We have extensive evidence that Earth has already been hit by asteroids many times throughout history-the most famous (or infamous) example is probably the asteroid or comet that created the Chicxulub crater in the Gulf of Mexico and may have contributed to the extinction of the dinosaurs at the end of the Cretaceous Period 65 million years ago. A more recent but less devastating example, called the Tunguska event, occurred in 1908, when a meteor or comet exploded over the wilderness of Siberia, damaging farmland and leveling trees for miles around. Because most of the earth is covered by oceans, there may also be many small impacts that go unnoticed.
There are thousands of small bodies that we call asteroids or meteoroids in orbit around the sun. Many of these objects are called near-Earth asteroids (or NEAs) because they have orbits that repeatedly bring them close to, or intersect with, Earth's orbit.
Although the odds of any one particular asteroid ever impacting Earth are quite low, it is still likely that one day our planet will be hit by another asteroid. At the current rate of impacts, we would expect about one large asteroid to impact Earth every 100 million years or so. For that reason several programs, such as the Lincoln Near-Earth Asteroid Research (LINEAR) project at the Massachusetts Institute of Technology, have been undertaken around the world to discover and monitor potentially Earth-threatening asteroids.
When a new asteroid is discovered, astronomers analyze it to determine whether its orbit around the sun could bring it close to the Earth. They take successive images of the asteroid over the course of days after its discovery in order to predict its probable orbital path for the near future. The predicted orbit is then compared to the orbit and position of Earth to check for any times when they might pass close to each other.
Although scientists can calculate a most-likely orbit from these early observations, each single observation of the asteroid's position contains some uncertainty. Most asteroids are small objects, a few meters to a few tens of meters across, and even the resolving power of a large telescope cannot determine their positions exactly. The uncertainties in an asteroid's position lead to uncertainties in how well we can determine its speed and direction of travel. As a result, a large number of possible orbits for an asteroid can be predicted within these windows of uncertainty.
Careful computer simulations are used to calculate the future orbital path of the asteroid, with randomly chosen initial positions and velocities that fall within the margin of error of the telescopic observations to date. A large number of these simulations are generated for each asteroid. The probability that any particular one will actually hit Earth is given by the fraction of the extrapolated paths that leads to an impact. For example, if one million different possible orbits are calculated, and one of those leads to an impact, then we say that the odds of the asteroid hitting our world are one million to one.
The uncertainties in an asteroid's orbit are greatest in the hours just after its discovery, and thus the calculated probability of an impact also tends to be the highest at these times. As we monitor an asteroid over the course of the weeks or months that follow, its orbit becomes more and more certain, and we become more knowledgeable about its position at a given date in the future. We can then rule out many possible paths it may take. In most cases, monitoring the asteroid over a few weeks quickly leads to an impact probability of very nearly zero.
Answer originally published October 18, 2004.