Remember that MeFi post on earthquake prediction?

September 3, 2002 6:13 AM

Remember that MeFi post on earthquake prediction? They did it again. There was a 4.6 quake yesterday near Santa Barbara that hit the bullseye. Compare the map of where the quake hit to the prediction map. That's at least five accurate predictions since the scientific paper (pdf) was released earlier this year.
posted by insomnia_lj (10 comments total)

To correct myself for accuracy, the report contains more of an earthquake forecast than an earthquake prediction, but the fact remains that accurate forecasting of earthquakes could help save lives and might help us to predict quakes in the future. Cool stuff...
posted by insomnia_lj at 6:30 AM on September 3, 2002

>tufte<Well, some context is needed here before we herald in a new era of seismology. How many quakes did it miss, and how many false positives were there?

I'll take your word for the accuracy of the maps, but what are we looking at on the prediction map? There's no legend to define what I'm supposed to be seeing.
>/tufte<

Cool that we're getting better at this, though. Especially for places like Central Asia, which seem to have more disasterous quakes and less capacity to deal with them.
posted by mkultra at 6:38 AM on September 3, 2002

hmm... gotta work on those greater than/less than signs...
posted by mkultra at 6:39 AM on September 3, 2002

Excuse my geography, but how is Yorba Linda anywhere near Santa Barbara? The Santa Ana Mtns., perhaps.

And yes, this is certainly far from a prediction, which would be a lot more specific than "the period from 2000-2010" to be useful. I'm not entirely sure, either, that medium-intensity earthquakes appearing ... well ... where other medium-intensity earthquakes have appeared before yet proves the efficacy of this model, particularly since the map provided is for Richter magnitudes greater than 5, which this example wasn't. Insomnia, where were the other quakes?
posted by dhartung at 7:45 AM on September 3, 2002

Five accurate predications? How many other predications did they make that were incorrect?
posted by Captain Ligntning at 8:02 AM on September 3, 2002

why are we so willing to believe that they're forecasting these earthquakes? why aren't we considering that they could be causing them with some sort of secret satellite that was once stolen by evil guys who hid on a highjacked train but were stopped by a chef/ex-Navy SEAL?
posted by tolkhan at 10:14 AM on September 3, 2002

"Five accurate predications? How many other predications did they make that were incorrect?"

There's no way of knowing that until 2010, but as far as I know, there hasn't been a significant quake in California that it hasn't pinpointed yet.

Some of the quakes that it has predicted so far:
- a magnitude 5.7 quake on Feb. 22 in the Baja region of Mexico
- a magnitude 5.2 quake, near Gilroy on May 13
- the magnitude 5.1 Big Bear quake of Feb. 10, 2001 (which occurred after the work was completed)
- the magnitude 5.1 Anza quake of Oct. 31, 2001 (which occurred after the paper had been accepted for publication)

See this article for more details.

Even the scientists say that they expect the forecasts to be right only about 80 to 90 percent of the time... but that's still quite significant.
posted by insomnia_lj at 11:49 AM on September 3, 2002

tolkhan, isn't that the same satellite that makes crop circles?
posted by tomplus2 at 11:49 AM on September 3, 2002

"I'm not entirely sure"...."that medium-intensity earthquakes appearing ... well ... where other medium-intensity earthquakes have appeared before yet proves the efficacy of this model"

It's more involved than that. In many cases, it is using the lack of appearance of small, stress relieving earthquakes on faults to predict significant quakes. Some of the things that it doesn't forecast over the next 10 years are actually more interesting than those it does. For instance, it doesn't forecast a major earthquake for either San Francisco or Los Angeles. Apparently, the stress released by prior earthquakes have actually helped reduce the risk of major earthquakes in these regions.
posted by insomnia_lj at 11:56 AM on September 3, 2002

"In the final analysis, his method says there will be future earthquakes in areas that have had earthquakes in the past," says Mary Lou Zoback, who has worked in earthquake forecasting for the U.S. Geological Survey.

"They are basically the same places that have had major quakes over the past 10 years, and thankfully not in heavily populated areas," says euphorb who has absolutely no knowledge of the science of earthquake forcasting.

Look for the sun to rise in the east tomorrow and set somewhere in the vicinity of Santa Barbara.
posted by euphorb at 1:41 PM on September 3, 2002