Vladimir Keilis-Borok
|
SOUTHERN CALIFORNIA WONDERS
Will quake forecaster go three for three?
by anne burke
ucla today staff
Late last year, Vladimir Keilis-Borok, a UCLA professor in residence
who is arguably the world’s leading earthquake forecaster,
issued a startling prediction: A temblor of magnitude 6.4 or greater
had a 50-50 chance of striking the Southern California desert in
the nine months leading up to Sept. 5.
Keilis-Borok leads an international team of experts who experiment
with earthquake prediction. With two successful forecasts to their
credit last year — the Dec. 22 temblor that killed two people
on the Central California coast and a larger quake Sept. 25 in Hokkaido,
Japan — Keilis-Borok and his team were not lightly dismissed
by the seismology world.
The state Office of Emergency Services convened a blue-ribbon
panel of experts to look into the prediction while the U.S. Geological
Survey issued special bulletins. The experts agreed that while the
team’s methodology had merit, it was too early to tell whether
it would prove useful.
At UCLA, meanwhile, plaintive inquiries appeared in Keilis-Borok’s
e-mail inbox: “Should I take my newborn twins out of Southern
California?” “Should I sell my house?” “Is
it safe to use LAX?”
“I felt sympathy for them. Some of them didn’t want
a response, they just wanted to share their feelings,” said
Keilis-Borok, 83, who talked from his office in the Geology Building.
With six weeks to go before the prediction window was to close,
Keilis-Borok’s chances of going three for three seemed vastly
diminished. But the Russian-born scientist was not discouraged.
He said that even an unfulfilled prediction will be useful as his
team works to improve the method.
“If we made no mistakes, it would mean the problem is easy,”
said the scientist, who has the air of someone who would rather
be working than chatting. “To test the method, we need to
study tens and tens of earthquakes. We study our failures so that
we can eliminate them in the future.”
Of course, he added self-deprecatingly, the first thing he may
do on the morning of Sept. 6 is slap himself on the side of the
head.
The team’s method, which Keilis-Borok calls “tail
wags the dog,” looks backward to make earthquake predictions.
The professor and his colleagues seek out long chains of small earthquakes.
They then look at the seismic history of the area during preceding
years. If they detect a certain seismic pattern, they sound a nine-month
quake alarm, passing along the information to disaster management
authorities.
If valid, the method would represent a major breakthrough in earthquake
prediction. In the past, quake forecasters have done no better than
predict seismic events years in advance. The new method would shrink
the window to months.
In 1986 Keilis-Borok’s team predicted a moderate earthquake
within five years in the region of the Loma Prieta temblor, which
struck in 1989. In 1992, the team warned of a 6.6 or higher quake
within 120 miles of Landers. That forecast missed the Northridge
temblor by 21 days. The latest prediction covers a 120,000-square-mile
area that takes in large portions of the Mojave Desert, the Coachella
and Imperial valleys, and eastern San Diego County.
At UCLA, Keilis-Borok has the respect of his colleagues, though
not all share his confidence in the “tail wags the dog”
method.
“Of course, the proof is in the pudding, and Professor Keilis-Borok’s
methods have now delivered several correct and impressive forecasts,”
added John Vidale, interim director of the Institute of Geophysics
and Planetary Physics and professor of earth and space sciences.
Perfecting the science of earthquake prediction is a consuming
passion for Keilis-Borok, who is a widower and often works 10- to
12-hour days. A mathematical geophysicist, Keilis-Borok was the
leading seismologist for decades in Russia before he joined UCLA’s
faculty in 1999.
Whether or not the ground shakes between now and Sept. 5, Keilis-Borok
said he will not give up. His hope is that scientists will one day
able to predict earthquakes weeks or even days in advance.
|
