Chris Rowan is a geologist specialising in the dark arts of paleomagnetism, and getting people to pay him to travel to exotic destinations for fieldwork. Having drilled up New Zealand during his PhD, and South Africa in his first post-doc, he now works at the University of Edinburgh.
Anne Jefferson has a love of all things water-related and blends hydrology, geomorphology, geology, and climate change in her work. She has a Ph.D. from Oregon State University and is now an assistant professor at the University of North Carolina at Charlotte.There's an interesting news story in Nature* about a distributed computing project with a seismological twist. The proposed aim of the Quake-Catcher project is to hack and collate data from laptop accelerometers - designed to protect the hard drive when your precious portable is about to meet the pavement - to detect and track the propagation of earthquakes
When a computer signed up to the program senses shaking, it calculates the intensity and pings the information back to the servers at Stanford in less than a second. If enough computers detect ground shaking in the same area, the system could send out a warning to users who haven't felt it yet that an earthquake is on its way, [Jesse Lawrence, one of the seismologists developing the system] says.
If it works, it will be the cheapest seismic network on the planet and could operate in any country. It wouldn't be as sensitive as traditional networks of seismometers, but Lawrence says that's not the point. "If you have only two sensors in an area, you have to have a perfect system. If you have 15 sensors in a system it [can] be less perfect. One hundred, one thousand, ten thousand -- your need for the system to be perfect becomes much smaller," he says. "That's really our approach -- just to have massive numbers."
I've mockingly questioned the actual use of getting a few seconds warning before (and some people rightly pointed out in the comments to that post that even those seconds are potentially very valuable), but I'd imagine the real value of such data will be in tracking how seismic waves actually propagate through particular areas - the effects of both sub-surface structure and topography are an important control on the shaking intensity in any particular place, and are quite difficult to model. But this sort of system is not without it's own computational challenges, such as producing software that will discriminate earthquakes from knocked desks, slamming doors and passing lorries.
Although I can't find reference to it on the Quake-Catcher website (their pdf brochure seems to be MIA), the Nature article also seems to refer to a distributed computing application in the more traditional, SETI@home sense:
Little is known about how seismic waves travel and refract deep in Earth's crust, and modelling this movement accurately takes enormous computing power, which can be generated by combining many different users on the network.
This is possibly referring to producing velocity models of the earths interior by processing teleseismic arrivals (earthquake waves from the other side of the planet that have travelled through the Earth's interior, as opposed to more local ruptures) from official seismometer networks; I wouldn't have thought that your lap-top sensor could pick them up such faint tremors from background, although it would be pretty cool if they did. I wonder if it would be practical to cheaply build an instrument that could?
*(I also see that Julian has already blogged about this, knows one of the people involved, and may even get to help beta test...)
Environment
Physical Science
Comments
Oh yay, the article's out!
I've actually been alpha testing the system since the beginning of February! So far, I've only had it running during one earthquake, and it didn't tell me that it had picked up anything unusual. I should ask if there's some way to view the computer's old output, rather than only the last minute, or as long as you can stand to hold down the scroll-back button.
It uses the BOINC distributed computing system, which is the same thing SETI@home uses.
One of the ways it supposedly discriminates between an earthquake and passing vehicles or rampaging cats or what have you is mostly from the number of computers reporting the shake. I'm guessing that makes an alpha test with a small number of computers tricky to manage, but in the long run, it'll only register triggers from a bunch of computers at once as possible earthquakes, and I think (don't quote me on this), it's supposed to cross reference with a bigger seismic network.
Posted by: Julian | March 27, 2008 3:10 PM
The significant problem I see with such a system is that laptops *are* portable. Unless they also have built-in GPS, localization can't be accurate.
Posted by: chezjake | March 27, 2008 4:38 PM
I was wondering about the laptop problem, because being part of the seismic network sounds interesting to me (if IE7 doesn't always eat up all my computing capacity). Not only are laptops portable, but my IP address doesn't correctly tell where I'm located, just where ATT happens to have their server.
Posted by: Silver Fox | March 27, 2008 10:07 PM
I'm not cool enough to be alpha-testing this, but I saw a talk a few weeks ago... anyway, they let you define a few locations/corresponding IP blocks, so the system will know if you're at home or work or your favorite coffee shop. The locations aren't as precise as seismometers, but even if they just pin you down to the nearest DSL hub, with thousands of computers they can still do an okay job.
I'm not convinced that this system will ever offer much in the way of performance improvements over conventional seismic networks, but even if it doesn't, it's a fantastically cool outreach project.
Posted by: Maria | March 28, 2008 1:34 AM
I've been slow off the mark on this one. Thanks, Chris, for bringing it to my attention! It sounds like a cool idea, though I'd have to think a little more about how to distill useful information from thousands of poorly located low gain accelerometers.
As for the distributed computing idea, after SETE@home came out, I thought I could use a similar system to exhaustively forward model Earth structure. An IT crazy friend and I even designed and constructed a platform-independent system to do just that. Since then (it was five years ago) access to big parallel machines and computing clusters has become much easier, so the project never really made it beyond the testing stage. Should anyone be interested, we wrote up some blurb about the system here.
Posted by: Alessia Maggi | March 29, 2008 2:06 AM