Slashdot Mirror
GPS Could Speed Tsunami Warning
wwood_98 writes to tell us that Wired is running a story about how GPS could serve more than its traditional role. From the article: "International organizations like the Pacific Tsunami Warning Center, or PTWC, in Hawaii currently depend on coastal seismic stations to record deep-sea earthquakes that could cause giant waves. But according to Jeff Freymueller, a geophysicist at the University of Alaska Fairbanks, data from GPS receivers could provide quicker, more accurate estimates of the magnitude of a tsunami-causing quake, buying time for evacuation. Freymueller presented his findings at this week's American Geophysical Union conference in San Francisco."
Excellent, if you live in a coastal city, you'll get to know that you have 20 minutes left to live...
MoM++ - A Classic Expanded - [Master of Magic 1.5]
http://mompp.sourceforge.net/
But looks like there may be a way to detect a tsunami caused by an under sea event;
Sounds promising but is it possible to tell the difference between air movement caused by a tsunami and just a sudden gust of wind? How dense would the sensor array need to be to prevent false positives?
Does anyone have any further information regarding what grade of GPS is being used? Based on the price point, I would have to guess that it is Mapping grade GPS equipment, but the problem would be that in order for Mapping grade GPS to update, the person (or mount) holding it, must be moving to get the corrections from the satellite. Survey grade may work, but it wouldn't be under $10k per receiver.
Any help?
From "Wiki: "The accuracy of the GPS signal itself is about 5 meters (16 ft) as of 2005 and has steadily improved over the last 15 years. Using differential GPS and other error-correcting techniques, the accuracy can be improved to about 1 cm (.4 in) over short distances.". Given a fixed GPS sensor, the relative accuracy in ground measurements would be more than adequate for seismic warnings.
I don't think the core idea is to be able say "In 23 minutes and 12 seconds a tsunami of height 13.6 meters will hit blah blah blah...". I believe they are building a warning system - "Within the hour conditions will be extremely favorable for tsunami conditions to occur, and residents should take precautions...".
From TFA:
k .html
s sen_etal2002c.pdf
"With GPS, the displacements are measured second by second," said Bock, who also presented at the American Geophysical Union conference. "Within 70 seconds you have a good idea of the final deformation." In addition to predicting tsunamis, he thinks GPS modules could be used to monitor the activity of volcanoes and landslides in real time. [emphasis mine]
I thought GPS was already used extensively in volcano studies.
http://vulcan.wr.usgs.gov/Monitoring/GPS/framewor
No specific mention of real time data whacking in that link? A quick Google, and we find this, for example:
http://www.gmat.unsw.edu.au/snap/publications/jan
Computational Chemistry products and services.
The article is interesting in the use of GPS recievers to gather information. Let's look at two datasets.
From Wikipedia: "The accuracy of the GPS signal itself is about 5 meters (16 ft) as of 2005 and has steadily improved over the last 15 years. Using differential GPS and other error-correcting techniques, the accuracy can be improved to about 1 cm (.4 in) over short distances."
From NASA: "Large earthquakes often cause permanent movement of the Earth's surface, a result of the motion that occurs deep underground. The tsunamis spawned by the magnitude 9.0 earthquake on December 26, 2004, were the result of motions of the sea floor above the earthquake fault. Seismic measurements and computer models show that the Burma Plate slipped up to 20 meters (66 feet) at the location of the earthquake, 18 kilometers underground. The sea floor above moved less, up to 5 meters (16 feet) vertically and 11 meters (36 feet) horizontally."
So, the practical uses of this, even without error-correction, are theoretically viable for creating an early warning system for Tsunamis.
The article states that it should only really take 70 seconds for "a good idea of the final deformation". Linking this data to website and government run servers, the early warning system for Tsunamis would be far greater and accurate that say, tornado early warning systems. Consider the following exerpt from PBS's NewsHour: Developing a Global Tsunami Warning System: "STUART WEINSTEIN, Geophysicist, Pacific Tsunami Warning Center: I think the 'holy cow' moment didn't occur until we started getting the first preliminary reports over the wire services that, in fact, a damaging wave struck Phuket, Thailand and Sri Lanka.
BETTY ANN BOWSER: Were you frustrated?
Stuart WeinsteinSTUART WEINSTEIN: Very frustrated. Frustrated and to a certain extent humiliated. It's humiliating for me as a geophysicist working for a tsunami-warning program to learn first of a tsunami from a wire service than from a tide gauge. That -- it doesn't get any worse than that, quite frankly.
