It looks inconsequential enough, the faint little spot moving leisurely across the sky. The mountain-top telescope that just detected it is taking it very seriously, though. It is an asteroid, one never seen before. Rapid-survey telescopes discover thousands of asteroids every year, but there's something very particular about this one. The telescope's software decides to wake several human astronomers with a text message they hoped they would never receive. The asteroid is on a collision course with Earth. It is the size of a skyscraper and it's big enough to raze a city to the ground. Oh, and it will be here in three days.

Far-fetched it might seem, but this scenario is all too plausible. Certainly it is realistic enough that the US air force recently brought together scientists, military officers and emergency-response officials for the first time to assess the nation's ability to cope, should it come to pass.

They were asked to imagine how their respective organisations would respond to a mythical asteroid called Innoculatus striking the Earth after just three days' warning. The asteroid consisted of two parts: a pile of rubble 270 metres across which was destined to splash down in the Atlantic Ocean off the west coast of Africa, and a 50-metre-wide rock heading, in true Hollywood style, directly for Washington DC.

The exercise, which took place in December 2008, exposed the chilling dangers asteroids pose. Not only is there no plan for what to do when an asteroid hits, but our early-warning systems - which could make the difference between life and death - are woefully inadequate. The meeting provided just the wake-up call organiser Peter Garreston had hoped to create. He has long been concerned about the threat of an impact. "As a taxpayer, I would appreciate my air force taking a look at something that would be certainly as bad as nuclear terrorism in a city, and potentially a civilisation-ending event," he says.

The latest space rock to put the frighteners on us was 2008 TC3. This car-sized object exploded in the atmosphere over Sudan in October last year. A telescope first spotted it just 20 hours before impact - at a distance of 500,000 kilometres - and astronomers say we were lucky to get any warning at all.

Thankfully, 2008 TC3 was far too small to do any damage on the ground, but we are nearly as blind to objects big enough to do serious harm. We have barely begun to track down the millions of skyscraper-sized asteroids zipping around Earth's neighbourhood, any one of which could unleash as much destructive power as a nuclear bomb on impact.

Asteroid impacts are not as rare as you might think. It is widely accepted that an asteroid or comet 30 to 50 metres across exploded over Tunguska in Siberia in 1908, flattening trees for dozens of kilometres all around. The chance of a similar impact is about 1 in 500 each year (Nature, vol 453, p 1178). Put another way, that's a 10 per cent chance of an impact in the next 50 years (see Should we panic?).

"Fifty-metre asteroids scare me to death," says Timothy Spahr, director of the Minor Planet Center in Cambridge, Massachusetts. "I could easily see a 50-metre object hitting in three days causing absolute pandemonium."

During the US air force planning exercise, the participating scientists explained that with so little warning there would be no hope of preventing an impact. Even Innoculatus's smaller 50-metre asteroid would weigh hundreds of thousands of tonnes, requiring an enormous push to change its trajectory appreciably - so much so that detonating a nuke near it in space would not provide a sufficient impulse so late in the game to cause a miss. To deflect an asteroid sufficiently, force would need to be applied years in advance (see Could we nuke it?).

In fact, it could make things worse by breaking the asteroid into pieces, some of which could be large enough to do damage, and even create a blizzard of meteors that would destroy satellites in Earth orbit.

Panic on the streets

Realistically, though, the nuclear option would not be on the table in the first place: the nuclear-tipped missiles sitting patiently in silos around the world are not designed to track and home in on an asteroid or even survive for more than a few minutes in space. Instead, we would simply have to brace ourselves for the impact.

The good news is that even a little warning makes a big difference, simply because it would allow us to predict the time and location of impact. In the case of 2008 TC3, just a few hours after the asteroid's discovery, NASA scientists completed calculations that predicted an atmospheric plunge over an unpopulated desert area of northern Sudan, with timing accurate to within a minute.

But participants in the planning exercise worried that if an asteroid posing an imminent threat to a populated area were discovered, and the situation were not handled properly, panic and lack of coordination could lead to chaos on the roads.

Spahr was not involved in the exercise, but shares those concerns. "With a three-day warning, you can walk away and be safe. But it scares me, given how poorly we've handled things of this nature in the past," he says, citing the failure to fully evacuate New Orleans ahead of hurricane Katrina in 2005. "I'm picturing people panicking and driving the wrong way on the freeway, screaming 'Oh my god, it's going to kill us!'"

To prevent panic and disorganised movement, it is crucial for authorities to develop an evacuation plan and communicate it to the public as soon as possible after discovery of the dangerous object, since such discoveries are posted automatically online and would cause a media firestorm.

Such measures should ensure the streets would be very quiet as an object such as Innoculatus plunges into the atmosphere and makes its final approach to Washington DC. The compression of the atmosphere in front of the asteroid and friction with the air would cause rapid heating. At lower altitudes, where the air is denser, the heating becomes so intense that the asteroid vaporises and explodes. For the Tunguska event, this happened at about 8 kilometres above ground.

Supersonic shock wave

If you were unfortunate enough to be looking up from directly below, the explosion would be brighter than the sun. The visible and infrared radiation would be strong enough to make anything flammable ignite, says Mark Boslough of Sandia National Laboratory in Livermore, California. "It's like being in a broiler oven," he says. Anyone directly exposed would quickly be very badly burned.

