More than junk science?

April 15, 2004 1:11 PM

Quake to hit LA "by September 5," predicts a geophysicist at UCLA's Institute of Geophysics and Planetary Physics. Some skeptical, while others say it's not junk science.
posted by valerie (34 comments total)

nice 'drudge' link
thanks,
posted by clavdivs at 1:26 PM on April 15, 2004

This is actually somewhat interesting but one link to a news article is not that terrific.

Why not link to the original UCLA press release? Or to more information on the actual prediction algorithm, known as M8, given here and here?

Or, that this is the same guy who uses pattern-matching algorithms to predict the outcome of presidential elections?

Now that would have been a good front-page post... ;)
posted by vacapinta at 1:38 PM on April 15, 2004

From the article:

And he has made a chilling prediction that a quake measuring at least 6.4 magnitude on the Richter scale will hit a 31,200-square-kilometre area of southern California by September 5.

The question to be asking here, is how often an earthquake of magnitude 6.4 or greater has hit the given area, on average. Once a year? Once every ten years? The less frequently they happen, the more impressive this prediction is if it turns out to be right.

Well, this page doesn't give us exactly what we want, but it does give us a basis for some back-of-the-envelope calculations.

99 earthquakes of magnitude 6.0 or greater were recorded in the US in the period of 1990-2003, an averge of about 7 per year. The page doesn't break it down by tenths of points on the Richter scale; lets guess that half of those--3.5 per year--are 6.4 or greater. (If you want to get really into it, you can probably get a reasonably good function of the number of earthquakes vs. the magnitude based on the data here, but let's not get that in-depth.)

The number of quakes in Alaska vs. California vs. anywhere else is broken down only for quakes of magnitude 7.0 or greater, but for those we have 5/13 in California. If that holds for quakes down to 6.4, that's about 1.3 earthquakes of magnitude 6.4 or greater in California per year.

How many of those occur in the Southern California region in which this quake is expected to occur? I don't know.
Well, we can look at a map like this one to see where some recent small quakes occured. I don't know exactly what the 31200km² region the prediction is for, but given that that's about 7% of the state's total area, it looks like if we take about 7% of the state's total area, in Southern California, including as much earthquake-dense region as we can, that might include about 1/3 of the total earthquakes in the state. So let's say there's an average of 0.4 quakes per year, magnitude 6.4 or greater, in the given area.

That's per year, but the prediction only covers today through Sept. 5--144 days. So in an arbitrary 144 day period, we'd expect somehwere around 0.16 such earthquakes on average. Or, in other words, if I knew nothing about earthquakes beyond historical information about their frequency, and I made a prediction like this, I'd have a 1-in-6 chance of being right, by luck. So if he does turn out to be right, it's not conclusive evidence, but on the other hand it's not too shabby evidence in support of his hypotheses.
posted by DevilsAdvocate at 1:57 PM on April 15, 2004

c'mon, Arizona Bay...
posted by keswick at 2:00 PM on April 15, 2004

"New Madrid" by Uncle Tupelo, March 16-20, 1992

All my daydreams disasters.
She's the one I think I love.
Rivers burn and then run backwards.
And for her that's enough.

They all come from New York City
And they woke me up at dawn
She walked with me to the fountain
She held onto my arm.

Come on do what you did
Roll me under New Madrid
Shake my baby and please bring her back
Cause death won't even be still
Carom's over the landfill
Buries us all in its broken back.

There's a man with conviction
And although he's getting old
Mr. Browning has a prediction
That we've all been told.

So come on back to New York City
Roll trucks in at dawn
Walk with me to the fountain
And hold onto my arm.

Come on do what you did
Roll me under New Madrid
Shake my baby and please bring her back
Cause death won't even be still
Carom's over the landfill
Buries us all in its broken back.
posted by Mo Nickels at 2:05 PM on April 15, 2004

There's another useful issue in here. From a public service standpoint, should people have this information? On the one hand, he could be panicking people unnecessarily (although frankly, Angelenos should always be ready for a quake). On the other hand, is it his responsibility to broadcast this info if he feels he has a good scientific backing? Is it worth having a quake warning that's similar to a raised security alert, ie, we're not SURE exactly when something bad will happen, but we have some pretty good info that suggests that it will happen around here, around this time?

