An earthquake will destroy a sizable portion of the coastal Northwest. The question is when.And here's the caption below the accompanying illustration.
The next full-margin rupture of the Cascadia subduction zone will spell the worst natural disaster in the history of the continent.Now let me point out that the Yucatan peninsula is part of "the continent" and that a meteor hit the peninsula 65 million years ago (during the "history of the continent") and that event caused the extinction of the dinosaurs. See what I mean about sensationalism. As to the simplicity part, the article name checks Portland and Seattle. Why? Because they are big cities that people have heard of. Would they be destroyed by a Cascadia subduction zone earthquake happening off the coast of Washington and Oregon? No! But if you need to manufacture a mega-disaster, wiping out Seaside, Oregon just doesn't cut it.
So, if the New Yorker article is just the latest example of the "a giant earthquake is going to wipe out civilization as we know it (or at least a big chunk of it) and it's going to happen real soon now" school of journalism, what's the real story? That's what this post is all about.
Seismology, the study of earthquakes, has been around for a long time now. Its origins can be easily traced back to ancient China. And devastating earthquakes have been with us for all that time. Recently we have had about 20,000 people killed in Japan, Before that an earthquake in Haiti killed more than 100,000 people. Before that the Indonesian earthquake killed about 250,000 people. And that just covers a period of about a decade. It would be nice if earthquakes could be predicted so that measures could be taken to reduce the death and destruction. But earthquake prediction has not advanced much beyond what the ancient Chinese were capable of. Why is that? Let's take a deep dive and see what we can find out.
We now know a lot more about where earthquakes, particularly the big ones, come from. They come from plate tectonics. The earth is like an onion. It has layers. At the center is the aptly named core. The core of the earth is hotter than the surface. This is caused by radioactive decay, if you were wondering, but only the "hotter" part is important here. The core is surrounded by the mantle. Floating on top of that is a thin surface skim called the crust. Heat can relatively easily escape from the crust to space through the atmosphere. This causes the crust to cool and harden into the rocky material we are familiar with.
Everything we see is part of the crust. In round numbers it is 50 miles thick. That sounds like a lot but the radius of the earth is 4,000 miles so it isn't. The crust and the inner part of the core are solid. The rest is not. What that means is that it can bend and flow. All earthquakes originate in the crust. They are caused by the fact that the crust is stiff. Earthquakes are caused by parts of the crust grinding and breaking against other parts of the crust. And that's were the plates come in.
We are all familiar with the slate pavers that are sometimes used to create a garden path. If two pavers are jammed against each other it is easy to image grinding and breaking going on. And that's the simplest model of earthquakes. There is one other part. What's causing the jamming? Temperature differences cause convection currents. Material flows in an attempt to equalize the temperature. What's flowing in this case is mantle material. In some areas you have an upwelling of warm material. In other areas you have a sinking zone of cool material. Finally, material flows across the top of the mantle from the upwelling areas toward the cooling zones. This mantle flow carries parts of the crust along with it. It's really as simple as that.
So that's our modern model of what is going on. Mantle flow drags pieces of the crust along. This causes the crust to grind and break into pieces that smash and bash into each other. And the smashing and bashing causes it to break into chunks called plates. Remember, this has been going on for billions of years so it has had a lot of time to settle down into what we now see. This model is relatively new. It only dates back about 50 years. And as a general model it works very well. But it doesn't tell us much about the details. And even at this very general level of detail it misleads us in an important way.
Pretty much all of us think of the various plates like they are giant slate pavers. Our mental model is of really big slate pavers. And this results in us thinking the plates are rigid like your typical slate paver. It is strong and stiff and really big. Pretty much all you can do to a slate paver is to break it and that's hard to do. So when we think of the North American plate we think of something that is extremely strong and extremely stiff and hard to break. But a big piece of something doesn't behave the same way a small piece does.
Consider the two by four. If it is a few feet long we would consider it pretty unbendable. But how about one that is twenty or forty or sixty feet long? Now all of a sudden it becomes quite bendable. The same thing is true of rock. Thinking of a piece of rock as being just like the paver works just fine if it is a foot long or maybe even a hundred feet long. But what if it is ten or a hundred or a thousand miles long? The properties change.
And slate is slate is slate. But this whole grinding and breaking thing has been going on for a long time. If you look around, even over a distance of a few miles, you will usually find different kinds of rocks and soil with different amounts of strength and stiffness and breakability. So if we study plates that are big enough to encompass an entire continent we find that things are actually much more like a garden as a whole than they are like a single slate paver. In a garden over here you have some pavers. But over there you also have some gravel or some cement or a deck. And over there you also have soil or shrubs or lawn. In a garden you have all these different kinds of things in close proximity.
