One of the main weaknesses of the environmental movement has been its unfortunate predilection for using doom-laden language and catastrophic superlatives to describe problems that are serious but not immediately disastrous. But one calamity that truly deserves such a description is almost never talked about. There are tens of millions of asteroids in the solar system, and several thousand move in orbits that take them close to Earth. Sooner or later, one of them is going to hit it.

©NASA

Several have done so in the past. Earth's active surface and enthusiastic weather conspire to scrub the tell-tale impact craters from the planet's surface relatively quickly, but the pockmarked surface of the moon - where such scars endure for much longer - testifies to the amount of rubble floating in the solar system. Earth's thick atmosphere makes it better protected than the moon: asteroids smaller than about 35 metres (115 feet) across will burn up before hitting its surface. Nevertheless, plenty of craters exist. The Earth Impact Database in Canada lists more than 170.

Fortunately, such impacts are relatively rare, at least on human timescales. Statisticians calculate that the risk to lives and property posed by meteorite strikes are roughly comparable with those posed by earthquakes.

Although the chance of an impact may be small in any given year, the consequences could be enormous. The effect of an impact depends on an object's size and speed. A meteorite a few metres wide could level a city. The largest (a kilometre or more in diameter) could wreak ecological havoc across the entire globe. David Morrison, a NASA scientist, argued at a recent conference that a large meteorite strike is the only known disaster (except perhaps global nuclear war) that could put civilisation at risk.

Examples give a more visceral illustration than statistics. The Chicxulub crater, buried beneath modern Mexico, is 65m years old and 180km (112 miles) across. Some think that the ten-kilometre meteorite that created it threw so much dust into the atmosphere that it blotted out the sun and led to the extinction of the dinosaurs. In 1908 a comparatively tiny piece of space-borne rock, 30-50 metres across, exploded above Tunguska, a remote part of Siberia. The blast - hundreds of times more powerful than the atom bomb dropped on Hiroshima 37 years later - felled 80m trees over 2,150 square kilometres. Only blind luck ensured that it took place in a relatively unpopulated part of the world. Astronomers are currently trying to work out whether a 270-metre asteroid named 99942 Apophis will hit Earth in 2036 (probably not, but it would be nice to be sure).

Happily for humanity, technology has advanced to the point where it is possible, in principle, to avoid such a collision. In 1998 NASA agreed to try to find and catalogue, by 2008, 90% of those asteroids bigger than 1km in diameter that might pose a threat to Earth. Any deemed dangerous would have to be pushed into a safer orbit. One obvious way to do this is with nuclear weapons, a method that has the pleasing symmetry of using one potential catastrophe to avert another. But scientists counsel caution. A nuclear blast could simply split one large asteroid into several smaller ones, some of which could still be on a collision course.

Other plans have been suggested. One is to use a high-speed spaceship simply to ram the asteroid out of the way; another is to land a craft on the rock's surface and use its engines to manoeuvre the asteroid to safety. A subtler method is to park a spaceship nearby and use its tiny gravity to pull the asteroid gradually off course. For now, all such suggestions are theoretical, although the European Space Agency is planning a mission, named Don Quijote, to test the ramming tactic in 2011.

These schemes offer consolation, but any effort to deflect an asteroid requires plenty of advance warning, and that may not always be available. NASA has so far catalogued only the very largest, "civilisation-killing" asteroids. Plenty of smaller ones remain undiscovered, and they could inflict considerable damage. In 2002 a mid-sized asteroid (50-120 metres across) missed Earth by 121,000km - one-third of the distance to the moon. Astronomers discovered it three days after the event. Comets, which originate from the outer reaches of the solar system, are faster moving and harder to track than asteroids, but carry just as much potential for catastrophe.

But perhaps the biggest problem is humanity's indifference. Currently only America is spending any money on detection, and even there, politicians have other priorities. Much of the work is done by Cornell University's Arecibo radar in Puerto Rico, which is facing federal funding cuts. The telescope costs roughly $1m a year to operate. As an insurance policy for civilisation, the price looks cheap.