Public: Communication: Warning
Zoo-cam Earthquake Early Warning System   (+4, -1)  [vote for, against]
ZEEWS

According to reputable zookeepers (see link), animals can sense earthquakes before they occur. Imagine a system that recognizes simultaneously occurring unusual behavior in the zoo's animals, as is said to occur before an earthquake. It takes as input video feeds from cameras in each of the animals' cages/pens/tanks/lairs. Upon algorithmically identifying "unusual behavior", output is piped into a big horn which honks to warn the city, or perhaps to a more modern computer-driven messaging system. According to said reputable zookeepers, this will provide people with around 5 seconds of warning time before the shaking starts -- more than enough time to crawl under your desk.
-- swimswim, Aug 25 2011

Zoo mystery: How did apes and birds know quake was coming? http://www.washingt...QAZrXQcJ_story.html
[swimswim, Aug 25 2011]

How, indeed? http://en.wikipedia.org/wiki/Recall_bias
[mouseposture, Aug 25 2011]

2008 Lincolnshire earthquake http://en.wikipedia...olnshire_earthquake
Just a little one. [8th of 7, Aug 26 2011]

"Toads predict earthquakes: Official" http://www.theregis...1/earthquake_toads/
[swimswim, Dec 01 2011]

//algorithmically identifying "unusual behavior"//

Ah, there's an algorithm. That's alright then.
-- hippo, Aug 25 2011


I assumed those details would be obvious and boring. It borrows liberally from the airport security industry.
-- swimswim, Aug 25 2011


Would this not tempt a flashmob to simultaneously startle animals in zoos country-wide, such that much hilarity ensues?
-- MaxwellBuchanan, Aug 25 2011


I saw some unusually behaving animals about 53 hours ago and remembered the same fact about impending disaster. What I REALLY need now is some sort of natural signal that it's okay to come out from under my desk.
-- AusCan531, Aug 25 2011


[+] because it's a challenge to publish a report of these animals' unusual behavior *before* the earthquake, rather than after it is already know to have occurred <link>.

Worthwhile whichever result you get. [marked-for-croissant] good science.
-- mouseposture, Aug 25 2011


I don't accept the premise. Animals act weird all the time, but people only make note of it when something like an earthquake or tsunami happens later.
-- tatterdemalion, Aug 26 2011


I've met a fellow who can sense earthquakes coming.
The accuracy of his pre-quake reportings to authorities doesn't seem to factor in. (+)
-- 2 fries shy of a happy meal, Aug 26 2011


//there's an algorithm// Is that the same thing as learning the intonations of one of Al Gore's speaches? (probably an old joke - but I like repeating old jokes)
-- xenzag, Aug 26 2011


//I've met a fellow who can sense earthquakes coming.// This one's going to be 5.6 on the Rectum scale.
-- xenzag, Aug 26 2011


True though.
I gots ta tell it like it is...
-- 2 fries shy of a happy meal, Aug 26 2011


I went through several small quakes with animals around, and noticed none of them reacting early in any way. I also read about dolphins getting washed ashore and stranded by the 2004 quake/tsunami.

Still, if the animal-sensing did work, you not have to analyze behavior. You would only have to put a camera in each pen and analyze how much each succeeding image has changed from the one before (some video compressors do exactly that). If every animal was thrashing around at once, a computer would detect varying image values, and you could announce that it was probably feeding time.
-- baconbrain, Aug 26 2011


How do you know they are reacting to an impending earthquake, and not a volcano/rapture/invasion by anunaki from nibiru?
-- DIYMatt, Aug 26 2011


You wouldn't.
An animals actions just let you know when they are uneasy. They don't always know why they feel that way, and unlike us, they don't question it.

If they act funny prior to an earthquake without any other stimulus other than what they consider normal and it happens more than twice then... is that not proof enough?

I think it would be for most long time pet owners.

My wife and kids tell me our last dog of fifteen years would start pacing the entrance about ten or fifteen minutes before I came home from work every day, but since I never know my schedule, how could she?

Animals can do many things we don't give them credit for.
It doesn't jive with the scientific method though... so therefore, must not exist.
-- 2 fries shy of a happy meal, Aug 26 2011


Baconbrain, you basically nailed the gist of the algorithm, except that i had envisioned it scripted in Python -- and also, since feeding time is typically staggered across the multiples of beasts, the simultaneous thrashing of animals (and atypical locations, body temps, vocalizations, etc) couldn't be linked to that.