How accurate does he have to be before he meets social (as opposed to scientific) levels of expertise?
posted by synapse at 2:22 PM on April 15, 2004

Preposterous! Devilish witchcraft. If we go to our grave then God wills it so.
posted by stbalbach at 2:32 PM on April 15, 2004

Hooray. Bring on the California-hating. We've already got the obligatory "Arizona Bay" comment. Thar's real genius, keswick.
posted by scarabic at 2:38 PM on April 15, 2004

good lord, what if it happens during the coachella festival?? could the pixies reunion be THAT powerful????
posted by armacy at 3:25 PM on April 15, 2004

His techniques seem very impressive...until they're wrong.
posted by jacobsee at 3:46 PM on April 15, 2004

Will someone please explain the mechanics of a strike/slip fault system to all those "Arizona Bay" land speculators? :)

Or maybe, read this.
posted by zoogleplex at 4:42 PM on April 15, 2004

From the press release that vacapinta linked to:

Keilis-Borok is a member of the National Academy of Sciences, and the American Academy of Arts and Sciences, as well as the Russian Academy of Sciences, and the European, Austrian and Pontifical academies of science.

Wow. Those are some serious credentials. It's also impressive that someone this old (82) is still cranking out the research.
posted by mr_roboto at 4:44 PM on April 15, 2004

What? You mean SoCal wouldn't actually fall into the ocean?!?!?! What next? A president who lies to us?!?!?!

scarbic: in the future, i'll run all my comments by you first. oh, wait, what is it is everyone says to me about IraqFilter? oh yeah. "don't read it if you don't like it."

and i don't hate california. just southern california. it's a time-honored tradition amongst northern californians, so respect my cultural heritage, mmmkay?
posted by keswick at 5:00 PM on April 15, 2004

I didn't know until now that "temblor" is an acceptable alternative for "tremblor". But as a Southern Californian, I'm glad to have found out.
posted by Slothrup at 5:29 PM on April 15, 2004

And, of course, this has no connection at all to the upcoming NBC television movie, 10.5.

LA Times Story

I call shenanigans!
posted by dewelch at 6:09 PM on April 15, 2004

California is not going to slip into the sea.

Rather, it's all the rest of the continental USA that'll be going under.
posted by five fresh fish at 7:38 PM on April 15, 2004

Fools! NBC has seen this coming for weeks!

On preview...damn you dewelch! I blame that horrible Apprentice finale for my slow wit.
posted by graventy at 8:18 PM on April 15, 2004

From my old Murphy's Law poster I had when I was 11:

"Everything East of the San Andreas fault will eventually fall into the Atlantic Ocean."

Horrible finale? Sam offered Trump $250 grand if Trump would let him work for him!
posted by onlyconnect at 8:34 PM on April 15, 2004

Doesn't this guy realize that "nonlinear dynamics" and "chaos theory" are out? Piezoelectric crystals and magnetic field anomalies are the trendy new areas of fringe seismology! They even incorporate mechanisms for earthquake prediction by animals. Keilis-Borok is totally going to be snubbed by the cool nut theorist crowd at AGU.
posted by yami_mcmoots at 8:38 PM on April 15, 2004

And we'll sink with California/
As it falls into the seeeeeeeeeeeeeeeeeeeeeeeaaaaaaa
posted by Ufez Jones at 9:19 PM on April 15, 2004

DevilsAdvocate: According to The Southern California Earthquake Data center there has been exactly one earthquake > 6.4 since 1995 in the Mojave area and to the south, a 7.1 in 1999. This makes his prediction a little more notable if correct, I think.
posted by spacehug at 9:28 PM on April 15, 2004

Indeed it does. That's exactly the sort of thing I was looking for to base my calculations off of, but obviously I hadn't found it. Thanks!

