Asteroids don't have to strike Earth to wreak havoc. Building-size rocks can blow up overhead, unleashing shock waves that pummel the ground. However, don't lose sleep over this threat: A new study based on satellite images of small asteroid bursts in the atmosphere suggests that the most harmful explosions happen only about once a millennium--about four times less often than previously thought.
The most recent violent airburst struck in 1908 near the Tunguska River in Siberia. It was a weak asteroid, perhaps 50 meters wide, that never hit the ground. But the shock from its detonation flattened trees over hundreds of square kilometers. The blast's energy packed as much punch as 10 megatons of TNT; a similar event over an urban area today could kill millions of people. Earlier estimates from counts of asteroids and craters imply that a Tunguska-size airburst should happen somewhere on Earth every 200 to 300 years--a rare but not vanishingly small risk.
Images from government satellites have helped researchers refine that calculation. Meteor scientist Peter Brown of the University of Western Ontario in London, Ontario, and his colleagues analyzed previously classified data from U.S. Department of Defense and Department of Energy satellites. These sentinels watch for nuclear tests, but they also catch hundreds of bolides--small asteroids that explode into fireballs.
Brown's team studied the light from 300 such airbursts in the last 8.5 years. The scientists converted the optical signals into an energy for each blast, calibrated by other data on sound waves and speeds for a dozen of the events. In the 21 November Nature, they report that a bolide equal to 5 kilotons of TNT blows up in the atmosphere about once per year. That's a spectacular show, but it's not big enough for a shock wave to hit the ground. As for the bigger, badder bolides, the team extrapolated between the energy ranges of the 300 airbursts and those of the far rarer impacts. Damaging 10-megaton explosions strike Earth's atmosphere just once per 1000 years, on average, they deduced.
That extrapolation is a convincing use of the best existing data, says planetary astronomer Robert Jedicke of the Spacewatch program at the University of Arizona, Tucson. "But variations in the impact rates are inevitable," he adds, because disrupted comets or collisions among asteroids can create more intense streams of objects. Brown agrees, noting that 8.5 years of data isn't nearly long enough to account for such episodes.