Experts criticize quake-forecasting claims

A California firm is forecasting a magnitude 2.0 to 3.5 earthquake for New Madrid, Mo., between today and Sunday -- part of a marketing attempt to gain subscribers for its Web site, geoForecaster.com.

In response to the mass e-mail and telemarketing in connection with the forecast, government scientists and emergency management officials are making it clear they've seen no evidence geoForecaster Inc. can accurately forecast an earthquake.

The New Madrid Fault -- which runs across parts of Missouri, Kentucky, Tennessee, Illinois and Arkansas -- experiences more than 200 small earthquakes annually.

"Most events in the New Madrid area are smaller in magnitude," said Michael Kozuch, Ph.D., president and co-founder of geoForecaster. "We chose to publish the sample forecast and issue a notice on this event because the higher end of the expected range in magnitude is outside the norm of what we've been seeing in that area for some time."

However, New Madrid city manager Ferguson Hunter said no one from geoForecaster warned him his town could have an earthquake, and he finds the commercial aspect of the issue unsettling.

"It seems kind of odd to have to pay for that service," he said.

Kozuch said money is necessary to "keep funding the science."

GeoForecaster says it does not predict exact times and locations of earthquakes but rather forecasts a short time period and magnitude range. Kozuch refuses to say how the company forecasts quakes, pointing to proprietary reasons.

GeoForecaster has no endorsements from any of the nation's science or government agencies that deal with seismic activity.

Bruce Kinerney, spokesman for the USGS, said his agency isn't aware of any reliable method to predict earthquakes. The USGS monitors earthquakes in the nation and would be responsible for issuing warnings if a reliable prediction technique became available.

"We would be prepared to examine the company's research and methods if they were published in the open scientific literature," Kinerney said.

Seismologist Paul Bodin of the Center for Earthquake Research and Information at the University of Memphis in Memphis, Tenn., said he's skeptical, too. He went to graduate school with two of the creators and has reviewed the site.

"It is about as far removed from science as an endeavor can get without being labeled fantasy," Bodin wrote in an e-mail.

Nicholas Tibbs, a professor of geosciences at Southeast Missouri State University in Cape Girardeau, was not aware of geoForecaster's purported ability to forecast seismic activity.

"You can't accurately predict earthquakes," he said. "The technology to do that doesn't exist. It sounds like a scam to me."

No published research

The company claims to have accurately forecasted 14,484 magnitude 3 earthquakes since Jan. 1, 2002. But no published comparative results are available, and no independent agency audited these previous forecasts, either.

"That was purely an internal testing," Kozuch said. He said geoForecaster wants to develop an easy-to-understand system for users before making the comparisons available.

The Web site went online earlier this week. Kozuch would not reveal the number of subscribers, though the company is targeting the media, utilities and government and disaster-relief agencies.

Fees may be paid monthly or annually and vary depending on how large of an area the subscriber wants.

The New Madrid Fault gained international notoriety in 1990 when climatologist Iben Browning predicted a devastating 7.5 to 8.0 magnitude earthquake could occur on Dec. 3 of that year. That big quake never happened, but there is a 7 to 10 percent chance of one within the next 50 years, according to the USGS.

mwells@semissourian.com

335-6611, extension 160