Earthquake prediction

I_am_learning

The seismic waves of earthquake travels at around 5 km/s. Wouldn't it be possible to setup network of detectors through out the world, and pre-warn people about approaching earthquake in real-time?

Evo

Are you suggesting that once an earthquake happens that it starts traveling to other areas?

Bobbywhy

I had the misfortune to be asleep in my bed in Sherman Oaks (San Fernando Valley) of Los Angeles in January, 1994 during the "Northridge Earthquake". My building was totally trashed and I was uninjured, but in shock. (which is actually a "mental injury")

According to our information I was about 10 km from the epicenter of the earthquake. That translates to around two seconds warning time according to your numbers. Two seconds notice would not have helped one bit. Now those located 50 km away (ten seconds) only felt some mild shaking and no damage whatsoever.

There is a massive network of seismic sensors, especially along the San Andreas Fault.

“Getting information from the sensors is easy. Tying that information into a system where Inland Empire (metropolitan Los Angeles) residents might have as much as a 30-second notice that seismic waves were approaching is a much harder task. Such systems are already in place in Japan, but USGS geophysicist Doug Given, who is overseeing this project, said a similar network here is years away.”

http://www.pe.com/local-news/reports...fault-line.ece

Evo

That's terrible, glad you are ok.

Please remember to put information taken from other websites in quotes.

I'll admit to being skeptical about the benefits of earthquake warning systems. Of course it's too soon to know how much actual benefit they can be to the general population. Unfortunately those closest to the epicenter are unlikely to get any warning.

http://www.pcworld.com/businesscente...and_quake.html

I_am_learning

Of course, people closest to the epicenter won't benefit, but lets think about who can benefit and how much?
Suppose 8 Richter scale earthquake occurs in certain place. What is the usual radius of fatal damages? Although the radius might be small for developed cities with strong building, but for weak buildings of developing countries, it could be 100s of kilometers.

apeiron

You do get some advance warning from the faster moving, very noisy, P-waves ahead of the main S-wave shake. For an 8, that could be as much as a minute's notice.

Not much use for a public warning system, but enough to start thinking about where you're going to dive if you recognise the signs.

Edit: ooh, just had a 3.9 jiggler just then.

Astronuc

That's detection, not prediction.

Prediction infers knowledge in advance, preferably well in advance.

USGS is working on a warning system.
http://www.usgs.gov/blogs/features/u...to-california/

Separately, seismologists and geologists are attempting to develop predictive methodologies, but requires a lot more sophisticated monitoring and modeling.

http://earthquake.usgs.gov/aboutus/nepec/
http://www.ess.washington.edu/SEIS/P...rediction.html
http://www.certifiedchinesetranslati...ake/aki00.html
http://scec.ess.ucla.edu/~ykagan/perspective.html

More resources
http://earthquake.usgs.gov/learn/top...pic=Prediction

I_am_learning

Thanks Astronuc.
But seems like the networks are still regional, not global.

Evo

This is part of what confused me about your OP, what do you mean by "global"? These systems are very expensive, I read that japan's system cost ~one billion dollars, and not every country has problems with recurring earthquakes that would warrant such an expense.

Astronuc

Most networks are national. Instrumentation is placed in those areas deemed to be of greatest risk.

As Evo indicated, these networks are quite expensive. They require instrumentation and the communications infrastracture. If one reads the literature, one finds that one of the greatet challeges is 'where to put the monitoring devices' that provide some clue as to the state of the earth. Also, note that earthquakes can be a few km to several hundred km below the surface. We see typical well depths of a few km. The deepest well bore on record is only ~12 km. Ref: http://en.wikipedia.org/wiki/Kola_Superdeep_Borehole

There exists the Global Seismographic Network (GSN) - http://earthquake.usgs.gov/monitoring/gsn/
http://pubs.usgs.gov/fs/2011/3021/

The US has the Advanced National Seismic System (ANSS) - http://earthquake.usgs.gov/monitoring/anss/

One approach to placing monitoring stations would be to use hazard maps in order to locate instruments in places most likely to produce seismic events.
http://earthquake.usgs.gov/hazards/products/

However, faults are long continuous structures, and there may be numerous parallel faults in a particular seismic zone.