Even before the sound of the blast reaches you, your body would be smashed by a devastating supersonic shock wave as the explosion creates a bubble of high-pressure air that expands faster than the speed of sound. Planetary scientist Jay Melosh of Purdue University in New York once experienced a shock wave from an experiment that exploded 500 tonnes of TNT, a tiny blast in comparison with the blast from an asteroid. "I was standing on top of a hill about 1.5 kilometres away wearing earplugs," he recalls. Melosh says you would see the shockwave in the air due to the way it refracts light. "It's a shimmering bubble," he says. "It spreads out in complete silence until it reaches you, then you hear a double boom."

Melosh was at a safe distance, but at ground zero below an exploding asteroid, the shock wave would be powerful enough to knock down buildings. It would arrive about 30 seconds after the blazing hot flash of light, and could also knock any nearby planes out of the sky, Boslough says. Any surviving buildings would be pummelled by raging winds blowing faster than any hurricane can muster.

Of course, two-thirds of Earth's surface is ocean. While our atmosphere is likely to protect us from asteroids smaller than 100 metres across, anything larger hitting the ocean - including chunks of Innoculatus's rubble pile - would cause a giant splash that could smash coastal buildings with high-speed volleys of water. The tremendous damage and loss of life that would ensue if multiple cities around an ocean basin were flooded led NASA scientists in 2003 to rate ocean impacts by asteroids as far more dangerous than those on or over land.

Recent computer simulations offer some hope, though. They suggest that the monster waves generated by ocean impacts would typically break far from shore, dissipating most of their energy before they could reach cities - unless the impact was very close to the coast, of course. Another ray of hope is that 100-metre asteroids hit Earth only about one-tenth as often as 30-metre objects.

Lasting just one day, the 2008 US air force exercise could barely scratch the surface of the incoming-asteroid problem. Not surprisingly, it discovered that should the nightmare come true, there is no plan for how to coordinate the activities of NASA, emergency planners, the US military and other parts of government. Further planning exercises are needed: the time saved through early preparation will be crucial if an evacuation is ever required at short notice.

Our chance of having any prior warning at all for an approaching 30-metre asteroid is no better than 25 to 35 per cent with existing sky surveillance, calculates astronomer Alan Harris of the Space Science Institute in Boulder, Colorado. The sun washes out half of the sky with daylight, blinding us to 50 per cent of threatening objects. Even glare from the moon can hide unwelcome incoming guests.

What's more, two of the world's three leading asteroid surveys are based in Arizona, including the Catalina Sky Survey, which discovered 2008 TC3. The region tends to cloud over between July and September. "Shift 2008 TC3 back to July and forget it. It wouldn't have been seen," says Spahr.

Now picture this ugly scenario, which worried some participants in the air force exercise: an asteroid flies out of nowhere and explodes over a sensitive nuclear-armed region, like South-East Asia or the Middle East. There's a reasonable chance that such an airburst could be misinterpreted as a nuclear attack. Both produce a bright flash, a blast wave and raging winds.

Such concerns were one reason why, when NASA found 2008 TC3 in its sights, it not only issued a press release but also alerted the US State Department, military commanders, and White House officials, says Lindley Johnson at NASA headquarters, who oversees the agency's work on near-Earth objects. "If it had been going down in the middle of the Pacific somewhere, we probably would not have worried too much more about it, but since it was [going to be] on land and near the Middle East, we did our full alerting," he says.

There is one major way to improve our prospects - point more eyes at the skies. The European Space Agency wants to get into the monitoring game and may set its telescopes at the European Southern Observatory in Chile on the problem. This could fill a gap in the NASA-funded surveys, which are limited to watching the skies of the northern hemisphere, says Richard Crowther of the UK's Science and Technology Facilities Council, who is a consultant for ESA and heads a United Nations working group on near-Earth objects.

Be prepared

"Up to now, the US has taken the majority of the responsibility for dealing with this issue and I think it's time for other states to take on a more equitable share of that," he says.

Help will also come from two new US observatories designed to survey the entire sky visible from their locations every few days. The Panoramic Survey Telescope and Rapid Response System (Pan-STARRS), will consist of four 1.8-metre telescopes, the first of which is already up and running in Hawaii. Plans are afoot to construct the 8.4-metre Large Synoptic Survey Telescope in Chile by 2015, though the project is still raising funds. These will improve the chances of an early detection and potentially extend warning times for 30-metre objects to more than a month. But even so, every ground-based lookout suffers from interference from the sun and moon.

A dedicated space telescope would fix this problem, but such a mission could cost more than a billion dollars. "We're talking about investing in an insurance policy," says Irwin Shapiro of the Harvard-Smithsonian Center for Astrophysics in Cambridge, Massachusetts.

Shapiro is leading a US National Research Council panel that by year's end will recommend a strategy to better address the threat from near-Earth objects. That study, along with the air force's report on its asteroid impact exercise, is intended to help the White House develop an official policy on the near-Earth object hazard by October 2010, which Congress has requested.

While asteroid impacts are much rarer than hurricanes and earthquakes, they have the potential to do much greater damage, Johnson warns: "It's not something I think there needs to be billions of dollars per year spent on, but it does warrant some priority in the list of things that we ought to be worried about." The cash would at least give us a better idea of when the next asteroid might strike. "From what we know today," he says, "it could be next week."