BETTY ANN BOWSER: Thousands of miles away at NOAA's Pacific Marine Research Lab in Seattle, tsunami researcher Vasily Titov was also frustrated. It took him until 4 a.m. in the morning of the next day to run this computer model, because he didn't have tsunami readings either."
Considering the earthquake hit at 00:59 GMT, and the wave first makes landfall at Sumatra 01:30 GMT, then 02:30 GMT in Thailand, then 03:00GMT in Sri Lanka and India... having a result from this system at 01:00GMT (70 seconds) automatically piped to the national emergency centers of governments, could have at least mobilized aid faster in Sumatra, and could have evacuated thousands in Thailand, Sri Lanka, and India.
A total of approximately 275,000 died in the Indian Ocean Tsunami of 2004. At a cost of even $10,000 per detector, 5000 detectors for $50million USD would have only cost $181 for every person that died.
I8-D
"In 23 minutes and 12 seconds a tsunami of height 13.6 meters will hit blah blah blah...".
Actually it is the core idea.
Tsunamis not caused by weather. They are generally sparked off by earthquakes, which we cannot predict reliably. However, once one is sparked off and we know the nature of the event(what kind of earthquake, magnitude, location) we can predict the tsunami once we crunch the numbers. However, this takes time.
Tsunamis are basically shock waves. They travel at 500-1000 kilometers an hour while in the deep ocean. That means that you are going to have much less than an hour in most situations. If you're talking over an hour of travel time you're talking about a minor tsunami for anything less than a huge asteroid hit. Near shore it slows down to 'dozens of kilometers' as it builds into a huge wave.
Seconds count in this situation. We're talking sirens annoncing the need to take immediate shelter, or travel inland/out to ocean. The most dangerous area is the coast region. If you're far enough out to sea, the wave can be handled, if you even notice it. Coastal mega-hotels need to be built strong enough to take the wave, though I'd see evacuating all of the ocean side and lower level rooms.
20 minutes can save a huge number of lives, though. A moderatly healthy person should be able to walk at least a mile and a half in that time. It's also the sort of situation where you're likely better off walking than trying to take your car. So many other people will be trying to drive, panicing, that accidents and traffic will slow you to a stop.
I don't read AC A human right
Until the tsunamis get ahold of these.
I think this article (and the researcher) missed the point of a far more profound, yet more far-off, use of GPS in earthquake detection. Spectrum recently ran a story describing how the GPS system among other things could sense impending earthquakes days, weeks, and even months in advance of an earthquake. The basic concept is the cracking of rocks in the crust leading up to an earthquake frees electrons which rush downward creating positvely charged holes. These holes also appear at Earth's surface and attract electons out of the ionosphere. The phase difference of the two signals the GPS satelites transmit is affected by this change and the change from normal values can be detected. This is still a long way off, and the use of GPS to measure this phenomena is only one and definitely not the best. But it sounds better than the couple hours mentioned in the article.
Check out the Southern California Integrated GPS Network for an example of highly accurate (sub millimeter) uses of GPS over a large area. The receivers being used in this network are 10 years old and still returning excellent data which we can use to compute annual tectonic plate motion. In 1999, after the Hector mine earthquake, we were able to determine 17cm of slip at a sight 40km away from the epicenter. This was done in a few hours.
Also, take a look at the Plate Boundary Observatory which is being built now.
--Keith
Just wanted to add that these GPS measurement stations put up for this purpose could give valauable information that could also be used for other many purposes, I can mostly think of geophysics purposes as that is what I work with myself, but I imagine such a web could be used for many other things
I remember a paper about the isostatic rebound after the icecap in Scandinavia where GPS recievers were used. A curiosity I remember from the paper was that at the coastal areas there were quite a lot more uncertainty on the vertical movement, which the authors said probably was caused by the bigger amount of snow that assembles on the GPS devices and thereby causing refraction of the waves used for the GPS measurement
GPS information could is also used to determine the absolute movement of plates and I imagine that this system could also with time, be used to predict big earthquakes more precisely and thereby give even more time to prepare for the tsunamies etc.
The postprocessing software is GAMIT (GPS at MIT)
--Keith
If this GPS idea works as advertised, it'll be a very useful incremental improvment to the tsunami warning networks. But it'll hardly be revolutionary. Being able to determine which events will produce a tsunami just from the seismic data - that would be revolutionary.