If you look at tens or hundreds or thousands of miles of land you have all these same kinds of differences. Materials vary and each material can have a different amount of strength or stiffness. And this makes it hard to generalize about the attributes of an entire geologic plate. They are not just like a big piece of slate. Instead the basic properties of the plate vary wildly from place to place. Our mental model of a plate steers us wrong. Scientists try to take this into account. But it is complicated and difficult and they are not as good at it as they would like to be.
Let's now look at earthquakes in more detail. Plates move. A typical speed is an inch a year. That might not seem like much but in a thousand years it amounts to over eighty feet. That too might not seem very fast but in 100-200 million years it can and did create the Atlantic Ocean. And would you really notice if two buildings a few miles apart got an inch closer together or further apart over the course of a year? You wouldn't. It has only been in the last few decades that it has been possible to measure distances accurately enough to detect changes that small that happen that slowly.
So plates are moving. and they are smashing and bashing into each other. What does that actually mean? Well, the edges of these plates are called fault lines. The most famous is the San Andreas fault. The San Andreas is a transverse fault. The relative motion is along the fault line. The North American plate is moving south while the Pacific plate is moving north. They grind against each other as they pass. But what about faults where the motion is perpendicular to the fault line? There are three general cases. Running down the middle of the Atlantic Ocean is the Mid-Atlantic Ridge. It is completely under water except in a couple of places like Iceland. It sits on top of one of those upwellings I talked about. So warm material rises and uses volcanoes to push material to the side. Crustal material is pushed away from the ridge and the Atlantic slowly gets wider. And, of course, if something is getting wider something else must be getting narrower.
Where plates are jamming together we have collisional faults. The entire country of India sits on the Indian plate. The Indian plate is moving north. And that causes it to collide with the Asian plate. The result is the Himalayan mountains. If two plates just smash into each other you get mountain ranges. What's happening in my neighborhood (and what was the subject of the New Yorker article) is a little more complex. The Juan de Fuca plate is smashing into the North American plate. But it's not just a smack. It is diving underneath it. This is called a subduction fault. This might seem less violent then the straight smack situation but it isn't. According to recent research this fault generated a magnitude 9 'quake in 1700. How big is that? The earthquake that killed 20,000 in Japan was a magnitude 9 'quake as was the Indonesian one. The Haiti 'quake was only a magnitude 7. So what do these numbers mean and how can a smaller 'quake kill more people? One question at a time
The magnitude numbers you see in the press typically run from 1 to 9 (often with a single decimal place added - 6.2) and are based on the modern equivalent of the Richter scale. Here's my handy dandy (and completely unscientific ) guide to the Richter scale:
- 1-4 - A seismometer registers it but in most cases people don't.
- 5 - You definitely feel it. It's like being in a "fender bender" car crash.
- 6 - It's like being on a roller coaster.
- 7 - It's like being in a serious car crash where the air bags go off and maybe the car rolls over.
- 8 - It's like being in a roller coaster that jumps the tracks at the top of the starting ramp and crashes to the ground. People get killed.
- 9 - There is lots of death and destruction. It's like what happens in a big disaster movie.
So why did 20,000 people get killed in the magnitude 9 Japan earthquake but hundreds of thousands got killed in Haiti in the much smaller magnitude 7 earthquake? That's the other thing. Generally speaking the farther you are away from the epicenter the less damage there is. The epicenter in Japan was about 50 miles offshore and about 20 miles underground. The Haiti earthquake was only about 10 miles underground and was only about 15 miles from Port-au-Prince, the Haitian capitol. Another factor was that the Japanese have stringent construction standards whereas the Haitians don't. The point is that the magnitude of the earthquake doesn't tell you everything you need to know. The New Yorker article was written as if the "big one" was happening right under Seattle and Portland. In reality it would happen about 50 miles off the coast and both Portland and Seattle are significantly inland. As far as I can tell Sendai, 81 miles away and the closest large city to the epicenter of the Japan earthquake, suffered little damage.
So far we are talking history, about earthquakes that have already happened. What about earthquake prediction? Well, that's a problem. Let me first get into the new modern scientific approach. We can now measure plate movements. The basic idea is a simple one. Faults lock up but plates keep moving. The rock should act like a giant spring, building up more energy as it gets bent by plate motion. Then a plate unlocks violently, i.e. an earthquake happens. Just like letting a big spring go, the rocks shift until they are no longer bent. There they lock back up and we start the whole process over again. If we can measure plate movements we can figure how far the rocks need to shift to relive the stress. That tells us how much energy is involved and that tells us how big the earthquake will be. That's the theory. Of course it is hard to figure when the fault will unlock but we should at least be able to figure out how big the quake will be when it does unlock.
