over here, in the Muertos Trough, this one may be on a different fault there. But an interesting thing, the Enriquillo Fault system when it ruptured, it ruptured in 1751 once, 1751 twice, and 1770. And those were all big earthquakes. Most people figure that the earthquake that we just had was most similar to the 1770 earthquake. But, of course, these are historic records from captains who boats were at harbor in the port there and things like that. So it’s a little difficult to really make comparison closely.

Male speaker:

Just one question, the size of it, was that indicative of the magnitude?

Wayne Pennington:

It’s indicative of the rupture area, as far as we can tell.

Male speaker:


Wayne Pennington:

And that corresponds, correlates to magnitude but those are all magnitude seven or greater events, probably.

Now a different viewpoint, and this is from Roger Dillon, who is -- those of you who saw the Discovery Channel special on the Haiti earthquake, he was the rock star involved with that. He was the one they followed around through the – I’ll come back to him again later. Exactly where these earthquakes were is obviously a matter of some scientific debate. Not a big debate, but a little debate. And here we see the earthquakes that have in the past generated tsunamis with the red stripes, and those that did not generate tsunamis with the green stripes. And we see that there is a real and realistic tsunami hazard in the Caribbean. The Haiti earthquake itself killed a few people from a minor tsunami, I believe it was in the Dominican Republic, where they were killed. But otherwise, there was negligible tsunami from that earthquake. In part, because it was a strike slip mechanism --

Male speaker:

The 1770 one, you’re talking about?

Wayne Pennington:

Sorry. No, the one that just happened.

Male speaker:

This one

Wayne Pennington:


In part, because it was a strike slip mechanism, and that does not displace water very much. There was a bit of a normal -- of a vertical component to it and we’ll see that later. And that displaced enough water, to cause enough wave, to cause a little bit of a problem, but not anything we would really call a tsunami. The earlier ones, many of them did cause tsunamis that were recorded in the area. Now, given all this history, and given the fact that we are crawling with seismologist, in this country, at least; how many seismographs do you think were in Haiti at the start of the year 2010? You’re probably guessing the right answer –

Male Speaker:


Wayne Pennington:

Zero. Isn’t that something? Zero seismographs in the country.

So, that causes a problem in trying to figure out what to do. We try to locate the earthquakes very closely, very precisely, but with all the seismograph stations being quite a bit farther away, it’s difficult to locate them precisely.

Now, at some point in time after the earthquake, within a couple of weeks, this map was generated. And this is the aftershock distribution. The main shock occurred here and most of the aftershocks are out there. Now, this scatter; some of that is due to location error, because we don’t have nearby seismographs. So the error then, in the location, is somewhat big. They could all be along the Enriquillo Fault. Some of them, there’s some suspicion, may be along a series of nearby faults, that are a little bit different behavior, and that would be at this end here. But there’s still some debate about that. The largest aftershock occurred at this end here. Now, the Enriquillo Fault comes along and it does do a little bit of a job there, where these aftershocks seem to be clustered and the largest aftershock was. So there is some suspicion that the earthquake would have, in a sense, liked to have gone further, but it got hung up on this little jog in the system there. That might be the case, it might not be. The epicenter was very close to Port Au Prince, but almost all the rupture was to the west of the epicenter. So, the closest to Port Au Princ,e was the epicenter. Often it’s not that case, often the epicenter is near the center of distribution., in Chile that was the case. But in this case, the rupture moved to the west from that. So that means this part of the fault is currently unruptured, has not had displacement yet.

Male speaker:

[Unintelligible] I wonder if I could ask a basic question, which is, the epicenter -- to me the concept of having the epicenter one place but the rest of the earthquake occurring --

Wayne Pennington:

I will actually go through that in a second here, it turns out.

Male speaker:


Wayne Pennington:

But that is a good question. The epicenter, technically, is the name of the map location, it also has some depth, then we would call it a hypocenter or focus. The epicenter is the map location of the start of the seismic radiation and that’s what we use to locate the earthquake. But the earthquake -- an earthquake this size, has some physical dimension to it and so this one turned out to have the physical dimension, that was that region. The aftershock distribution usually maps that out. There are other ways that seismologists can go about determining the rupture size, and we’ll see that later.

Male speaker:

Right.. Does this, then it does not, by definition, mean the point of greatest magnitude?

Wayne Pennington:

No, it doesn’t necessarily. In fact, in this case, we’ll see the greatest magnitude was over here.

