MIAMI—To avoid Armageddon, we may have to invoke Armageddon. You know, the Bruce Willis version.

That's the opinion of David Dearborn, anyway, who says we may need to tap our nuclear arsenal if a life-threatening asteroid suddenly comes into view. Dearborn, a research physicist at Lawrence Livermore National Laboratory, laid out the nuclear case in a talk here Tuesday at the semiannual meeting of the American Astronomical Society.

Dearborn's research on nuking asteroids is a fairly natural outgrowth of his other work, which has involved weapons development and testing, as well as three-dimensional modeling of astrophysical processes. He has run numerical simulations of how a nuclear detonation either near or on the surface of a threatening near-Earth object could divert or fragment it, and has found that with a little bit of lead time the weapons could do the job rather well.

Unlike the asteroids themselves, whose makeup and physical properties are not always well understood, nukes have been well characterized through testing. "We know what the nuclear part is going to do," Dearborn said. Gentler diversion schemes such as tugging the object onto a safer path tend to run into problems, he argued, because asteroids are such a diverse class of objects, and some might not hold together if grabbed in some way. "What's their structure? Rock or rubble?" Dearborn asked rhetorically. "The answer is yes."

A directed-energy weapon could do the job, Dearborn said, but not in the state that the technology is in today. A laser such as that at the National Ignition Facility, a nuclear fusion experiment at Lawrence Livermore, could adjust a typical asteroid's course enough to avoid a collision—changing its velocity by about one centimeter per second—but doing so would take about 6,000 years. "I'm not saying that our children's children won't know a lot more than we do, and I certainly hope that they do," Dearborn said, but laser technology isn't up to the job just yet. A non-nuclear blast—or a simple ramming mission—could also work, but those approaches would require numerous launches to match the power of a single nuclear device.

So what would a nuclear warhead actually do to an asteroid? It depends on how much time humankind had to deal with the threat. With a decades-long heads-up, a velocity shift of only a fraction of a centimeter per second would be needed, Dearborn said. In that case a detonation near—but not on—an asteroid would give it a sufficient nudge while leaving the object mostly intact. With less time to address the inbound impactor, a direct detonation would be needed. The downside is that some of the fragments might still hit Earth, although they would at least be reduced in size compared to the original object. Dearborn has simulated a last-ditch effort to destroy a 270-meter asteroid similar to Apophis, which has a very small chance (about 1 in 233,000) of striking Earth in 2036. With a detonation even 15 days before impact—"about the closest you'd want to do this," Dearborn said—only a few percent of debris would remain on course to hit Earth. But all simulations make some assumptions about the composition of the asteroid—when Dearborn ran other models assuming an asteroid with a core the strength of granite, he found that even a nuclear blast could still leave a dangerous solid core. If humankind had 30 years to deal with the problem, then, the best thing to do would be to launch a characterization mission to the asteroid, Dearborn said.

Of course, nuclear weapons make many people uneasy in any context, and not everyone is on board with the nuclear approach to asteroid deflection. But Dearborn maintained that the radiation from a detonation in deep space, where the radiation environment is already intense, would have no effect on life on Earth. "You wouldn't even be able to measure the difference in radioactivity," he said.

With any luck, all of Dearborn's modeling will remain an academic pursuit that never gets put to the test. "You shouldn't feel tremendous worry," Dearborn said, noting that impactors capable of regional destruction only come around on an average timescale of tens of thousands of years. And whereas an asteroid of that class could flatten everything from Los Angeles to Sacramento, a global killer only comes around every million years or so, Dearborn said. Perhaps most reassuring of all, Congressionally mandated programs have already catalogued what appears to be the vast majority of the largest objects, none of which pose a significant risk in the foreseeable future. But plenty of people are working on mitigation schemes—far-fetched though some of them may sound—just in case the sci-fi scenario of Armageddon becomes reality.

Photo of Meteor Crater in Arizona: NASA