But it turns out that the theory doesn't work very well. Satellite and GPS measurements allow us to measure how far land has or has not moved. That should allow the fault loading to be calculated and geologists routinely do this. But low load faults unlock and high load faults stay locked all the time. A lot of the why of this is still a mystery. Geologists have lots of theories but little in the way of methods to determine how best to sort through them. A big problem is that they can't see what's going on.
Studying the surface is pretty easy. If you fly over the right part of California you can actually see part of the San Andreas fault. But it is rare that surface features tell you much. Most of the interesting stuff is underground. It is theoretically possible to drill holes to see what's going on but the deepest hole ever drilled went less than 8 miles down and most holes go less than 5 miles down. Drilling is very expensive so it has only been done a few times for solely scientific reasons. The result is that the detailed characteristics of the rock in and around a fault are mysterious. Without these detailed characteristics it is impossible to predict how much stress the rock can take. Effectively, geologists are flying blind.
But at least geologists can accurately measure plate movement, right? True but that is less help than you would think. Again using the simple theory, plate A moves X feet while plate B moves Y feet. We can now do some trigonometry and figure out how much the rock needs to shift to relieve the stress, right? Well, what about small earthquakes? They could have moved things and relieved some stress. Ok so let's factor those in. The problem is that to do this we need a complete inventory of all the earthquakes. No problem. Cut to the seismometer records and we are good to go, right? In theory yes. In practice no. There is a world wide network that measures all large earthquakes. But this leaves out most of the small ones. The data from the big earthquakes is better than nothing but it is not nearly enough to do the accounting accurately enough to figure fault loading.
And then there are silent earthquakes. These were only discovered a few years ago. The satellites and GPS stations were catching rock motion that didn't seem to match up with earthquakes. So a spot where this was happening was studied by installing high precision seismometers locally. They picked up earthquakes that were so small and so consistent that they had been missed. But over a period of months they were causing substantial rock movement. If you are loading stress into a fault via plate movement but relieving it via swarms silent earthquakes then that fault is not going to unlock when it is supposed to. Scientists have a long way to go and they know it. So what else is going on?
There is the historical method. This method is used more frequently by scientists than people think. But in this and many other cases it is entirely appropriate. In this case you find out what you can about past earthquakes. Japan has been keeping records of earthquakes, at least the big ones, for a couple of thousand years. Scientists were able to pin down the exact date of the 1700 'quake that happened around here by consulting Japanese records (see below). It was a big 'quake but no one around here at the time was keeping records. And even in parts of the world where they have been keeping records for a relatively long time only of the biggest 'quakes got recorded until very recently. So we have only a little information about 'quakes in the written record. Fortunately, this can be supplemented by the geological record.
We keep getting better at interpreting the geological record. The first clue as to the existence of the 1700 'quake was uncovered by noticing some weird land formations out on the Pacific coast. Closer to my home is Lake Washington. It turns out there are two separate forests at the bottom of the lake. They got there because two different land slides moved a bunch of trees from the side of a hill into the lake where they promptly sank to the bottom and got preserved. Tree rings dated each slide to within a couple of years and both slides were caused by earthquakes. This is now a recurring pattern in my neck of the woods. Geologists get access to better tools. They look around and find more evidence of earthquakes. And often a newly discovered earthquake is associated with a previously unsuspected fault. It turns out there are faults all over the place. This keeps making a complicated situation even more complicated.
Let me stop for a minute and summarize. We think we know the general idea. Plate tectonics push things around. This causes stress to lock into faults. The faults unlock, break, if you prefer, and we have an earthquake. So far so good. But in order to predict earthquakes we need a lot of very detailed knowledge. That detailed knowledge would allow us to predict how much rock is going to break (size), when it is going to break (timing), and where it is going to break (location). Geologists can currently do this in only the most general way. They look at the historical record. If a place has had an earthquake in the past it is a good candidate for a future one (place). The same indications give them a general idea of how big a 'quake is likely to be (magnitude). They can also use the geological and historical record try to calculate a "repeat rate", how frequently 'quakes take place.
