The New York Times (pg. D3)  [Printer-friendly version]
April 14, 2009

EARTHQUAKES' MANY MYSTERIES STYMIE EFFORTS TO PREDICT THEM

By Kenneth Chang

Almost all earthquakes are small. A small segment of a fault, miles
underground, jerks a little, the rumble imperceptible at the surface.
But with a few quakes, the fault continues breaking, the ground jumps
several feet and the world shakes in cataclysm.

"How does a rupture go from an inch a year to 3,000 miles per hour in
a few seconds?" asked Ross S. Stein, a geophysicist at the United
States Geological Survey.

No one knows.

This gap in knowledge makes earthquake prediction a frustrating and
chancy exercise, and complicates the effort to calculate the risk that
a human construction like a water reservoir or a geothermal power
plant could inadvertently set off a deadly quake.

Last month, Giampaolo Giuliani, a technician who works on a neutrino
experiment at the Gran Sasso National Laboratory in Italy, issued an
urgent warning that a large earthquake was about to strike the Abruzzo
region. The prediction was based on measurements he had made of high
levels of radon gas, presumably released from rocks that were being
ground up by the stresses of an incipient quake.

On April 6, a magnitude 6.3 earthquake hit L'Aquila in central Italy,
killing almost 300 people. Mr. Giuliani claimed vindication for his
prediction, which had been discounted by officials.

But earthquake experts like Dr. Stein are skeptical. Scientists
studied radon as a possible earthquake warning signal as far back as
the 1970s, and while they found convincing cases of radon releases
before some earthquakes -- for example, levels of radon in groundwater
were 10 times normal before the earthquake that hit Kobe, Japan, in
1995 -- the correlations were not strong enough or clear enough for
useful predictions.

One instance of confusing radon signals occurred in 1979. Two
detectors in Southern California, 20 miles apart, measured unusually
high levels of radon beginning in the summer. The radon levels then
decreased in October, shortly before three earthquakes struck.

One earthquake, of magnitude 6.6, occurred 180 miles to the southeast,
and the two smaller ones, of magnitudes 4.1 and 4.2, were 40 miles
away. In addition, a radon detector close to one of the smaller quakes
did not observe high radon levels, although it did observe a radon
drop a few days earlier.

That left scientists puzzled about how they could construct a
prediction out of the rising and falling radon levels. Data on other
gases like carbon dioxide and on electromagnetic emissions that have
sometimes been detected before earthquakes are also confusing.

"You can't hang your hat on it unless it's a reliable precursor and
it happens before most earthquakes and it doesn't happen at other
times," said Susan Hough, a seismologist at the geological survey.

To complicate matters, Mr. Giuliani's prediction was off in time and
place. He had predicted that the quake would hit a week earlier in a
town 30 miles away. Had officials acted on his prediction, said
Richard M. Allen, a professor of geophysics at the University of
California, Berkeley, "you would have evacuated the wrong town and
you would have evacuated the wrong town at the wrong time."

While prediction remains elusive, scientists have learned that human
activity can set off an earthquake. In December 2006, a geothermal
energy project in Basel, Switzerland, started injecting water three
miles into the ground. Some tiny tremors were expected, but the water
was shut off when one of the quakes reached a still minor magnitude of
2.7. A few hours later, a larger quake, at magnitude 3.4, shook Basel,
causing minor damage to buildings.

A couple of months later, there were two more magnitude 3 earthquakes.
Researchers at the Swiss Federal Institute of Technology in Zurich
calculate that the area will experience a slightly greater number of
small earthquakes over the next 20 to 40 years as a result of the
brief geothermal project, which remains halted.

The worry is that one of these small earthquakes could cascade into a
big earthquake like the one that badly damaged Basel in 1356.
Conversely, the small earthquakes could instead be relieving stress
along a fault, reducing the likelihood of a larger quake.

"With the current knowledge, I can't really tell you," said Jochen
Woessner, one of the Swiss scientists.

Geologists do not know how the pieces of the Earth's crust that
usually squeeze together tightly with high friction slip past each
other smoothly during a large earthquake, as if sandpaper suddenly
changed to Teflon. "It looks like friction is more a complicated
beast than anyone would have imagined," Dr. Stein said.

A core dug up from the San Andreas fault in California revealed the
presence of talc, which could be acting as a lubricant during an
earthquake. But from one core, scientists cannot tell whether this is
typical of rocks around earthquake faults.

At a meeting of the Seismological Society of America last week in
Monterey, Calif., a lively debate continued about whether big
earthquakes are fundamentally different from small earthquakes or
whether a big earthquake is just a small earthquake that did not stop.
If big earthquakes are different, then it might be possible to detect
them in the first few seconds of seismic waves and send out a warning.
People would not have time to evacuate, but they might have enough
time before the heaviest shaking to move to a safer location in a
doorway or under a desk.

Reservoirs are also believed to induce some earthquakes. Most
seismologists believe that a magnitude 6.5 earthquake in India in 1967
that killed about 200 people was set off by the weight of water in a
reservoir that had been filled a few years earlier. A reservoir cannot
generate an earthquake by itself, but it can act as a trigger to
release accumulated tectonic stresses and hasten an earthquake by
years or centuries.

More controversial is the assertion by some scientists that a
magnitude 7.9 earthquake in Sichuan province in China last year that
killed about 80,000 people was set off by the 320 million tons of
water in a nearby reservoir.

Leonardo Seeber, a research scientist at the Lamont-Doherty Earth
Observatory at Columbia University, is not sure about the Sichuan
earthquake, but he believes that scientists and officials need to take
more account of the risk of induced earthquakes.

For example, Dr. Seeber wonders whether a swarm of magnitude 4
earthquakes a couple of weeks ago around the Salton Sea in Southern
California, close to one end of the San Andreas, might have been
caused in part by a nearby geothermal power plant.

Extraction of oil from the ground may have set off other earthquakes,
Dr. Seeber said. In the coming years, the proposed strategy to reduce
global warming by capturing carbon dioxide from power plants and
pumping it into the ground could create new earthquake risks.

So far, experiments in this kind of carbon sequestration have focused
on whether it will work to keep carbon dioxide out of the air for
centuries. But Dr. Seeber said this technology "has huge implications
for triggering earthquakes."