The subject of the threat to our civilisation from asteroid impact is an old one in SF, having inspired various novels (of which Arthur Clarke's The Hammer of God, reviewed here 12 Sep 2008, is probably the most informative, albeit now a bit old) and a couple of modern Hollywood films (Deep Impact and Armageddon, both released in 1998). Only the other week, a careful analysis of all of the data relating to the great dinosaur die-off 65 million years ago confirmed beyond reasonable doubt that the major blow was struck by the asteroid, ten kilometres in diameter, which created the 180 kilometre wide, 900 metre deep, Chicxulub crater in Central America, with catastrophic consequences for the planetary environment.
Such a huge impact happens about once every 100 million years on average, and at present there is nothing that humanity could do to prevent it from happening again - or even to deal with a much smaller incoming asteroid which could still cause a major regional disaster. A collision with a 200 metre wide body takes place about once every 10,000 years. Several asteroids have been identified which will pass close enough to the Earth to create a small risk of collision, the most worrying being Apophis, with is 270 metres wide and has a one in 45,000 chance of hitting us in 2036. So it's worth devoting some thought to how we might protect ourselves in the future.
Current thinking on these issues was discussed in a couple of New Scientist articles by David Shiga last year: How to save the world from an asteroid impact (28/3/09) and No need to worry about asteroid tsunami disaster (18/4/09).
The author summarises three possible means of preventing or reducing the scale of the destruction: blasting it apart with nuclear weapons (which may mean that many smaller chunks hit the Earth with lesser but still serious consequences); forcing it off course by hitting it with another heavy object travelling at speed (technically difficult and risky); or nudging it more gently out of the way without breaking it apart, by detonating a nuclear device at a distance or pushing it with high-powered lasers.
Work has taken place at the Lawrence Livermore National Laboratory in California to evaluate the possibility of pushing an asteroid aside with a nuclear blast. This is complicated by the fact that most small asteroids are believed to be loose agglomerations of material rather than solid objects. They were able to demonstrate that it would be possible to change the velocity of a one kilometre wide asteroid sufficiently to miss the Earth by detonating a 100 kt bomb some 250 metres behind it - provided that this was done thirty years before the impact. An alternative being considered is to detonate a much smaller (less than 1 kilotonne) weapon just below the surface. These seem to be the best techniques available to us for the time being, provided that we have enough notice: the shorter the warning time, the more difficult it becomes.
If we have only a short warning time we would need to revert to Plan A and attempt to fragment the asteroid with a massive nuclear device detonated under the surface. If done three years before impact, only a very small fraction of the resulting debris cloud would hit the Earth - but this may not work if the asteroid is one solid chunk of rock.
A more technically difficult but potentially less risky proposal is to use lasers. These would be mounted in spacecraft, several of which would be sent out to rendezvous with the asteroid. They would focus their beams on one point on the surface, creating a plume of vapourised rock which over a period of months or years would act as a side-thruster, nudging the asteroid onto a safer course.
But suppose all these attempt fail, what happens if a large asteroid strikes? Given that the Earth's surface is 70% water, there's a good chance that it would land in the sea and cause a massive tsunami, but there is some dispute as to exactly how big this might be. If a 200 metre wide body struck the deep ocean it would displace billions of tons of water, creating waves hundreds of metres high. However, unlike the earthquake-induced tsunamis with which we have become familiar, these huge, steep waves are likely to collapse and rapidly reduce in height with distance, declining to perhaps 10 metres at a distance of 1,000 km. Given the way in which oceanic waves pile up and increase in height as they approach the land that still sounds very serious, but the shorter wavelength of impact waves (less than two minutes rather than eight minutes for a typical tsunami) means that they would be unlikely to penetrate so far inland. Such an impact, while still devastating to land areas nearby, may therefore have a more limited effect than earlier studies suggested.
This may be slightly reassuring, but having seen the consequences of much smaller waves from the Indian Ocean tsunami of 26 December 2004, it's still worth putting significant scientific effort and funding into methods to avert such a disaster.