Successful Earthquake Prediction

An anonymous reader writes "Although a touch late, it appears that today's earthquake was successfully predicted by the Keilis-Borok team. (The prediction was covered previously on Slashdot.) Purists might argue that the gap between both distance and magnitude is too large to count in favor of the prediction, but this non-geologist is certainly impressed. Here's hoping they continue to receive funding."

Comments

Let me be the first to

[TykeClone]

Predict that there will one day be another earthquake in California.

And, sometime in the future, another hurrican will hit Florida.

Re:Let me be the first to

[Radical Rad]

I predict that Windows will be hit by a virulent new strain of computer virus. You heard it here first!

These guys are amazing.

[kagaku]

The prediction is for a magnitude 6.4 or greater earthquake to occur between January 5 and September 5, 2004, within a 12,440 sq. miles area of southern California ...

Last time I checked, this was called an educated guess.

One quake too north, too late?

[Justen]

The Agence France-Presse says otherwise, sadly.

It is true that seismologist Vladimir Keilis-Borok predicted a quake around this magnitude for this year. This quake missed his six-month timeframe by just over three weeks. And it is notably further south.

That would still be a little too close for coincedence for me, except the day before the deadline for the earthquake to occur Dr. Keilis-Borok announced the prediction was based on false data.

So was he covering his tail and reputation back then at the expense of being a little too north and a little too early?

We'll have to wait to see what he says, I guess.

justen

Re: One quake too north, too late?

[Justen]

Eek. Error correction: the quake was notably further north, not south.

Re:One quake too north, too late?

[Txiasaeia]

Well, we *are* talking about huge underground plates here, not the weather. I mean, come on, they can't even get the *weather* right where I'm from! Even though he missed it by three weeks, I'm still impressed. Hopefully this line of research can be focused so that scientists can say "Get the hell out of SF before 5pm tonight!" and save some lives, you know?

Re:One quake too north, too late?

[Graelin]

Hopefully this line of research can be focused so that scientists can say "Get the hell out of SF before 5pm tonight!" and save some lives, you know?

Of course, with such little time between warning and the real event we'd likely have as many (or more) injuries and deaths as a result of the rush out the Oakland bridge or the panic that would ensue.

Having lived in an Earthquake prone state (Alaska) and somewhere near Tornado Alley (Kansas) I have been through both events many times and I would gladly take Earthquakes over tornados. Earthquakes come and go with very little warning so there isn't much you can do but run under a door or hide under a table. They're also very short lived and don't usually do too much damage. Tornados, OTOH, can be predicted to some extent leaving you huddled in fear for hours while a major storm cell passes overhead. And if you're unfortunate enough to have that cyclone of death drop on you-you are guarunteed to lose everything you own and maybe (although it seems unlikely these days) even your life.

Today was a gentle reminder of home for me today. I was on the phone with a guy in Cali when the quake hit. I can't remember the last time I had to tell someone "Hold on, earthquake.......... ok."

Re:One quake too north, too late?

[Lars T.]

Any warning along the lines of "Get the hell out of SF before 5pm tonight!" would probably cost more lives in the ensuing chaos than it would save from the earthquake.

From a purist....

[hopemafia]

Ummmmm...no.

This isn't a success. The earthquake today was on a completely different segment of the fault, and was significantly weaker than the prediction, not to mention over three weeks after the generous nine month time period.

Don't get me wrong, I like the research, they've had some surprising success in the past, and I hope they continue. But even by the generous error margins allowed for primitive earthquake predictions, they're wrong this time...and if you asked them I'm sure they'd say the same thing.

And yes, I am a geologist.

Re:From a purist....

[WhiteBandit]

Ummmmm...no.

This isn't a success. The earthquake today was on a completely different segment of the fault, and was significantly weaker than the prediction, not to mention over three weeks after the generous nine month time period.

Don't get me wrong, I like the research, they've had some surprising success in the past, and I hope they continue. But even by the generous error margins allowed for primitive earthquake predictions, they're wrong this time...and if you asked them I'm sure they'd say the same thing.

