How to save the world from an asteroid impact ?

Year 2036. A large asteroid is on a collision course with Earth. Unless it is stopped, it will crash into the Pacific Ocean, creating a devastating tsunami. What should we do?

We could blast the asteroid with a nuclear bomb, but that would risk shattering it into smaller pieces that could still threaten Earth. Or maybe we should try to force it off course by slamming into it with a heavy object - an unproven and therefore risky technique. Now there may be a third option: gently nudging the asteroid away from Earth without breaking it apart, either by exploding a nuclear device at a distance or zapping it with high-powered lasers.

Astronomers have found thousands of asteroids that pass near Earth's orbit, and a few of these are on trajectories that give them a small chance of hitting Earth. The most worrying is a 270-metre-wide asteroid named Apophis, which has a 1 in 45,000 chance of hitting us in 2036.

To investigate the best way to deflect this and other asteroids onto a harmless path, a team led by David Dearborn of the Lawrence Livermore National Laboratory in California has modelled the impact of a nuclear explosion on an object's trajectory. Their virtual asteroid was 1 kilometre in diameter and made of rocky rubble loosely bound together by gravity, which is considered by many planetary scientists to be the most likely composition for small asteroids.

Thirty years before the asteroid was set to collide with Earth, a nuclear blast, equivalent to 100 kilotonnes of TNT, was set off 250 metres behind it. The nudge from the explosion increased its velocity by 6.5 millimetres per second, a slight change but enough for it to miss us.

The technique also reduced the risk of a break-up - just 1 per cent of the asteroid's material was dislodged by the blast, and of that only about 1 part in a million remained on a collision course with Earth. Dearborn adds that the technology for this method is already established, unlike for the use of a heavy object to shove the asteroid onto a different path - the "kinetic impactor" strategy. "Should an emergency arise, we should know that [the technology] is available, and we should have some idea of how to properly use it," he says.

He has now begun simulating the effect of nudging an asteroid with a smaller nuclear explosion - less than 1 kilotonne - 1 metre below its surface. This would reduce the device's weight, making it easier and quicker to launch. He will discuss the work next month at the 1st IAA Planetary Defense Conference in Granada, Spain.

A less established and gentler approach would be to nudge the asteroid away from Earth using lasers. In this theory, being investigated by Massimiliano Vasile of the University of Glasgow in the UK and colleagues with funding from the European Space Agency, a fleet of eight or more spacecraft, each carrying a laser, would be sent to rendezvous with the asteroid. Hovering a few kilometres away, each craft would unfurl a 20-metre-wide mirror made of a flexible material such as Mylar. The mirror would focus the sun's rays onto the spacecraft's solar panels, powering the laser.

All eight lasers would then be simultaneously fired at a single spot on the asteroid's surface, vaporising that region and creating a plume of gas that should provide enough thrust to push the asteroid off course (see diagram). This relatively gentle nudging, over a period of months or years, would not break the asteroid up into any smaller pieces, the team say.

Vasile, who will also be presenting his idea at the conference, touts the flexibility and reliability of the approach. "You have a formation of satellites and if one breaks you have the others [for back-up]," he says. "And it's scalable, so if you have a bigger asteroid or you want to have a faster deflection then you add more spacecraft."

Whichever option is ultimately chosen, reliability will be essential for a task as critical as asteroid deflection, says Bill Ailor of the Aerospace Corporation in El Segundo, California, who is chairing next month's conference. "Launch vehicles fail at a rate of about 1 in 100, and new spacecraft might fail at the rate of 1 in 3, [which] has to be factored into the overall design of your deflection," he says. "We're in a sense betting the planet that we're going to make this work."