Male speaker:


Wayne Pennington:

So, the epicenter was there, the rupture started there and propagated in that direction. Now this is the map that shows the, as of a few weeks ago, it has probably been refined now, where the rupture on the fault plain was the greatest. Now, what we’re doing here, we ‘re looking at the fault plane head on, from the, let’s see here, get it right, we’ re doing it from the north. So the east is there and the west is there. And Port Au Prince then, is off on this side and the surface of the earth is up here. Now, and in fact it’s up higher than this is showing. And we see that the epicenter was at some depth and the greatest rupture was a bit up from the epicenter, and toward the west. As we go farther away from that, more and more of the motion was upward and less of it was sideways. Now, that is kind of interesting. This is a little bit complicated then, this earthquake. But as we get closer and closer to the surface, which is not shown at the top of this figure, the displacement got less and less. So there is not, known yet anyway, a surface rupture of the fault. Now, how did it spread from there? Here is a little animation, speeded up five times. The ruptured front spread like that. And like I said, this has probably been refined; this was a geophysicists model a few weeks, a couple of weeks, after the earthquake occurred.

Now, aftershocks. Aftershocks always occur after large earthquakes like this. Bruce and I were just at a meeting in Memphis where that seemed to be a surprise to a lot of people. So, seismologists can never repeat it enough, a large earthquake will have aftershocks, guaranteed. At the meeting we were at they were saying, “Well, if aftershocks occur,” seismologists” had to stand up and say, Not if, when.” So from a magnitude seven earthquake like this, just as a rule of thumb, we could expect at least a magnitude six aftershock, quite a few magnitude five aftershocks [inaudible] magnitude four aftershocks. Magnitude four’s are big enough to scare the daylights out of you, if you have recently experienced a large earthquake. Did you experience a magnitude four while you were down here?

Female speaker:


Wayne Pennington:

Yeah, okay. If you haven’t just experienced a big earthquake and are tend -- panic as a result, magnitude fours are actually a lot of fun. They’re just, they go, “What’s that? Is that an earthquake? Oh my god check – Oh, it’s getting over, it’s done, whew, what a relief.” It’s little bit like a roller coaster ride, just enough to scare you, but not enough to hurt you. But if your building is damaged, if you were run out of the house, if members of your family have been killed, and things like that; magnitude fours still scare the living daylights out of you.

So, we do see aftershocks after every large earthquake of this nature and the specific timing of the aftershocks, as far as we can tell, random. It follows a statistical pattern though, that is quite standard. And they decrease in number, with time, following a simple mathematical equation. And then the size distribution of earthquakes always follows a sort of pattern. Where if you have one big earthquake, and it turns out that the logarithmic scale works this way. That if you have one magnitude seven, you’re likely to have, say, about ten magnitude sixes, and likely have about 100 magnitude fives, if you average it over a long enough time. Now, it doesn’t work quite exactly that way for the aftershock pattern, for one magnitude seven expecting ten six aftershocks, no you’d expect one magnitude six aftershock in this. But that relationship is described by this arithmetic. This relationship is described by that arithmetic. You can put these together and solve for some constants, after you’ve had a week or so of aftershocks to see this earthquake, how is its aftershocks sequence behaving. The USGS did that and made some predictions.

Now, they made a prediction that, initially, for the period of the 21st of January to the 22nd of February, they did the calculations on the 20th of January and released it on the 21st of January, and they predicted that there would be a 3% probability of a magnitude seven or greater. Now remember, the earthquake itself was magnitude seven, so this is a little scary, of course, but there is and this happens. A few percent of the time, there is another earthquake as large as the main shock. The figure there is a 25% chance of a magnitude six, within that month period, and a 90% chance of a magnitude five and that would probably be two to three magnitude five or greater events. Well, this time period’s passed. How did they do? Well, the listing of events that we have in that time period is the day before they made this announcement, but after they made the calculations, there was a magnitude 5.9. Does that count or not? Well, technically, it’s not exactly in that time period, but when you realize we’re talking about probabilities here, and these things are fuzzy, that’s actually pretty good. And then, there were two magnitude 4.9’s and two magnitudes 4.8’s. So those aren’t exactly magnitude five, but given that these are probabilities, they actually did pretty well. Their prediction was really pretty much spot on. So, more time has gone by and their most recent assessment is now -- covers a longer time period. For the 30 day time period, that is just about to end, there is a 55% chance of a magnitude five or greater, 7% chance of a magnitude six or greater, 1% chance of magnitude seven or greater. As the time period gets extended, it’s more likely to have any of these events. So, this is the current best guess -- not guess, it’s following the arithmetic and calibrated to the aftershock sequence, for this earthquake and its aftershock sequence.