And MaxB, you are correct that the system would be an attractive nuisance; but the consequences of such hilarity inducement would be no different than pulling the fire alarm in a building, and would hopefully serve as an effective deterrent.
-- swimswim, Aug 26 2011


// Animals can do many things we don't give them credit for. //

Anecdotal datum: During the Linconshire earthquake of 2008 <link>, a peacefully-sleeping dog located more than 150km from the epicentre, suddenly awoke and went to "Red Alert", about four or five seconds before the humans became aware of a prolonged low-frequency rumble.

This is in all probability due to high frequencies, to which dogs are sensetive, propagating faster than low frequencies (audible to humans).

The point, however, is that - certainly at night - [swimswim]'s "five seconds" is nowhere near enough time to take protective action.

Thirty seconds would be the absolute minimum effective warning period.
-- 8th of 7, Aug 26 2011


Five seconds? It's enough time to put the gimp back in his trunk.
-- DrBob, Aug 26 2011


// During the Linconshire earthquake of 2008 <link>, a peacefully-sleeping dog located more than 150km from the epicentre, suddenly awoke and went to "Red Alert", about four or five seconds before the humans became aware of a prolonged low-frequency rumble.//

Hold on just one moment. Do you have any idea how many dogs there are living within 150km of the epicentre?

There are about 7 million dogs in the UK, on a land area of 245,000 square kilometres. A 150km-radius circle encloses about 70,000 square kilometers, and should therefore be home to something like TWO MILLION dogs.

If our dog is anything to go by, the average dog goes nuts for no apparent reason at least a couple of times a year. This means that, within the 150km circle, there will be four million "inexplicable dog going nuts" events per annum. In any given 5 second window, therefore, there should be approximately one "inexplicable dog going nuts" event.

What I find very, very disturbing is the fact that there are often earthquakes *without* this canine five-second warning. Based on the numbers above, it's pretty clear that dogs are deliberately keeping schtum in the run-up to most earthquakes.
-- MaxwellBuchanan, Aug 26 2011


I'm gonna hafta say no on this one too. My dogs act funny for any number of reasons, but I don't believe any mathematical formula could ever sort out 'dogs going bonkers because there's an earthquake coming' from 'dogs going bonkers because it's a full moon, or there's a storm on the way, or a squirrel just farted outside'. Granted, my dogs are idiots, but I'll put down good money that it's the same way with elephants.

Also, it occurs to me that there are all these people called 'seismologists' who have lots and lots of expensive scientific machines that _can_ tell the difference between an imminent earthquake and a squirrel fart.
-- Alterother, Aug 26 2011


Squirrels don't fart. Their diet lends to a non-gaseous digestive system. I been told.
-- blissmiss, Aug 26 2011


"Another County heard from …"
-- 8th of 7, Aug 26 2011


[blissmiss], thanks. I forgot about that. 'Squirrel farts' have long been used as a scapegoat for Unexplained Canine Idiocy in my family, despite their non-existence. None of you know about that, however, so referencing them here makes me look* even sillier than usual.

*figuratively speaking
-- Alterother, Aug 26 2011


//earthquakes *without* this canine five-second warning// The curious incident of the dog in the night-time.
-- mouseposture, Aug 26 2011


As my steamed colleague [mouseposture] points out, it is a widely-known fact that dogs cannot detect earthquakes after dark.
-- Alterother, Aug 26 2011


But, and this is important, they CAN look up ...
-- 8th of 7, Aug 27 2011


I love this concept for its creepy short story potential. I envision an Outer Limits type thing. You have several parties with potential: the humans who think they know what they are making, various zoo beasts, and the AI they create.

Suppose they do a teaching set for the AI, using prerecorded animal behavior for earthquakes, then a prediction set. Fine. But what if the AI signifies animal behavior occurring and no earthquake happens. But something else is happening. At first the humans don't realize the AI is predicting the something else. Then they figure it out. And maybe the animals aren't reacting to the natural event, but something else that they can perceive in the unseen world which knows about the event. And other events. Possibly causing the events.

I am reminded of an awesome ghost story about a parrot that mimicked people, and then began to mimick something it could see but that no-one else in the room could.
-- bungston, Dec 01 2011


//But, and this is important, they CAN look up ...//
Unlike humans in horror movies.
-- AbsintheWithoutLeave, Dec 02 2011



random, halfbakery