Magnitude:
From: - To:
Location: John Vidale
Top Latitude: - Bottom Latitude:
Right Longitude: - Left Longitude:

some colorful quotes:

San Bernardino Sun
newspaper on July 17:

Cracking the quake code

Forecaster at UCLA goes for three out of three in temblor predictions

By DAVID SCHWARTZ, Staff Writer

Prediction of earthquakes has, until recently, been left to the crackpots.
Scientists have been stymied.

The last time they tried it, the people in lab coats couldn't even hit a
nine-year window. That window since has stretched to 19 years and counting.

Thanks to pattern recognition and Russian geophysicist Vladimir
Keilis-Borok, all that could change sometime between now and Sept. 5.

The 82-year-old scientist is predicting that a magnitude 6.4 earthquake has
a 50-50 chance of striking somewhere in a 12,440-square-mile area that
includes the eastern Mojave Desert, the Coachella Valley, Imperial Valley
and eastern San Diego County.

What's different this time is that Keilis-Borok is going for a hat trick.
He's already been right twice.

And other scientists are paying attention, if somewhat skeptically.

"Either he's onto something or he's very lucky,' said Egill Hauksson, a
seismologist at Caltech Seismological Laboratory. "I think it's just
chance.'

Lucy Jones, scientist in charge of the U.S. Geological Survey for Southern
California, said, "One thing is different.

"This is the first time since (a failed 1985 attempt at quake prediction)
that the scientific community didn't say this is nonsense. (The California
Earthquake Prediction Evaluation Council) looked at it and didn't say this
is a crackpot. That is new.'

"It's certainly a good experiment,' said John Vidale, seismologist and
director of the UCLA Institute of Geophysics and Planetary Physics. "He's
trying it the right way, and I think there's something to it. I'm not sure
it works quite as well as he thinks it does.'

Even a success rate of 50 percent, which Keilis-Borok has said is the chance
of an earthquake hitting the region, i s significantly better than the 10
percent likelihood that is predicted now using historical data.

True or not, the prediction for a 6.4 or greater temblor has people talking
again. The public's eyes widen. Government officials offer cautious words
that, when decoded, translate to: "He could be right, he could be wrong.
Don't panic.'

Keilis-Borok's team p redicted two other major earthquakes. In September,
Hokkaido, Japan, was hit with a major earthquake, satisfying a prediction
the team made for a very active fault. Then, in December, an earthquake hit
Central California, another one that the team predicted.

If prediction No. 3 comes true, Californians will be relieved to learn that
his next two prognostications are targeted for sites outside the Golden
State. Keilis-Borok just hasn't said where yet. Colorful history

Earthquake prediction has had a colorful history. Scientists believed they
were close, only to be proven wrong.

In the 1980s, the U.S. scientific community peaked out when it was predicted
that within nine years a large earthquake was 95 percent certain to hit an
area centered around the small Central California hamlet of Parkfield.

Nine years turned into 10, then 15. Now, 19 years later, there is still no
earthquake.

Earthquake prediction fell out of favor, with many arguing that it was
nearly impossible to develop a method that could reliably predict an
earthquake. U.S. graduate students, who had once thrown themselves into the
specialty of earthquake prediction, turned to the more practical world of
understanding earthquake behavior, to better learn how to build buildings to
withstand the next major temblor.

That left the art of prediction to those who would thumb through the
newspaper classifieds, tracking the number of pets that have run away from
home as a precursor to an earthquake. Those who think the ringing in their
ears is a signal that the Big One is on its way.

Quake prediction, which has been called the "Holy Grail' of earth science,
began to look more and more like the hunt for the Loch Ness monster.

Earthquake prediction fell to "the nut cases,' said Christopher Scholz, a
professor at Columbia University's Lamont-Doherty Earth Observatory.
"Keilis-Borok is a little bit of an oddball, too, but he's got (a)
scientific career, so what can you do?'

Some in the scientific community say that eventually a method of prediction
will be found, but are not convinced that Keilis-Borok is on the right path.

There are doubters who say that the answer is not just elusive, it's
impossible; the mysteries buried deep in the earth will remain there.

Still others say that predictions will only be getting stronger with better
computers and technology like GPS, which can pinpoint land movements.

"They stopped calling it predictions and they're calling it forecasting,'
said Mark Benthien, spokesman for USC's Southern California Earthquake
Center. "They're working on the physics of earthquakes.'

Meanwhile, government officials fret about panic and try to get the public
more focused on dealing with the inevitability of earthquakes than trying to
predict them.

Because Keilis-Borok's method is based on patterns more than physics, some
scientists complain that the science can't be checked or argued, leaving the
world to sit, speculate and see if the earthquakes come.

Even Keilis-Borok seemed to cringe that the prediction was made public while
his model is still being tested.

"Yes, I'm honor-bound to keep serious predictions only among professionals
because predictions, published among the common public, can cause a rash of
anxiety,' Keilis-Borok said. "So far, it seems to be not caused much harm.
... In the future, we'll communicate our predictions to professionals only
and disaster-management agencies.'

UCLA's Vidale said, "He intended to keep the prediction quiet. When we put
it in a press release, I guess that ended that.

"He would just like to talk to government officials, but it doesn't work
that way in the U.S.'

Keilis-Borok practiced seismology for years in Russia before coming to UCLA.
He had developed algorithms and prediction models based on patterns of
earthquakes, but abandoned them when they proved too inaccurate.

Scientists are constantly making earthquake predictions. The USGS is asked
to make evaluations of earthquake risks, and comes out with statements like:
"There's a 62 percent probability of at least one magnitude 6.7 or greater
quake, capable of causing widespread damage, striking the San Francisco Bay
region before 2032.'

Yet even vague predictions like that, set within a timeline based on a
30-year mortgage, are mostly guesswork, using the past to try to divine the
future.

Getting specific

What sets Keilis-Borok's prediction apart is how specific he is about the
magnitude and the short time frame, which will make it easier to judge if
he's right.

Forget predicting an earthquake in nine years, as the scientists thought
they could do in Parkfield. Keilis-Borok and his team predicted an
earthquake for a nine-month window.

That was in December. Yet despite the fact that as the months creep closer,
his chances of success to go down, the real value may be in proving that
earthquake prediction has a chance.

"Earthquake prediction ... has been considered impossible by many
scientists,' Keilis-Borok said in a statement when UCLA issued its press
release in January. "It is not.'

He said he can foresee major quakes by tracking minor temblors and
historical patterns in seismic hot spots that could indicate more violent
shaking is on the way.

That method tracks smaller earthquakes of magnitude 2.9 or greater and plots
them on a map. Then, within a certain area, he looks to see if another
earthquake of significant size happens in a period of time, within a certain
area. Then he looks for another, then another, until there's a chain of
earthquakes stretching more than 175 kilometers.

The pattern is run through a mathematical formula, using chaos theory, and
if a chain qualifies by having a high enough score, a prediction is issued
for the 9-month interval immediately following the last event of the chain.