Yesterday, a magnitude 6.3 earthquake struck L'Aquila, Italy, killing more than 150 people, injuring some 1,000, and leaving thousands of people homeless. Soon after the deadly temblor hit, news outlets including Time magazine, Reuters, and The New York Times reported Italian authorities had previously removed from the Internet a warning that a big quake was imminent. The prediction had been posted weeks earlier by a techician at the Gran Sasso National Laboratory in Abruzzi, Italy.*

The technician, Giampaolo Giuliani, who could not be reached for comment, had predicted that a massive earthquake would strike based on measurements of radon emissions. Skeptical Italian officials--believing that he was sowing unwarranted panic—publicly refuted the warnings and admonished Giuliani in court.

Did authorities err in contradicting Giuliani's prediction? Could lives have been saved had they heeded the warning?

"I am skeptical of the claim," says Shawn Larsen, a geophysicist at Lawrence Livermore National Laboratory in California. "Radon has been claimed to be a precursor of earthquakes for some time, since the late 70s. However, there has been no concrete evidence that it is indeed a predictor of earthquakes."

According to John Rundle, director of the California Institute for Hazard Research, a joint program between different University of California (U.C.) schools, the frequent release of that gas results in far too many false alarms to make the system trustworthy. Radon has been associated with seismic events, and earthquakes can cause the release of radon and other ground gases. The problem is, he says, many phenomena other than large earthquakes also result in radon release, including rainfall and atmospheric pressure changes.

"It has too many false positives to be useful," says Rundle, a professor of physics, geology and engineering at U.C. Davis. ""We actually do forecasting but we do that using probabilities. We can't say that an earthquake is going to happen at this point in time and space."

Rundle is an expert in forecasting quake probability over a range of time, and even published a map in 2006 identifying the spot of yesterday's earthquake as a likely target for a magnitude 7.0 or greater earthquake. That forecast identifies a 10-year-span rather than a single day or time when an earthquake might occur.

In fact, no current technique can accurately predict the timing of an earthquake more narrowly, Rundle says. For instance, in 1983 the U.S. Geological Survey (USGS) noticed that magnitude 6.0 earthquakes struck Parkfield, Calif., every 30 years or so. After observing the fault, the USGS predicted that another earthquake would strike between 1988 and 1993. There was another earthquake at Parkfield—but not until 2004.

The only technique that shows the promise of one day predicting earthquakes more precisely, according to both Rundle and Larsen, involves measuring electromagnetic (EM) radiation released by quartz veins compressing under the stress that precedes earthquakes.  But they note that technique remains years away from practical use, because scientists still have problems separating the electromagnetic (EM) radiation released by the quartz from EM radiation generated by other nearby sources.

The bottom line, according to Larsen and Rundle: If anything, Giuliani made a lucky guess.

"The ability to predict an earthquake would be tantamount to saying what the weather is going to be like in this city next November," Larsen says. "Maybe someone will find something in the future, but so far, that has not happened."

*Correction (4/8/09): Giuliani was originally identified as a seismologist in the headline and throughout the story.