On January 4, 2007, the Wall Street Journal published an op-ed entitled “A World Free of Nuclear Weapons,” written by an impressive array of statesmen: former secretary of state George Shultz, former secretary of defense William Perry, former secretary of state Henry Kissinger and former senator Sam Nunn of Georgia. In the article the authors worried that the likelihood of international terrorists acquiring nuclear weapons is increasing. They asserted that “unless urgent new actions are taken, the U.S. soon will be compelled to enter a new nuclear era that will be more precarious, psychologically disorienting and economically even more costly than was Cold War deterrence.” Invoking President Ronald Reagan’s call in the 1980s for the abolition of all nuclear weapons, they endorsed “setting the goal of a world free of nuclear weapons and working energetically on the actions required to reach that goal.”

Recently, however, a counterargument has been advanced—by NASA. In 2005 Congress ordered the space agency to analyze the alternatives that it could employ to divert a near-Earth object (NEO)—an asteroid or comet—if one was found to be on a collision course with our planet. Last March, NASA submitted a report entitled “Near-Earth Object Survey and Deflection Analysis of Alternatives,” having first coordinated its response with the White House, the Department of Defense and the Department of Energy. In its report NASA chose to analyze only the highly improbable threat posed by large NEOs, which very rarely strike Earth, in lieu of the more realistic danger of a collision with one of the cohort of smaller NEOs, which are far more numerous. What is more, the report emphasized the effectiveness of nuclear explosions in providing the force to deflect an NEO from a collision course, but it completely neglected the need for precision in such a procedure.

This analysis is seriously flawed. It is important not only to deflect an NEO from a collision course with Earth (primary deflection) but also to avoid knocking the object into a potential return orbit that would cause it to come back a few years later (secondary deflection). Nuclear explosions are not controllable in this way. But a nonnuclear kinetic impact—that is, simply smashing a spacecraft into an NEO—can provide the primary deflection for the vast majority of objects, and a precise secondary deflection, if necessary, could be performed by an accompanying gravity-tractor spacecraft, which would be needed in any event to observe the NEo deflection and its aftermath [see "Gravitational Tractor for Towing Asteroids,” by Edward T. Lu and Stanley G. Love, in Nature; November 10, 2005].

Nuclear explosives would be needed only for deflecting the largest NEOs, which are the least common and most easily detectable objects. Scientists are not concerned about a collision with an extremely large NEO—say, 10 kilometers in diameter—because all these objects have been discovered and none currently threatens Earth. Big things are easy for astronomers to find; the smaller objects are what we have to worry about./Of the estimated 4,000 NEOs with diameters of 400 meters or more—which includes all objects that might conceivably require nuclear explosives to divert them—researchers have so far identified about 1,500. And if NASA meets the search goals mandated by Congress, it will locate 98 percent of these objects and calculate 100-year projections of their orbits by 2020.

As NASA continues to find big NEOs, the calculations of risk change accordingly. A decade ago, before astronomers began to systematically locate NEOs larger than 400 meters in diameter, they estimated that we faced a statistical risk of being struck by such an object once every 100,000 years. But now that researchers have identified and are tracking about 37 percent of these NEOs, the frequency of being hit by one of the remaining large objects has dropped to once in 160,000 years. Unless NASA finds a large NEO on an immediate collision course by 2020 (a very unlikely event), the frequency of a collision with one of the 80 still undiscovered objects (2 percent of 4,000) will drop to once every five million years.

Thus, the probability that nuclear explosives might be needed to deflect an NEO is extremely small. And even this minuscule probability will diminish to the vanishing point as researchers improve nonnuclear interception technologies. After 2020 the need to keep nuclear devices on standby to defend against an NEO virtually disappears. As a result, the decision to move toward the worldwide elimination of nuclear weapons can be made strictly on the basis of human threats to global security. Extraterrestrial dangers need not be considered.