Now, we do know some things about this earthquake. We know how it slipped and we figured that out from the wave forms of the seismic waves that were generated and recorded in seismographs around the world. So, based on that, people can then -- you see where this is, this is Haiti. These gray dots are in fact population densities; the more dots, the more people there are there. Some of it’s obscured by the other figure here. But, we know where the earthquake ruptured; we know how much of it ruptured in different parts along the fault, from the main shock activity and from some of the aftershocks. And so they know that; okay, we’ve got this earth, we’ve got this fault and we’ve already taken this fault and we’ve moved it some. How does that stress neighboring parts of the fault? Well, you can imagine it must stress neighboring parts of the fault, at least somewhat. So, over here, yeah, it’s stressed, over here it’s stressed. Well, this is a bit of a concern, because, like I said, the population density is covered by this figure; that’s Port Au Prince. Remember, the earthquake that we are studying occurred away from Port Au Prince and ruptured farther away from it. The stress on the fault in the area of Port Au Prince has increased, as a result of that earthquake. And so, should we expect that this means there is a greater likelihood of a large earthquake closer to Port Au Prince in the near future? Probably. This is not an exact science. This is still, let’s put this hypothesis out here, let’s do the arithmetic, and it turns out that’s the result, that it’s under greater stress and more likely to experience now, than in fact it was before.

Our intuition would tell us that too. The historic record, unfortunately, also suggests that, from the 1751 and 1770 sequence of earthquakes. That when one part of the Enriquillo Fault goes, the other parts follow fairly soon, within a period of months to years, or a couple of decades. We have one historic data point, that 1751-1770 sequence. Was that just happened that way, or is it routinely that way. We have no idea. But it does make us more concerned, about large earthquakes on this segment of the Enriquillo Fault, than we might otherwise be. And you basically heard all the science behind it; it’s still a little bit of an art and conjecture.

So, let’s take a look at how scientists actually study these big faults, especially in the absence of local seismographs. Now, I am going to show examples from the studies in Haiti. We, first of all, people would use previous geological studies. Now, this is a figure from Paul Mann who is also often on NOVA and Discovery Channel and things like that, connected with the Haiti earthquake. And he’s a geologist and he and other geologists have mapped a lot of the area. And you can see that rock units come up and they end at the fault, and a different type of rock unit is at the other side. They try to line these things up; get some idea of how much displacement might have taken place over geologic time, millions and millions of years, and they get some ideas of that. Okay, a lot of work goes into that and, by the way, that’s a lot of fun, because you spend a lot of time outdoors, tromping around through the woods and the mountains. It’s just great. And then, if we know where the fault is, and this is an example of the Septentrional Fault, here in Dominican Republic, in the northern fault system, because it hasn’t been done yet in the Enriquillo Fault system. But if you know where the fault is and you got sort of a flat area, where material has been coming from the nearby mountain range, in this case, and depositing there quickly over geologic time, you can trench through that. Dig a trench and look at the layers and see when the layers are disrupted and date those, carbon 14 and things like that work in this time scale, we can figure out how frequently this fault ruptures. And this is Carol Prentiss, who occasionally is also mentioned on some of these TV shows and things. So here you can see, this is actually a Google Earth image of the area nearby, it’s not exactly where the fault was -- where that trench was, but you can see the fault on there and you can see farmlands, and you pay the farmer some money for permission to dig this up a bit. And you dig a trench across the fault and you map it in a profile and you get these different layers. And you can see, that “okay, up here, this is not disrupted, so no big earthquakes have happened in the time period representing this deposition. But what about the time period representing that deposition? Oh, okay, something happened then. Well, you can date these things, like I said with carbon 14, and figure out in a geologic column when the big earthquakes occurred and when other ones might have occurred, because it isn’t always real clear. And you get some idea; how frequently do these things occur over recent historic time. And in this case, the Septentrional Fault, it’s maybe every few hundred years or something like that, maybe a little more often. And the same sort of answer is likely from the Enriquillo Fault. And it can be refined, that’s how we know the recurrence rate of earthquakes, say in the Pacific Northwest of the United States. Did you know that, say, Seattle/Oregon coast is a seismic hazard area? It is, and we can figure that, because we know that the plates must be converging there. And people have done this sort of trenching and can recognize that the repeat time for large earthquakes there is about six or seven hundred years, if I remember correctly. Well ,we haven’t had historic recordings written down over that time period, so we don’t know from our own historic record how frequently earthquakes occurred, but from the paleo-record we can determine that. So we try to do that here too.

