Big Earthquakes
Posted by Hig on 03 Mar 2010 | Tagged as: events, geology, maps, video
I got my PhD studying tsunamis and working with other scientists who study tsunamis. One thing that almost every scientist studying tsunamis has in common is that they’ve never actually seen one.
For several of my former colleagues, this changed when the tsunami from Chile spread throughout the Pacific. Andy recorded 7 distinct waves using a ruler he’d just purchased at Home Depot in Santa Cruz harbor. Jody (my former advisor) and Tanya watched ice shift in tsunami waves in the frozen harbor of Petropavlovsk, Kamchatka (Russia). Jody has been studying tsunamis since the 80s. In an email she told us all, “I am SURE I eyewitnessed a tsunami, for the first time in my life!”
Tsunami scientist Tanya Pinegina watches gentle tsunami waves in the ice in Avacha Bay, Petropavlovsk, Russia.
Photo by Jody Bourgeois
All I can say is that I may have seen a tsunami, albeit a really small one. On a beach here in Seldovia I watched the dropping tide. Did it drop a little faster during the last 5 minutes? Has it slowed now? The tide gauge in town definitely saw something… a wave several inches tall with a period of 10 minutes or so (link here, many gauges on this page, including a more easily seen example of a tsunami recorded by a tide gauge in King Cove a few hundred miles southwest of Seldovia).
I also set up my camera. I took a timelapse video, accelerating reality by 300x. I’ve watched this video many times now, and I think I can see slight variations in how fast the tide drops down the beach near the center of the image. Then again, maybe not… you can judge for yourself. (The boat in the distance probably isn’t being moved by the tsunami, as the period of its motion is only 1-2 minutes, as opposed to about 10 minutes.)
Chile tsunami timelapse in Seldovia, Alaska from Bretwood Higman on Vimeo.
The Santa Cruz harbor had a much more obvious tsunami, and there’s a small amount of video of the devastating tsunami in Chile.
What are the chances? Global occurrence of large earthquakes
The earthquake in Chile was a really big one. It’s amongst the largest ever measured, with the energy of a billion tons of TNT, enough to change the rotation of the earth. Decades pass without a single earthquake this large anywhere on the planet.
Does it seem like there are a lot of big earthquakes lately? Two recent deadly events, one in Haiti and one in Chile, have gotten a number people wondering if that is more than a coincidence.
In the case of Haiti and Chile, it almost certainly is just a coincidence. The earthquake in Haiti was a giant in terms of human tragedy, but as far as seismic energy, it was quite small in comparison to many earthquakes that have happened around the world lately. The USGS catalog records 16 earthquakes as large, or larger than the one in Haiti in the last year. The Haiti earthquake was large enough to increase the danger of other earthquakes on the same fault, but not large enough to influence tectonics a quarter of the way around the world in Chile.
However, the largest earthquakes – those over magnitude 8 – do seem to cluster in time. The three largest earthquakes in the 20th century, all magnitude 9 or more, occurred in 1952, 1960, and 1964. (Some catalogs list the 1957 Andreanof Islands earthquake as a 9.1, making four over 9 in that time range, but the USGS rates this one an 8.6.) After 1965, there were only two magnitude 8.3 earthquakes, and none higher until after the turn of the millennium. The statisticians have taken a look (Bufe and Perkins, 2005), and they don’t think that’s random. I plotted the data below, and you can judge for yourself.
We’ve been measuring earthquakes since 1900, and the recurrence of the largest ones doesn’t seem random. There’s a clump of large earthquakes in the ’50s and ’60s, and then a lull through the turn of the Millenia. Things have been more active again in the past decade.
Click the graphic for a larger version, as well as data and vector graphic file.
And the past decade has been a big one for earthquakes. There have been five earthquakes above 8.3, including the 2004 magnitude 9.1 earthquake in the northeastern Indian Ocean. Each increase of 0.2 in magnitude corresponds to a doubling in energy released, so the 2004 magnitude 9.1 released as much energy as 16 magnitude 8.3 earthquakes.
What does this mean? We only have a short instrumental record (since 1900) and there’s a lot of variability, so it’s impossible to know whether we have another magnitude 9 just around the corner. But it seems likely that the period of tectonic quiescence starting in 1965 and ending with the gradual increase in seismicity in the late 90s is gone. It’s no time to dally on the science, both old-fashioned paleoseismic studies, and maybe some new methods that can help warn of impending earthquakes. And it’s no time to skimp on education and good infrastructure that can save lives during an earthquake. Likely the biggest difference that led to far fewer people killed in Chile than in Haiti was better building standards.
Why do the biggest earthquakes come in clusters?
At this point we don’t really know.
For an earthquake on a fault to happen, there have to be two things in place: The fault has to be under stress so it can provide energy for an earthquake, and some point has to fail, triggering the fault to move and that energy to be released. Stress increases over time and eventual failure is inevitable, but exactly when it happens is dependent on that trigger, which can be very subtle. The point where failure begins is the hypocenter (directly beneath the epicenter on the surface of the earth) and the entire portion of the fault that moves is the rupture area.