If there are no 'quakes in the historical and geological record for a location then it is likely to remain earthquake free. On the other hand, 'quakes routinely appear where they are not supposed to so calling an area 'quake free involves a bit of guessing. The record turns up lots of 'quakes in some places. So this should yield a solid repeat rate, right? Unfortunately, even in the most earthquake prone places the timing of 'quakes is very irregular. Scientists talk of a 'quake being overdue but this is more dressed up guesswork. So how good are geologists at this kind of guesswork? They are good enough that you should incorporate their guesswork in building codes but you shouldn't take any predictions about exactly when the next one will show up very seriously. And it is also important to keep in mind that proper engineering can not completely earthquake-proof a building even if the 'quake is within the design range. But even if the 'quake is outside the design range it will provide some protection.
The different recent experiences of Haiti and Japan bear this out. The Japanese are perhaps the best earthquake people in the world because they have so many 'quakes. But even they did not plan for a 'quake as big as the recent one. The Japan 'quake was larger than anything in Japan's historical record. And it is very expensive to earthquake-proof against very large 'quakes. Even though it was not enough, the level of preparedness in Japan substantially reduced the death and destruction. In Haiti's case, not that much money and decent building codes,effectively enforced, would have saved literally hundreds of thousands of lives.
Finally, let me turn to a related subject, tsunamis. The Haiti story is not a Tsunami story but both the Japan and the Indonesia stories are tsunami stories. The whole Fukushima nuclear disaster and much of the Indonesian death and destruction were cause not by the earthquake but by the tsunami each 'quake generated. So what's the state of the art with tsunamis? There are problems here but things are in much better shape than they are with 'quakes. First the theory. It's pretty simple. When an earthquake happens the land gets thrown around. It this happens on land that is pretty much that. But if it happens under water the water above the epicenter gets thrown about too. Throwing a lot of water around causes tsunamis. Its as simple as that.
Well, there is a little more to it. Let's start with the bad news. How much water gets thrown around and which way does it go? This is the mystery part. Most of the guess part of this process involves guessing the location and other specific details of the earthquake. That is part of the magic and mystery of earthquakes. But let's move on. What if we know (or can guess) the exact earthquake specifics. What then? Then we are into the good part.
Oceans have been tossing giant waves around for ages. So scientists have been able to thoroughly study them and they have a good understanding of them. This means that scientists can predict their behavior very accurately. That's good news. The bad news is that water is very efficient at moving large waves long distances. So a tsunami can devastate shorelines thousands of miles from its epicenter. Scientists can accurately predict where the tsunami is going to go and how long it is going to take to get there. They just can't do anything to stop it. A tsunami can do a lot of damage a long way away. On land even a giant earthquake tails off to nothing within a few hundred miles so the damage mostly happens close to the epicenter. The same is not true when a 'quake throws a big tsunami.
With a few measurements from the tsunami close to the epicenter scientists can very accurately predict what will happen. Scientists were caught flat footed in the case of the tsunami associated with the Indonesia 'quake. But even so they were still able to get the measurements they needed and to provide some decent guidance as to what would happen. The big problem turned out to be that there was no system in place for passing this information on to the relevant authorities. By the time of the Japan 'quake and associated tsunami much better procedures were in place. And outside of Japan itself the size of the tsunami was much more manageable so the damage outside Japan was minimal. So the modern approach is to measure where it happens and a few other specifics. From there computer models will be able to predict the where, when, and how big of the tsunami.
And about that 'quake of 1700 I was talking about above. It was an underwater 'quake that created a large tsunami. That tsunami crossed the entire Pacific Ocean and eventually hit Japan. At that point it was still powerful enough to do enough damage that the Japanese put it in their records. That information and a computer model of how long it took for the tsunami to cross the ocean allowed scientists to pinpoint the exact day the 'quake happened even though no record of it exists on this side of the ocean.
If a similar sized 'quake to the 1700 one happened today at roughly the same location, it would likely rattle dishes in Seattle and Portland. It might even do some damage to the two cities. But it would not do the kind of damage outlined in the New Yorker story. And such a 'quake could and probably would create a large tsunami. And that tsunami would likely wreck havoc on the coastal communities of Washington, Oregon, and other coastal cities around the Pacific. But there are geographical barriers that would prevent it from doing much damage to Seattle or Portland.
There is a way to wreck havoc on Seattle to the extent outlined in the New Yorker piece? (I don't know the geology of Portland well enough to answer this question with respect to that city.) Yes! It's simple. You just have one of the several faults that passes directly under Seattle rupture in a big way. There is ample evidence in the geological record for this happening in the past so it could certainly happen again. And the amount of death and destruction resulting would satisfy even the likes of Ms. Shultz. Mayhem of biblical proportions is possible. But it wouldn't have happened the way Ms. Shultz outlined. It also wouldn't have been a "really big one" sized earthquake. I guess telling it that way just wouldn't have made as good of a story.
No comments:
Post a Comment