Now, we can go on to more modern studies, though. We can take GPS, and this is a little fancier than the GPS use to figure out if we are near a Starbucks or not, and we set them out and they sit there and record for days, in order to get really good signal to noise ratio, or sometimes the stations are there permanently. Eric Calais, who is a geophysicist from Purdue, and is the one most often mentioned in context with the prediction of this earthquake, has done a lot of this work. And he set stations out around Haiti, all of these dots are locations of GPS stations where they record them, first at one time and then they go back a year later and back a year later and so on, and they see how that station has been moving. How much has this station moved? How much has that station moved? Because they’re all moving on the surface of the earth, along with the surface of the earth. But if one is moving more and the other one’s moving less, than stress and strain is building up in between those two. So, we take a look at that. This figure shows us these arrows are slip directions from earthquakes, that we can determine. But these arrows are from the GPS re-measurements, and we see that okay, yeah, there is a bit of movement. Now, it didn’t move that far, hundreds of kilometers or anything. The scale is different, the arrows mean on the order of millimeters to centimeters of movement. And we see this quite a bit here, less there, less there, and less there, and then, we’re assuming in this case that North America is fixed, we’ve held this plate fixed, in order to make these calculations. Well, that’s very interesting. So, we can say that –how much stress is building on these faults and how much has to be released in a big earthquake. Let’s take a look at a cross section through Hispaniola and project everything onto that, from all of these stations. And this one is drawn a little bit differently; this is assuming that the Caribbean plate is fixed. We see that these stations have not moved relative to the Caribbean plate. These stations are on the Caribbean plate, they’re south of the Enriquillo Fault. These stations have moved more and those stations up there have moved more yet. It’s difficult to tell from the arrows, but if you plot them up like this, you see, okay, there is motion there. All of these south of the Caribbean plate are moving at the same rate, but as we cross the Enriquillo Fault, they are moving at a different rate. And as we cross the Septentrional Fault, they are moving at a different rate yet. Well, this is showing that there must be slip accumulating along that fault, because of the way that it’s gradational through there, of about seven millimeters a year and slip that’s accumulating along that fault that is not slipping, it’s strain that is building up there, of about nine millimeters a year, on the Septentrional Fault. That’s the strain that’s waiting to be released during the big earthquake. Presumably, because the numbers work out about right, this strain was released during that earthquake, so if we plot it again, we’d find these coming up to here and then those going there and right in between where the fault is, that’s where it really moved, the equivalent of seven millimeters a year, but all at once. It hasn’t done it here on the Septentrional Fault, yet, so of course we should be concerned about it slipping there. This is the basis of the prediction that the area was ready for a magnitude 7.2 at any time. And remember that prediction was made in 2008. Eric Calais did convey that to, in fact, even the President and Prime Minister of Haiti. He speaks French, he’s a native French speaker, and he advises the government on many issues. But what do you do in 2008, when you’ve already had buildings built there for the last 150 years; you can’t do anything that fast to fix it.

Now, what happened after the earthquake? Scientists went into the area after the earthquake, to do various types of studies. Some of them you might have seen on TV. They were looking for uplift and subsidence and deformation, searching for faulting at the survey. They did the GPS re-survey. They wanted to find out really quickly; how much did slip on the fault and how much slip is waiting to occur? Seismographs and strong motion recorders, remember there were no seismographs in Haiti, ahead of that. Marine surveys, part of that fault is offshore, under water. Can we find fault displacement there? Can we find slumping? Can we find land sliding that occurred in a marine environment, that often causes small tsunamis? Dating of coral heads that are now exposed, that’s part of the uplifting substance, and engineering surveys. Of course, engineers were around looking at what caused the damage to the buildings. Now this is part of the helicopter survey. Roger Bilham flew over the area and was looking for the fault start. Now we’re looking right down the valley, that is in fact caused by the Enriquillo Fault here. And you see signs of recent landslides, but nothing is connected, there is no actual start of the fault. It did not reach the surface and that’s a little -- people were a little surprised at that. They put a huge effort into trying to find it. Now, some people who were there just recently have said they might have found some areas where it did reach the surface and at a meeting next week, I’ll find out if that, in fact, has been borne out or not. But it did not, in general, reach the surface.