Better still: I searched this database 1981-2003, mag >=6.4, Lat. 33-35N, Long. 116-118W (which covers an area more or less the area covered by the prediction, in the most active area), and found only two such earthquakes--one in 1992 in addition to the one spacehug found. Only two in twenty-three years. So predicting one in a period of less than six months would be pretty decent evidence for his method--only a 3-4% chance of hitting that purely by luck.
posted by DevilsAdvocate at 11:08 PM on April 15, 2004

Would there be any correlation between me moving to LA in the first week of September and this earthquake ? Do I need to worry about some terrible curse ?
posted by swordfishtrombones at 1:50 AM on April 16, 2004

That'll be the day I go back to Annandale.
posted by kirkaracha at 7:52 AM on April 16, 2004

good lord, what if it happens during the coachella festival?? could the pixies reunion be THAT powerful????

There's been some discussion about the earthquake prediction on the Coachella message boards. The general consensus seems to be that since the area is relatively flat and the on-site campsites are on a big open green field, an earthquake during the festival would probably be pretty cool, at least to the festival-goers.

Besides, many people there will be rolling in one way or another, earthquake or no. :-)
posted by Asparagirl at 10:14 AM on April 16, 2004

yami_mcmoots> Doesn't this guy realize that "nonlinear dynamics" and "chaos theory" are out?

That would explain why my desktop double-pendulum is suddenly producing completely consistent behaviour and is no longer sensitive to small differences in the initial state of the system. Right, I'm off to Vegas.
posted by snarfodox at 11:01 AM on April 16, 2004

snarfodox, has discussing the behavior of your desktop double-pendulum ever helped you get laid? If so I will retract my statement about the relative coolness of nonlinear dynamics and cute earthquake pet stories.
posted by yami_mcmoots at 11:18 AM on April 16, 2004

yami_mcmoots: you just have to find the right girl. Spend some time at a university campus.
posted by snarfodox at 11:26 AM on April 16, 2004

snarfodox: I am the right girl! I've got a geophysics degree from a very nerdy institution and am quite familiar with the various ways people use science-talk on the pull.

Invoking the counterintuitiveness of chaos theory (or quantum mechanics, or anything else) is Not Sexy. It is particularly not sexy as a defense against having to explain why you're doing what you're doing, or the physical mechanism behind it all, or why the scientific community hasn't yet heaped sufficient accolades upon your work, which is what seems to be going on in the press release. Mostly, though, chaos theory is old enough, and has been popularized enough, that any possible witty remark will be tinged with the banality of the butterfly effect.
posted by yami_mcmoots at 12:36 PM on April 16, 2004

c'mon, Arizona Bay...
You did mean the foothills while the California's desert being beach front property.
posted by thomcatspike at 4:09 PM on April 16, 2004

yami_mcmoots: there isn't a single branch of science without groupies. Reading your blog though, I share your Stephen "fails-to-cite-anyone" Wolfram pain. On the other hand you're (obviously) overgeneralising, but its funny, so I forgive you. Speaking of which, isn't geophysics the study of what we don't know about natural disasters?
posted by snarfodox at 8:05 PM on April 16, 2004

Why would Wolfram cite anyone? The man is a god.

Only a god of cellular automation lifeforms, of course, but a god nonetheless. As long as those little dots worship him, all is groovy.
posted by five fresh fish at 10:06 PM on April 16, 2004

Wolfram's book makes an excellent monitor stand, though! Very sturdy.

And snarfodox, if you're picking up chaos theory groupies, you must be unusually charismatic is all I can say. I was actually into geophysics for the study of what we don't know about places we'll never be able to visit...
posted by yami_mcmoots at 12:24 PM on April 17, 2004

Hey yami, you mentioned my bunyips! You're my kind of TECSOL A-2 Solvent, Anhydrous!
posted by languagehat at 1:59 PM on April 17, 2004