To give one an idea of how complicated establishing a montoring system can be, see howhttp://www.canterburyquakelive.co.nz/QuakeMap/AllQuakesSinceSept4/

The seismic activity around Christchurch was relatively quiet until Sept 4, 2010.
http://www.canterburyquakelive.co.nz/Quake100Years/
http://www.canterburyquakelive.co.nz/QuakeYears/

Now it's quite active.
http://www.canterburyquakelive.co.nz/

TheMadMonk

One of the other problems with such a system is that some of the areas that would benefit most from such a system don't have the infrastructure in place to make it worthwhile even if they could afford it and site everything properly.

It is useful in a highly connected society like that in Japan where a warning can be communicated to the population quite rapidly but in a lot of other areas you simply couldn't inform people fast enough to make system work as hoped.

Andre

Strange that in a thread about earthquake prediction the word "radon" gets no hits. But the radon concentration could be increasing in the atmosphere due to fissures and cracks opening, shortly before an earthquake.

There must be something useful in this google search

billiards

Note we are talking about "early warning" and not prediction as Astronuc pointed out earlier. An electronic signal travels faster than the seismic signal by some orders of magnitude. Therefore it is theoretically possible, and indeed practically possible to detect an earthquake locally and to issue an early warning to the rest of the world. A good system will respond to such a signal by shutting down power to electric trains and closing gas pipes, to limit damage when the seismic waves arrive.

An interesting example of this phenomenon has arisen due to the rise in popularity of social networking. Some people noticed that news of an earthquake in Japan hit Twitter before it was up on the USGS website. Of course, this is because people in Japan with their mobile phones felt the earthquake before whatever seismometers are used by the USGS to auto-update their earthquakes page, simply because they were closer to the source.

davenn

There is a global network
LISS

but for what you are thinking its not going to work because anyone at substantial distance isnt going to feel the quake recorded by the sensors anyway

My system records easily records the big events from anywhere around the Pacific and the huge events world wide, BUT I am not going to feel any of them unless they are LOCAL or REGIONAL in which case the global network isnt going to give me any warning as I will feel the event well before they detect it.

seismic wave speeds, just to set you straight....

P waves ---- 7 to 8 km / sec
S waves ---- 5 to 6 km / sec
Surface waves ---- 3 to 4 km / sec

all give or take -+ ~0.5 km / sec


P waves are no noisier than any of the other waves generated
That still totally depends on how far it is from you and for a local event.
The detection of the the P wave would still have to be by sensor if the event was more than ~ 50 - 100km or more from you, as their amplitude tends to be realitively small in comparison to the S and Surface waves.
have a look in this thread where I posted my seismogram of the Mw 8.6 Indonesian quake several weeks ago, look how small the P and S waves amplitude is compared to the surface waves.
Indonesian event

Its the surface waves that do the majority of the destruction. Local to the epicentre of the quake, the P and S waves are not much more than passing pulses

cheers
Dave

davenn

As I pointed out in my previous post .... early warning of shaking to the rest of the world is pointless as they are not going to feel the event anyway. The warning of impending shaking needs to be where the shaking is going to occur, and if that is close to the epicentre, say out to ~ 100 km then the warning is going to be from none to maybe around 13 sec. 10 sec or so is going to give people a chance to dive for cover but those close in still wont have time to do anything

Dave

davenn

Radon is one useful predictor, but like a lot of predictors, it is not reliable. that is, there isnt always an increase of radon emission in the days or weeks prior to a quake.
Most seismologists I have had any discussions with, and from others being interviewed on TV in doco's or on the new etc, pretty much agree that, in the mean time, short term prediction lies somewhere between very difficult to impossible.
The main thrust of research these days is to come up with ways to mitigate damage and deaths. Stronger structures and retro fitting of older structures with base isolation techniques, teaching earthquake awareness etc

Dave

billiards

Obviously by rest of the world I do not mean to imply that shaking will be felt everywhere in the world.