And yes, I am a geologist.


As another geologist, I agree.

Not sure where the submitter got his information, but this earthquake isn't even related to KB's prediction *at all*. This quake occured 250 miles northwest of KB's proposed area. In fact, it fell within the range of the area for the quake he predicted last year (the San Simeon earthquake).

However, I would be willing to wager that this earthquake *is* the one that the researchers involved in the Parkfield Experiment have been waiting for, only 11 years overdue. Especially since that segment of the S.A.F. through Parkfield has consistantly produced M6.0's or greater roughly every 22 years from 1857 to 1966.

Re:From a purist....

[RockDork]

Another geologist here (as if you couldn't guess from the name).

I'd like to add that KB is not a seismologist, but rather a mathmatician. A seismologist, should he or she be so bold, would 'predict' earthquakes based on the physical processes that generate earthquakes. KB's prediction is based on pattern recognition in the distribution of microseismicity. Sure, there is merit in short-cutting the physical processes and looking at emergent patterns, but if so, then the tests of these 'predictions' should be held to a rigourous standard, because there is nothing else to test! KB's prediction (a.k.a. test of his model) failed. Period. Parkfield was not in the prediction area. Its not even close!

Re:From a purist....

[Madcapjack]

Yeah, I'm a little surprised the original poster saw a connection.

I like this list: http://pasadena.wr.usgs.gov/shake/ca/archives.html

Gives a good sense of earthquake activity.

Well

[Anonymous Coward]

The time, place, and magnitude predicted were all wrong, so I wouldn't call it a success.

I can only add

[Safety Cap]

What can Southern Californians learn from this prediction?
Southern Californians should take the Keilis-Borok predictions as strong reminders that earthquakes have happened and will happen again in the region.

As a geophysicist, I can say, "Um... duh."

Re:I can only add

[prof_peabody]

As a geoscientist, I can say "pffft, needs more work".

Prediction NOT correct...

[mOoZik]

Let's see...

1) They predicted the earthquake would be in the 9 months between Jan 5 and Sept. 5. Despite the giant "margin of error," they were off by almost a month. A prediction should not span nine months and still be wrong.

2)They predicted a magnitude of 6.4 or greater: the earthquake was 6.0. Again, they were wrong.

3)They predicted it would be within a 12,440 sq. miles area of southern California that includes portions of the eastern Mojave Desert, Coachella Valley, Imperial Valley (San Bernardino, Riverside and Imperial Counties) and eastern San Diego County. It seems to me that the epicenter of today's quake was located north of the vast area in which they predicted the earthquake.

To me, this is nothing more than a coincidence. They were off on all three and the mere fact that it took place in the state should not be inferred that it was a valid prediction.

Regardless, methods get better all the time. I am not opposed to this particular prediction method, just a bit annoyed that the slashdot submitter believes it proves the model to be true.

Re:Prediction NOT correct...

[Mr. Pillows]

I think a lot of you are missing the point. Many of you have pointed out that the time frame is wrong because he missed it by such and such. But one must consider the search space in which the predictions are in context of . . . all future time which is a large search space. It is obvious that the earthquake occurred outside of the time frame of the prediction (let alone the geographic region), but there is promise to this line of research for sure, especially in light of the small margin of error in light of the backdrop of the search space. And yes, it would be nice to get it down to a day so people take safety measures. To submit a more abstract statement, models (whether for earthquakes or materials science) such as these are growing in increasing importance as our evolving society learns to understand our surroundings.

Earthquke?

[Anonymous Coward]

What earthquake? What, did the whole Earth shake today or last night without me noticing?

UNsuccessful earthquake prediction

[Anonymous Coward]

I'm really disappointed to see Slashdot post something so obviously off the mark.

Say what?

[warrax_666]

A virulent strain of virus, no less? Your ideas are intriguing to me and I wish to subscribe to your newsletter.

Mod article overrated.

[Anonymous Coward]

'Nuf said.