One way to think of it is to imagine the fault as a large building. Perhaps it is an apartment building in Istanbul, and as new floors are illegally added the stress on the structure increases. This unstable structure has a lot of energy in it, all in the form of cement and other materials suspended high in the air by weak architecture. But when it finally collapses, that collapse starts somewhere. Perhaps a pillar designed for two stories and holding five collapses because someone uses it to tie up their dog. Now the nearby pillars and walls must suddenly bear more of the weight, and they collapse as well. The failure spreads from the original “epicenter” pillar, and consumes the entire building, analogous to the earthquake’s rupture area.
So was it the dog that caused the building to collapse, or the extra stories? I’d say the cause was the additional stress, while the trigger was the dog. Earthquakes are the same way… caused by gradually building stress, but triggered when some point gives way.
Bufe and Perkins, 2005, discuss how an earthquake in one area of the world might lead to another far away and years later. Their first possible explanation focuses on triggering, while the others suggest that stress might increase on distantly separated faults at the same time:
- Perhaps the vibrations from a distant earthquake increase the pressure of water trapped in pores in rock, and that pressure leads to a gradual weakening of the rock, initiating an earthquake.
- Maybe a slow pressure wave moves through the lower crust after a large earthquake, and this wave causes stress to increase on other faults that then fail.
- Perhaps large “silent” earthquakes, rarely detected because the deformation they cause is so slow, drive the system by running along plate boundaries and triggering their more violent kin.
- It could be that the changes in where and how much ice is in glaciers changes the stress on faults and leads to times of quiescence or greater activity.
- Or my favorite: Maybe the largest earthquakes change the stress on tectonic plates. As this stress moves around the planet it triggers more large quakes, until the crust has settled into a lower state of stress.
Question: are earthquakes an inevitable “side-effect” of a planet that can support life, or can you imagine a planet whose surface is stable enough that it never quakes the way ours does?
Just idle speculation: it had been nagging at me for a few days but I didn’t know who to ask.
Andrew
It would definitely be a problem for much of the life that is on our planet anyway. Tectonic processes that lead to earthquakes and volcanoes are responsible for chemical recycling of the crust, and for topography. It may be that some nutrient poor ecosystem in a global ocean could survive…
One way to look at it is to imagine what would happen if you shut off tectonics on earth and then watched what happened.
For a while (tens to hundreds of millions of years I’d guess) it wouldn’t be a problem. No more volcanic eruptions would reduce nutrient input into some ecosystems, but similar nutrients could be derived from mountain erosion.
Gradually erosion and deposition would lead to a planet dominated by ocean and broad coastal plains with rounded upland hills and low mountains. At this point the rate of erosion would be dropping off because the softer sediment would already be eroded, and steep slopes where erosion is enhanced would be reduced. This means that the mineral portions of soils, and dissolved minerals in groundwater, would be becoming less and less. Also without upland sources rivers would be sluggish and low oxygen. Beaches would be broad and muddy, and tides would be damped by broad shallow seas along the coastlines. This is a planet where diversity is reduced, but life is likely still abundant.
Next the land would vanish, eroded away by the ocean. The ocean floor is covered in a layer of clay, sealing off any geologic nutrient sources. Life would have to evolve to make use of a narrower spread of chemical building blocks, but would still have sunlight as an energy source.
This picture is built around the impossible scenario where tectonics on the earth get shut off. If you had a planet where the nuclear heat that powers tectonics was absent from the start, then there would be no topography to start with. Many scenarios for the beginning of life, and for life’s survival of hard times like possible “snowball Earth” periods, rely on geothermal springs. And without tectonic mixing of the earth’s mantle and crust much of the volatile water and CO2 might never reach the atmosphere… perhaps it would be a frozen dry rock.
One thing that might alleviate the issues of no tectonics is the occasional meteor impact. For a time after such an impact there would be fresh rock exposed, and heat from the impact would drive groundwater circulation and nutrient flow. Impacts can be a very large source of energy that is similar to volcanic processes… The K-T impact 65 million years ago likely lofted a huge quantity of sulfur into the atmosphere, similar to a very large eruption, and a much older impact generated the complex metal-rich mineral deposits at Sudbury… similar to the way magma intrusions can concentrate scarce minerals.
Nice post, Hig!
I didn’t know about the Bufe and Perkins article, but I had seen some plots of this work when I was visiting Denver USGS back in the late 1990s, and I remember reading Barbara Romanowicz’ article [cited in B&P] and thinking it made sense.
Which meant that, along with many others, I was waiting… I hadn’t focused on tsunamis and paleoseismology until the mid-1980s, but after that, I remember waiting for just any tsunami to happen. We were studying thin sand layers on the coast of Washington that might have been deposited by a tsunami about 300 years before, but what did a tsunami really do when it came ashore? Some people thought they would not carry and leave a deposit.