Roger Bilham was in the area, almost immediately after the earthquake. He went in as a private citizen, essentially, through the Dominican Republic, with a film crew from the Discovery Channel. And so, they went to the airport to get a helicopter and helicopters, of course, were busy with relief. But they found one that could go drop off the medical supplies they was doing and take a little bit circuitous route on the way back, looking for the fault. And that is how they did this, without being in the way of the relief operations. They were still able to do that.

The summary of Roger’s experiences there, what he saw -- the intensities of shaking; they’re given in Roman numerals here, to help remind us that they are not quantitative measurements, but a measurement of how it was perceived -- but also indications of what was observed about liquefaction, the soil can turn to essentially quicksand, a liquid, and boil up and structures that are on it can sink in. And that’s what happened to those cranes in the port. Many of you have seen the pictures, or maybe even seen the cranes themselves, that were listing and obviously sinking into the port, because the foundation on which that was made, the soil there liquefied and the thing just had no strength and it sank into it. And over here, places where coral heads were raised and some slumps and things like that, we’ll see some examples of that. This is a picture, area from the helicopter ride, of corals that are expressed. And see this boat over here. Now what boat owner would have somehow gotten his boat across these exposed corals and beached it over there. Of course not, this was all under water and that was under water at the shoreline, when that boat was put there. So these corals have raised up. They will soon die, if they haven’t, by now, died. And people go in, then, and try to measure how far up they are, how old they were, and so on, to see; “what is that uplift here?” So that we can get a better sense of how the fault moved, during that earthquake. We can actually do that from space, too. It’s difficult to do it where the corals are, because that’s -- the before image is underwater and we don’t know how far underwater. But this is a comparison of, essentially, a radar map of the area. Port Au Prince would be right here. So, it’s a radar map of the area. Subtract from that a radar map of the area a few months before the earthquake. And so we see the wavelengths of the radar where it doesn’t match up, where the ground has moved. And each one of these fringes is another wavelength. So, they can tell how far up or down it’s moved. Well, the up or down would make sense, if the satellites went straight over the fault, but we didn’t plan our satellites that way – we plan our satellites to go around the earth where they happen to be, more or less, and they are often looking kind of sideways at the fault. So, this is showing us which parts have moved closer to and farther away from the line of sight to the satellite, which is why there’s an arrow up there, telling us one direction. We can see that the fault must occur along in here and go offshore there. We know that it appears back onshore over there, someplace. If there were displacement at the surface, we would see all of these patterns disrupted drastically right at that location. And we don’t see that. So, it’s more evidence that the fault did not rupture to the surface, in spite of the fact that, if we just simply take a really good radar map of the image and don’t subtract it from any other previous one, we can see that fault as a huge scar running through the Port Au Prince, again is over there. The part that ruptured was from here, over and not at the surface.

These are all ways that we study the earthquake after it occurred and we try to understand how it occurred. Now what can we do to try to improve this? Well, first of all, we have to do something to improve hazard estimation, and second, it’s really become clear to more and more seismologists that the seismologists have to improve their communication, communication with those who do the planning.

So, the hazard estimation -- right now, there’s a proposal by the U.N. for a seismic project, that would be a pretty good sized one, initially led by Columbia University. And that would be a good group to do this. There’s a proposal for a Caribbean seismic workshop, in fact some people in the room know about this, to provide regional studies. Now, this proposal was written prior to the Haiti earthquake and it’s timely that it’s coming up now. But, the idea is to elevate the quality of seismological studies in the entire region, by, in a sense, introducing the seismologists to each other, developing networks of data sharing, and developing seismograph networks that cross political boundaries. Seismographs and seismograph systems are good at doing that sort of thing. And then communication and planning. Now, there’s a workshop being held next week, a couple of us will be there from this room. It’ll be in Miami. And the purpose of it is to get the scientists and engineers to sit down with the policy makers and land planners, including quite a number of Haitians, in order to make sure that the scientists and engineers are providing the information that the planners, civic people need. And if we aren’t, we need to go back to the drawing table, and hope to figure that out within those two days and then set a roadmap for what to do about that. As I said, there is a growing awareness among scientists and engineers, that it’s incumbent on them to make their studies accessible to the public. The problem has been, and I’ve been as guilty as anybody else, when I write a scientific paper on seismicity and tectonics of the Caribbean Plate, my co-author and I will publish that in the most reputable place we can find. And that paper specifically, we published in a special volume of the Geological Society of America. About as reputable as you can get. If I’m looking for tenure, I’m looking for promotion, I’m looking for other research grants to be approved by the National Science Foundation, that’s what I do. Do any of you, say, Haiti Task Team people and so on, look in journals like that for your information? Of course not. I’d write it in jargon, for the specialists to be impressed with my use of jargon.


