U.S. Weighs How Best to Defend Against Nuclear Threats

But many other analysts looking at the data, such as Harvard University proliferation expert Graham T. Allison, conclude otherwise.

Vayl Oxford, director of the Homeland Security office Bush created a year ago today to put nuclear detection efforts back on track, said critics' concerns reflect a Cold War assumption that solid intelligence can be obtained against a terror group. The country must also consider its vulnerabilities and the consequences of the worst catastrophes, he said, which in this case tip the scale toward action.

"If you don't see a direct intelligence report that says there is something there, someone will leap to the conclusion the threat is not there," Oxford said. "But I don't think it's political hype. It's prudent planning to take action on this count. Sitting in hindsight saying 'Why didn't we see it in the intelligence?' is not the kind of hearing I want to go to."

Prompted by influential advocates including Cheney, former NATO ambassador David M. Abshire and former Lockheed Martin Corp. chief Norman R. Augustine, President Bush signed the 14th Homeland Security Presidential Directive last April 15. It consolidated development of countermeasures to a smuggled radioactive weapon that had been split among the Pentagon, the Energy Department and other federal agencies into the new Domestic Nuclear Detection Office, headed by Oxford. The office is designing a national detection system and a global strategy.

The emerging effort calls for thousands of scanners of all types throughout the country. These include backpack or handheld "cellphone" devices, units mounted on vehicles, and stationary portals to scan railcars and shipping containers, Oxford said.

The United States is also working with Canada and Mexico on strategies to deploy detectors and with the United Kingdom to exchange technology, he said.

As spending in the area grows -- from about $20 million in 2000 to a proposed $536 million in 2007, the largest increase for any Homeland Security agency -- the question is whether the government can learn from its mistakes. A post-Sept. 11 program to install more than 470 radiation monitors at checkpoints, ports and mail facilities has built a troubled legacy.

"That was not based on any scientific, technical or cost-benefit analysis that was the next best place to spend our dollars," said one senior federal scientist at a national laboratory. "Somewhere in there, just the importance of showing that you're doing something . . . came into play."

The early devices cannot distinguish between sources of radioactivity and are prone to false alarms. Some 10,000 such false hits occurred between May 2002 and March 2005, U.S. customs officials said last year.

Closing one terminal for one hour can cost $500,000, according to the Port Authority of New York and New Jersey, which averaged 150 nuisance alarms a day last summer. As a result, the machines' sensitivity was often turned down to a threshold that compromised their ability to detect actual threats, congressional auditors reported last year.

Rivalry among government agencies also hobbled the effort, as did poor leadership, underfunding and political meddling by Congress, former Homeland Security inspector general Clark Kent Ervin said.

Experts also say the nuclear detection effort at U.S. borders must be only one component in a more aggressive effort to secure weapons-grade material at its source from places such as the former Soviet Union and Pakistan and to detect the flow of illicit materials along foreign trade routes.

In the face of such challenges, the Domestic Nuclear Detection Office expects to choose next month from 11 vendors competing for a $1.5 billion contract to deploy thousands of Advanced Spectroscopic Portals (ASP) by 2011, Oxford said. The goal of ASP is technology that cuts down false alarms and distinguishes among natural occurring radioactive materials.

But such detectors are not as effective at picking up highly enriched uranium or weapons-grade plutonium or at detecting shielded material. So this month the detection office also is accepting proposals for a second, similarly sized program to develop prototypes and eventually deploy automated cargo imaging systems that can actively scan for the presence of nuclear material, or shielding, such as lead, Oxford said.

The projects present a microcosm of technical questions that have bedeviled the effort.

Should the government throw tax dollars into increasing the use of existing technology, flaws and all, or put its eggs in the basket of breakthrough research? What are the limits of the technology? How much would improvements cost?

In 2004, a panel of physicists split on the answer. So the government both funded long-range research and deployed existing technology.

Having learned the limits of the equipment, Oxford said, his office has conducted extensive field tests, developed plans to cut down on false alarms and committed itself to speeding the acquisition of more sophisticated devices.

But scientists from one of 11 competitors for the lucrative new contracts, Ametek Inc., say the government is in danger of repeating past mistakes. Ametek says it has engineered ASP radiation detectors relying on high-purity germanium technology that are 20 times more accurate than sodium-iodide sensors used now.

"Post-9/11, money has been spent on nuclear radiation measurement technology that doesn't work. Here we are again," said William J. Burke, a spokesman for Ametek and its lunch-pail-size entrant, the "Detective." "Are we going to . . . take the technology that will actually do the job?"

Germanium detectors probably will cost 10 to 20 percent more than competing technology, Oxford said. Considered the gold-standard in the laboratory, they have been difficult to use because they require cooling to 280 degrees below zero. Computing advances, meanwhile, are improving other technology, Oxford said, which may be integrated.

Richard L. Wagner Jr., leader of the Nuclear Vision Project at Los Alamos National Laboratory and a leader in the effort, estimated that a domestic nuclear detection system can be built for about $20 billion over two decades, a bargain compared with such programs as the "Star Wars" missile defense system.

Matthew Bunn, senior research associate of Harvard's Managing the Atom project, supports spending on detection but added, "We shouldn't fool ourselves into thinking that provides a very reliable defense."

"These materials are small and easy to hide," Bunn said. "It's really like a football team defending on its own goal line, but the goal line is thousands of miles wide, and millions of legitimate vehicles and passengers pass over it every year."