In January 1989 we went to the coast of central Chile to look at the record left by the 1960 giant tsunami. We interviewed eyewitnesses [29 years later]. Some of that information is available here:
http://pubs.usgs.gov/circ/c1187/
I instructed our “ayudantes” [assistants], who conducted the interviews in Spanish (and they said the local people, mostly of indigenous Mapuche origin, were hard to understand), never to mention the word “sand” (“arena” in Spanish) but just to ask what people saw when they came back to their [obliterated] homes. “Arena” was commonly the first thing they said. So in addition to digging and finding a sand layer on the coast, we had corroboration from eyewitnesses.
But still I waited for a tsunami to happen in real time. That came in 1992. There really was a lull in large tsunamis between 1983 Japan Sea tsunami and 1992 Nicaragua. We went and surveyed the effects, both to people and to the coast, the erosion and deposition.
After that, it seemed that tsunamis happened almost in cascading frequency –but not really–they just occurred about once a year, on average. Lots in the 1990s, but mostly only of local impact. And so we waited, still, for the next “big one” — and it happened in December 2004… but not in the Pacific! Now we have Chile 2010, and I can’t help but wonder if the next big one will be somewhere around the Pacific Rim and within the next decade. Posmotrim [we'll see, in Russian, as currently I am in Russia].
And, as you quoted me, finally I eyewitnessed a tsunami, though I never imagined the circumstances of watching a wave from Chilean summer gently move the ice under a frozen bay in the Russian Far East winter.
I agree completely about the usefullness of global visual onsite observations of natural disasters. Good work done here! But a global diagram about magnitude vs. age may misleading future earthquake forecasts for some regions of the “shaking” circumpacific plates. – Because magnitude by Richter is only one component to classify energy of destruction on surface. The reason why in last Chilean earthquake is less destruction in buildings than in Haiti depends also of depth, where the plate energy has been “liberated” (hypocenter, 3D model). Chilean case history: 7.8 magnitude 5 march 1985 with 20km depth with more destruction, 8.8 magnitude 27 february 2010 (10x stronger!), but similar or even less destruction on surface(!) Put that into account in your future statistics, please. OK? 4 march, Santiago de Chile
Great discussion both of your original article and the idea of a tectonic-less Earth. In regards to the discussion of clustering of large seismic events, give the short record of measurements and being pattern-seeking could we just be overprinting a pattern bias on the limited data set? Its an intriguing idea that large quakes “cluster” because the overall plate system is trying to equilibrate, but almost impossible (or hugely unweilding) to prove. Another possibility is that we are seeing in the record, a reoccurring “clumping” pattern in distributions that is so common in the natural systems.
Cheers.
Bob Chesson
Karsten: Magnitude (specifically moment magnitude, not Richter magnitude) is a measure of energy, and not of human impact. There is actually a scale for that too… the Mercalli Intensity scale (http://www.seismo.unr.edu/ftp/pub/louie/class/100/mercalli.html). It would be interesting to make a similar plot for that scale. You might find a long-term increase in intensity because of increased population, though that might be reversed in areas like Chile where greater seismic preparedness has been achieved through time.
Bob: The clustering is, statistically speaking, real. The approach that Bufe and Perkins use is to take the empirical probability distribution for earthquake magnitude, and generate a large number of “monte carlo” earthquake sequences using the assumption that the timing of earthquakes is random. Then they look through those and see how many include clusters or gaps like the ones seen in the historic record. It turns out to be very rare, for example only 0.5% of the time does the monte carlo simulation produce a gap as prolonged as the one from 1965 to 2001 in magnitude 8.4+. There’s always the possibility that the historic record is remarkably coincidental, but the chances of that are low.
Thanks. Great discussion. I tend to favor the concept (as do yourself) that the stress caused by the large quakes tends to propagate to other boundaries in an attempt to equilibrate. Seems to be an idea that could be tested, although it would likely be very expensive and time consuming.
Thanks Hig, really good and informative — thanks for letting us judge for ourselves. Looks like clustering in big quakes, but I couldn’t see the tsunami in your time lapse.
There’s an asst prof position open in Earth Science at APU….would love to see you there! You’d be a good fit.
I am wondering if the pressure from the huge earthquake in the Indian Ocean in 2004 caused pressure directly across the plate on the other side of the Pacific Ocean, where this huge quake just occurred. I ask because I work in environmental landscape design and I see similar ’causes and effects’ when working, only on a much smaller scale. I am wondering if we will see more shifting. This morning I am reading that the city of Concepcion moved 3 metres, this is not small and is it equal to, in pressure and size of what happened in 2004?
Amy, I suspect that in some way there is a connection between the 2004 and 2010 earthquakes. But it’s not quite that simple. There are other plate boundaries between Sumatra and Chile, and if there weren’t the sort of rigid motion you’re talking about would result in decreased stress in Chile following the Sumatra-Andaman earthquake in 2004, not increased stress. Also, on this large scale it’s not accurate to picture the earth’s crust as rigid, it’s more elastic…
Roman, do you think they’d allow me to live in Seldovia, take off snowshoeing every time it’s nice, and spend most of my time running Ground Truth Trekking?
After I asked those questions, I thought ‘Quinn, it’s not that simple’! Thanks so much for your offering to someone not in the biz or buzz!
I think you may be on to something…