Earth is a precarious place, but few of us take much action in light of that knowledge. One concrete step was taken in 1999 by researchers who created the first numerical scale for cosmic impact eventsa "Richter scale" for meteorite strikes.
Cosmic Impacts, the Once and Future Threat
You can talk about global warming or giant volcanoes all you want, but the greatest natural threat to the human species is impacts from outer space. Only recently, once the dinosaur-killing asteroid of 65 million years ago was accepted as fact, have geologists taken impacts seriously. And once impacts are permitted in geologic history, they must be considered a possibility for the geologic present.
There is tantalizing evidence from various ancient sources that whole civilizations have been destroyed by comets before the invention of writing. And if you can entertain the idea that the legend of Noah's Flood is the long-preserved story of a real geologic event, then you have to admit that the other stories, including the legend of Atlantis and the myth of Phaethon, are suggestive too.
And the spectacle of comet Shoemaker-Levy smashing into the planet Jupiter in 1994 got a lot of people thinking. It seems prudent for a civilization like ours, which is starting to think about better ways of living with disasters for the long term, to get ready for that sort of thing. An estimated 2,000 near-Earth objects are orbiting out there, not all of which have been found, that could collide with Earth and wipe us out. The U.S. Congress funded programs to discover the largest of these, and in due course all the rest should be observed and characterized. Several ongoing projects around the world are following up, Spacewatch being the one with the catchiest name. NASA's Near-Earth Objects program is a good home page for this effort.
The Torino Scale of Possible Cosmic Catastrophes
In 1995 Prof. Richard Binzel of the Massachusetts Institute of Technology drew up a simple scale, from zero to 10, to indicate how threatening one of those objects might be. A comet rated zero is good for a show. An object rated 10, though, is one that is certain to strike us and cause a global-scale catastrophe.
Binzel presented his scale at a scientific meeting, discussion proceeded, and in June 1999 a gathering of specialists in Torino, Italy, ratified the final version of Binzel's scale. It's named the Torino Scale.
The scale has been used a few times, and the public response was such that Binzel felt the need to reword it in 2004, emphasizing that events "merit attention" by astronomers, not concern by anyone else (and that includes the credulous press). That version is now the official one that appears in the Earth Science Scales list.
The numbers in between signify greater and lesser chances of collision and different sizes of cataclysms. The first five levels, in fact, are nothing to worry about. The worst possibility is that public officials will need to keep informed in case something needs to be done within a decade. For the rest of us it's enough to, as they said in the movie, keep watching the skies.
As of early 2011, only two objects even rated a 1 on the Torino scale. That's the "Normal" level, used for "a routine discovery in which a pass near the Earth is predicted that poses no unusual level of danger." NASA's Near Earth Object (NEO) program lists them in its Sentry Risk Table page, along with hundreds of zero-level objects, some of them only 2 meters across.
What Scientists Use: The Palermo Scale
The Torino scale is really only useful for once-in-a-lifetime exceptions and for consciousness raising. Astronomers need something to help them sort out the thousands of ordinary objects they see and pick the most significant ones for regular routine monitoring. An index called the Palermo scale was devised in 2002 for this purpose.
The Palermo scale is an index that takes note of an object's size (large is bad), the odds of a collision (high is bad), and the time until that event (short is bad) and weighs these against the background risk of a comparable event. Because any one event has a much smaller risk than the background level of risk from all possible events, the index is typically a small fraction of 1. The scale is the logarithm of that fraction, so that a significant event whose risk is as great as the background level (index value = 1) scores a zero on the Palermo scale. An event with a risk 1/100 (1 percent) of background has an index of 2. For cases when a space object will approach Earth many times during the next century, the Palermo scale can also be cumulative, adding up the risk from all those encounters.
The NEO program lists hundreds of objects, with cumulative Palermo scores ranging from 1.12 to 11, meaning risks from roughly one-tenth to one-hundred-billionth of background. An event with a Palermo rank above +1 will typically also have a Torino rank of 1 or higher, but the two scales don't match in any other way. Fortunately, astronomers have an endless interest in these insignificant things in the night. That helps the rest of us sleep well.
PS: If the thought of cosmic collisions keeps you awake and you feel spurred to just get off the planet, out of their way, visit the Space Frontier Foundation. Unlike geologists, who generally are happy with robot-centered space science, the SFF pushes for human settlements by whatever means necessary. They quote Larry Niven to great effect: "The dinosaurs became extinct because they didn't have a space program."