I don’t write it for other people. We’re beginning to realize that we’ve got to do that. NSF has realized that and forces, as part of their criteria for proposals to be awarded, what they call the broader impact, some way of getting this information out to the public. Most scientists consider that to be a pain in the neck, but I think they are coming around, each disaster, one at a time, to realizing that part of our job is to get it out there. Now I think the part that is not working in this system, is the tenure and promotion system. Tenure and promotion couldn’t care less about this and we have got to work on that. Okay, but there is this awareness.

Let us take a look at -- if we weren’t doing really good seismological studies, and we just said let’s look at the seismicity that’s occurred here in the recent several years since the -- since really good seismograph models have been in place around world. Say 1973 - 2009. All the shallow events greater than magnitude 4.4. You say, “Hey! Haiti is a pretty safe place, not many earthquakes there. Let’s build crummy construction there and get away with it. It is okay.” And in fact, we might even come up with a sort of official map of the hazard in the area. And we will use colors where red is worst and yellow is the least bad, even less bad, yellow still has some risk. And we will say, “Hey, Haiti is pretty good. No problem.” But if we take a look at where the earthquake and the aftershocks actually occurred, “Oops!” Our time window for this study was not long enough to include earthquakes like this one, and our earlier maps showed us that. These earthquakes occurred in the 1700’s. So it takes some other sort of study, not just mapping earthquakes, to do that. Haiti is not an isolated case. I am sorry that there is no coastline here, to make it easy to see. This is northern Pakistan; this is the line of control between Pakistan, India, Afghanistan here and the Wakhan Corridor over there. Now this map, I think you can see it well enough from where you are, shows red up to here, shows red from there over to here. In fact, part of it is red, because of one of the big earthquakes I was in. That was right here. But, it shows whatever this color is, orange or something, not the greatest hazard, right there. Well, the star tells you exactly where the Pakistan earthquake of 2005 occurred and this is the series of aftershocks from it, telling you what area it covered. And look at that, it fills in between these two red areas. So again, the seismic hazard map, done from simple-minded approaches, is absolutely wrong. In fact, it is almost exactly the opposite.

Can we come up with another example? Well, Chile is on our mind; let’s take a look at Chile. And I think for this I will turn off the light, because the colors are a little harder to see. Here we see that it’s recognized as a strong, large hazard, south here, recognized large hazard here near Santiago and here in between it is considered to be lower hazard. That’s where the recent earthquake occurred. And I don’t have an aftershock map yet, to show you of that. But, if we look at where the shaking was the worst, it turns out it essentially spans the area between this red zone and that red zone. And in fact in this earthquake, the rupture did go in both directions out from that distance. So, it is not just isolated to an odd earthquake here and there. The simple approaches of just looking at where earthquakes have occurred, that we can count on from instrumental locations and saying those places are high hazard, that is not good enough. We need better studies to be able to do that.

There is a study proposed, it is already underway, various countries are buying into it and have to support it. The proposal is with us right now, in fact, waiting to be revised, so that it fits the criteria that we have got. But the part that we have been talking about here is the seismic hazard part. And this proposal, Gem Global Earthquake Model, is going to essentially farm out to all of the experts, the appropriate work to be done in computing the seismic hazard of each area. For example, if they were still worried about Haiti that work would have probably gone to Paul Mann and Harriet Collete [spelled phonetically], the people whose figures I have been stealing left and right for this figure, for this talk. And that would go into the seismic hazard in a much more quantitative way than anything that I have shown you. But it would also take into effect;where the people live, where the structures are, and how vulnerable they are to that, what is the social network and so on. They are less vulnerable in, say, Japan, than they are in, say, Haiti, for example. And put that into a socioeconomic impact and that total thing then, is the risk. This is a long multiple year project. The part that I am most expert in is this part and they are planning that just right. Presumably, people who understand risk and exposure can figure that part out, as well.

