Over the past couple of years the U.S. space program has gone through a huge shake-up, leaving the nation’s goals in space unclear. I have a suggestion. NASA, working with other national space agencies and private organizations, should take on the job of ensuring that no destructive asteroid ever hits Earth on our watch. What project is more worthwhile in the long term or awe-inspiring in the short term than protecting humanity from ruin?

At first glance, asteroids may seem like a distant threat. But the hazard is well documented, and the consequences could not be more severe. The history of life on Earth has been shaped by asteroid impacts. One million of them wider than 40 meters in diameter orbit the sun in our vicinity, by some estimates. An asteroid of that size struck Earth over Siberia in 1908 and laid waste an area 150 times larger than the Hiroshima atomic bomb did. The odds of a repeat in this century are about 50 percent. On the larger end, asteroids greater than about one kilometer across would have global effects that threaten human civilization.

The first step in prevention is prediction. We must find, track and predict the future trajectory of those million near-Earth objects. Astronomers have already catalogued the orbits of most of the kilometer-scale objects they think are out there, and none are known that will hit Earth in the next 100 years. Yet the great majority of smaller ones, those big enough to destroy a country or unleash a tsunami that devastates coastal cities, remain untracked. This unfinished business should be tackled next.

Asteroids are warmer than the background sky and therefore stand out in the infrared. Telescopes have blind spots, however: they cannot look in the direction of the sun, which limits the effectiveness of telescopes stationed on or near Earth. The National Research Council recommended in 2009 that NASA place an infrared survey spacecraft in a Venus-like orbit around the sun. As it looked outward, away from the sun, the observatory would spot asteroids that go unseen from Earth. Once completed, such a survey would remain valid for about a century—the timescale on which the measured orbits begin to change because of gravitational interactions with planets—before we would have to do it again. The cost of such a mission would be several hundred million dollars—expensive, to be sure, but a bargain compared with NASA’s current budget, let alone the damage of an asteroid strike.

Should astronomers find an asteroid on a collision course, our task would be to reach out and alter its orbit to prevent that impact. If we find the asteroid early enough (decades ahead of its projected impact), several existing technologies might work: tow it, ram it, nuke it or employ some combination. (My colleagues and I used to advocate pushing on the asteroid with a rocket [see “The Asteroid Tugboat,” by Russell L. Schweickart, Edward T. Lu, Piet Hut and Clark R. Chapman; Scientific American, November 2003], but recent results on asteroid properties and orbits have made us reconsider.)

Yet no one is really sure whether these options would actually work. Surely the time to test them is before they are needed for real. NASA and other organizations should build and try out a system to deflect a nonthreatening asteroid in a controllable way. Given that astronomers have not even begun a complete asteroid survey, there is a real risk they will find an incoming asteroid before we have time to do a dry run. So this work must begin now. It would not take large increases to NASA’s budget.

All civilizations that inhabit planetary systems must eventually deal with the asteroid threat, or they will go the way of the dinosaurs. We need to predict in advance when impacts are going to occur and, if necessary, shift the orbits of threatening asteroids. In effect, we must change the evolution of the solar system.