|Source: U.S. News & World Report, Oct 25, 1999 v127
Title: Can the U.S. ever really know for sure?(problems in verifying
Full Text COPYRIGHT 1999 U.S. News and World Report, Inc.
Apart from the ferocious politics of the issue, the test-ban-treaty debate has some substantive disputes at its core. Most notably: Can existing technology detect other countries' nuclear tests? And can the U.S. government ever be sure that the nation's nuclear weapons will work without actually testing them?
Treaty opponents argue that there's no way to distinguish small nuclear tests from conventional explosions or even earthquakes, enabling other countries to cheat. But prominent geophysicists and seismologists disagree. Ernest Moniz, the under secretary of energy and the department's chief scientist, acknowledges that tests of crude bombs could go undetected. But he insists U.S. sensors can pick up any tests of advanced nuclear weapons. "The issue is whether the verification regime [can] detect militarily or strategically significant tests. Our answer is, yes," Moniz says.
Equally perplexing is whether it's possible to determine the safety and reliability of the nation's nuclear arsenal by simulating blasts instead of setting off bombs underground at the Nevada test site. "It is intuitively difficult to ensure that something works if you don't try it," says Victor Reis, a nuclear-weapons expert. But he notes the Energy Department has been developing a massive program to ensure that the nation's aging stockpile of nuclear weapons is up to speed.
John Browne, director of Los Alamos National Laboratory, said at a recent Senate hearing that the effort--estimated to cost $45 billion over 10 years--will be an "unprecedented technological challenge," as tough as it was to put a man on the moon. The reason: Nuclear weapons are extraordinarily complex, containing some 4,000 different parts and materials. What's more, many of the nation's nuclear arms are now past their prime, and nobody knows how plutonium, the key ingredient, will perform as it ages.
Until testing was banned in the United States in 1992, the way government scientists made sure nuclear bombs were effective was to explode them. But now they can only test individual components and use computer models to simulate actual blasts. So every year, the Energy Department's national laboratories randomly select about 10 of each type of weapon, dismantle them, and test the parts. They perform exhaustive tests to determine how the weapons will weather certain conditions and plug their findings into supercomputers. Is the process foolproof? "It's an excellent bet, but it ain't a sure thing," says Bruce Tarter, head of Lawrence Livermore National Laboratory. That's the trouble.