How hard could it be to find hundreds of tons of radioactive nuclear material? We've certainly got plenty of motivation to keep tabs on this stuff. There's the threat of terrorist groups like Al Qaeda, the standoff between Pakistan's and India's arsenals and North Korea's Kim Jong Il. Iran, the next big nuclear challenge, already has missiles that can strike Israel and a thriving civilian nuclear-power program. It claims to have no ambition for nuclear weapons, but verifying this is critical. We must know how much to press for a diplomatic solution or how seriously to consider a military strike.
Nuclear intelligence, however, is problematic. Despite all the high-tech gear that intelligence agencies have developed, facts on the ground are so thin that the whole question of what countries like Iran are doing with nuclear weapons is vulnerable to manipulation by policymakers. Who can forget how Condoleezza Rice, as head of the National Security Council in September 2002, declared that the United States could not let the "smoking gun" of Iraq's nuclear program become a mushroom cloud? Too often intelligence about clandestine nuclear programs is spotty and can easily be misread.
Many people might think that technology allows spies to easily snoop on weapons plants, but that is not true. Satellite images can flag buildings suspected of being part of a nuclear-weapons program, but they cannot reveal what goes on inside. It's easy to hide a gas-centrifuge plant, which is needed to take ordinary uranium and turn it into highly enriched uranium for weapons. Spy-satellite photos of new construction at pre-1991 Iraqi nuclear-weapons sites helped convince the CIA that Iraq was building a uranium-enrichment plant. It turned out to be a repair shop for military radar.
The chief evidence U.S. intelligence agents mustered on Iraq was a shipment of thousands of aluminum tubes thought to be key components of a gas centrifuge. Intercepted nuclear hardware and communications are powerful intelligence tools, but they're often ambiguous. In the case of Iraq, the aluminum tubes were probably intended for artillery rockets. Another tactic—placing "bugs" inside such dual-use equipment and arranging to have it traded on the black market—uncovered secret installations in Libya. But it failed in Syria; the dual-use items could not be linked to nuclear weapons.
Since nuclear materials give off radiation, spies can use sophisticated detection equipment to identify them. But these detectors work only within short ranges—they have to be placed or flown near a nuclear site. U.S. intelligence agents identified Pakistan's gas-centrifuge enrichment plant in the 1980s by secretly placing radiation detectors in fake rocks and roadside mile markers. But this won't work in Iran and North Korea. Since nobody knows where to look for the plants in these countries, getting detectors within range is nearly impossible.
Spies and defectors can reveal secret sites and the inner workings of nuclear programs, but this method is hit or miss. Collaborators helped reveal Pakistan's and Taiwan's nuclear-weapons programs in the 1980s. But defectors proved to be misleading in Iraq and have not been useful for years in Iran. Placing spies inside programs is a daunting challenge, especially in regimes such as Iran and North Korea and in terrorist groups.
Looking back on Iraq and North Korea, we see plainly that information given voluntarily to the International Atomic Energy Agency has been far more accurate than anything intelligence agencies have been able to get by stealth. The IAEA uses commercial satellite imagery, but it can also demand access to suspect sites and ask tough questions. Inspectors on the ground at declared nuclear sites and suspect sites can swipe surfaces for traces of enriched uranium or plutonium. Whereas the intelligence community often generates suspicions, the IAEA can actually follow up.
The poor track record of nuclear-weapons intelligence makes it almost impossible to know what to believe about Iran. Those who favor a diplomatic solution take comfort in claims of U.S. intelligence agencies that Iran won't obtain nuclear weapons until 2010 or even 2015, but the basis of this estimate has not been made public. Hawks have tried to show that Iran is closer to having nuclear weapons, but they have no greater claim on the truth. Recently the rhetoric has gotten heated. Some Israeli reports claim that Iran could have enough nuclear-explosive material for a bomb by the end of this year. (The Iranian opposition group the National Council of Resistance says that Iran has already crossed this threshold.) U.S. Vice President Dick Cheney warned in January that a nuclear-armed Iran was a "serious prospect." Both Cheney's and Israeli comments are insinuations, not proof.
The most worrisome trend is recent efforts by some U.S. officials to denigrate the IAEA for not declaring Iran's imminent weapons capability. The IAEA, acutely aware of the threat of a nuclear Iran, holds that evidence must be collected systematically and objectively. Heeding these experts might help repair the damage caused by getting it wrong on Iraq.