Let’s do a final comparison here, everybody likes to do this. What about Chile, what about Haiti? How does that comparison look? We quickly point out that the Chile earthquake had a magnitude 8.8, that is a big earthquake by anybody’s standards. A magnitude seven in Haiti is not all that big by most standards, unless you happen to live in crummy construction right over. But, this is a big earthquake, this is much smaller. But, this was a rural sort of area, it didn’t happen to hit the big cities much , and it was a little bit deeper. This was urban and this was shallower. So, it is hard to make that comparison. But we can take a comparison where the amount of shaking, and this is the Mercalli Intensity Scale that I mentioned before, how is the shaking perceived, and, in this case, they would call that severe shaking, with this color. And how many people were exposed to the area of severe shaking, this is product called Pager from the USGS. And we see that, for the Chile earthquake, three and a half thousand people were exposed to severe shaking. And we look at the Haiti earthquake. Well, 626 thousand people were exposed to severe shaking, but another 2.5 million, almost, exposed to violent shaking. So depending on how you want to look at this, well not depending, this is the answer. But, depending on how you want to interpret it. I have heard some people say that, “Well, the number of people exposed to severe or greater shaking in both earthquakes was actually quite comparable, three to three and a half million.” And yet, of course, in the Chile earthquake roughly a thousand people died, and in the Haiti earthquake roughly 230 thousand people died. So, the conclusion could be that you are 230 times more likely to die in Haiti than you are in Chile, for the same amount of shaking. But, that is a little misleading. We have to be a little more careful, because a lot of this three million actually experienced severe shaking – sorry, actually experienced violent shaking, about 2.5 million. And in Chile none did. So, we can’t really quite make that comparison. We might see it either way in the newspaperm but here you have got the real figures.

What we need to do in the long run, for any place in the world, is to come up with good seismic hazard maps. This is a map of California, showing the probability of exceeding a certain amount of acceleration, due to shaking from earthquakes. And then this is another map, zooming in at a very close up tiny little area, of how susceptible to liquefaction are the soils here and how susceptible to landslides are the soils on the slopes here. These sorts of maps then help -- this, combined with that, help land use planners determine to what sort of standards must buildings or any other structures be built. What sort of hazards exists and what would have to overcome them. You can build buildings to withstand huge earthquakes; it just costs a lot more, if they are big buildings. If they are small homes, it actually doesn’t cost much more, it just takes the knowledge of how to do it and the effort to make it right.

In the meantime, we don’t have such a thing for Haiti. People are working on it; I will see the first draft Monday, at the meeting. But, for demonstration purposes, a simple first pass for Haiti could look something like this. Take the Enriquillo Fault, take the Sententrional Fault, make a band about them and say a band about that is pretty high risk. A band about it farther away is still very high risk, but not as high risk. And maybe we extend it a little bit more, where we’re pretty darn sure that the area is susceptible to liquefaction, because we know that this is essentially a swamp filled in by streams coming from the mountains, and that is the type of soil that is very susceptible to liquefaction. So that might be a very crude first pass for a seismic hazard map, for planning in Haiti. And people are working on things like this and moving on from that. Now, a summary, seismologists did understand the risk in Haiti. Planners, apparently did not. Calamities planning and building codes failed in Haiti, for a number of reasons, but one of them is that there was a lack of awareness of the seismic risk. The threat of new events is in fact great; we can not underestimate the risk that exists on other parts of the Enriquillo Fault system. Scientists are currently trying to quantify it. It is research based; they will not come up with a definitive answer. But, we are all concerned about it. Many technologies come into play. I gave you a little guided tour of some of those. And scientific and engineering communication has to be improved because we, the scientists and engineers, know this stuff, but we aren’t getting the word out and the dialogue is not there; it is very poor, it must improve. It has to be a dialogue, not just the scientists and engineers trying to force feed their results to the population.

Well, thank you very much. It is one o’clock. I am sorry to have gone so long. If you have any questions?


[End of transcript]

Back to Top

Do you already have an account on one of these sites? Click the logo to sign in and create your own customized State Department page. Want to learn more? Check out our FAQ!

OpenID is a service that allows you to sign in to many different websites using a single identity. Find out more about OpenID and how to get an OpenID-enabled account.