The Nukes We Need: Adapting our arsenal to today.

When President Bush peered across the DMZ into North Korea recently, the most important things to see were out of sight. The North Koreans have two, related proficiencies: weapons production and tunneling. They built an underground city to conceal work on the No Dong ballistic missile, tested in 1993. In 1998, a tunnel complex big enough to house a plutonium production plant was discovered near a nuclear research center supposedly shut down under the 1994 U.S.-North Korean Agreed Framework. Meanwhile, the North Korea forward staging areas near the DMZ have more than 4,000 tunnels and bunkers.

The North Koreans may specialize mainly in backwardness, but in this they are on the cutting edge. Russia, China, Iraq, and other countries all have a new appreciation for the bunker mentality. The Chinese learned from NATO air campaigns in the Gulf and the Balkans that digging is the best way to counteract NATO's mastery of the air. As for the Russians, they have a tradition of digging going back to the Cold War, with some bunkers in Moscow estimated to be 1,000 feet deep, and one facility under Yamantau Mountain in the Urals reportedly as large as the area inside the Washington Beltway.

As the war on terrorism has now also become -- at least in the president's rhetoric -- a war on weapons of mass destruction (WMD), this drive underground cannot be ignored, especially in U.S. nuclear policy. The U.S. is finally -- a decade late -- taking account of the end of the Cold War by drastically reducing its operational strategic nuclear force from roughly 6,000 warheads to 2,000. But it makes no sense to react to the changed international environment only by scrapping the old force. The arsenal should also be updated to deal with new realities, most importantly by developing an earth-penetrating nuke, designed to target deeply buried WMD sites.

William Schneider, chairman of the Pentagon's Defense Science Board, explains that there has been a revolution in the economics of digging in recent years, thanks mostly to work on the Channel Tunnel. Run as a commercial venture, the Chunnel project emphasized innovation, producing technologies cheaper and more efficient than the traditional blast-and-cut methods. Now, according to Schneider, for a few million dollars a country can buy a Japanese, Finnish, or German machine that can dig an 18-meter-wide hole at a rate of 70 meters a day.

The U.S. has been working to counteract this new underground capability with conventional weapons. The GBU-28 "Bunker Buster" -- a 5,000-lb. laser-guided bomb rushed into production for the Gulf War -- has penetrated over 20 feet of concrete and more than 100 feet of earth in tests. Even bigger and better weapons are in the works. One is called "Big BLU," a sort of plus-size daisy cutter (the 15,000-lb. bomb designed to blast clear a 600-meter area that has had a starring role in Afghanistan).

With these bombs, the military is essentially attempting to create something that has the power of a nuclear weapon without actually being nuclear. But the explosive force of conventional weapons can be pushed only so far. In addition, some bunkers are simply too deep and too hard. A recent Pentagon study concluded, "Even with the current strategy and acquisition initiatives, the United States will still not be able to hold all known or suspected Hard and Deeply Buried Targets at risk for destruction, especially the deep underground facilities."

This means that the only conventional force to which some targets will be vulnerable is an invasion or special-forces raid. But not all future conflicts will resemble Afghanistan or the Persian Gulf War, when the U.S. had total control of the skies and could operate almost at will. The ground- force option, in addition to risking American lives, would almost always fail what Keith Payne, head of the influential National Institute for Public Policy (NIPP), says should be a three-pronged test for taking out a dangerous WMD site in a crisis: It should be prompt, predictable (for our leaders, not the enemy), and definitive. A conventional raid might well be none of the above.

Which leaves nuclear weapons. "From the public record, I don't know of any non-nuclear way of dealing with this underground threat promptly and conclusively," says Payne, whose work has been the basis of the Bush administration's recent reevaluation of U.S. nuclear strategy. A nuke would have several advantages. It passes the prompt-predictable-definitive test. It also might not require intelligence as precise as that necessary for a conventional weapon -- the explosive force provides room for error.

And it would destroy the targeted WMD agents rather than spread them as a conventional blast might. As a report from NIPP recently put it, chemical and biological agents "are extremely difficult to destroy (or sterilize) definitely, as opposed merely to disperse, except by means of the extraordinary heat and neutron flux generated by nuclear explosives." A nuke, of course, would create another hazard -- radioactive fallout -- but a low- yield weapon could be designed to minimize it.

The problem is that we don't have this kind of weapon. Given that we have been in the nuclear business for 50 years, how is that possible? A host of strategic and technical reasons account for it, together with a perverse arms-control orthodoxy that has attempted to keep the U.S. arsenal as massive, inaccurate, and potentially horrific in its effects as possible.

Mutual Assured Destruction relied on the "balance of terror," on the willingness of the U.S. and the Soviet Union to hold their populations hostage. Any highly accurate or earth-penetrating weapon that instead would have been effective against specific military targets was considered "destabilizing" -- a "war-fighting" weapon rather than a weapon of generalized terror. So, U.S. nukes tended to be designed for killing lots of Russians rather than destroying narrow military targets.

This was also simply easier as a technical matter. Getting a warhead to drive into the ground, then explode, is a technical challenge on the order of getting a car to drive through a wall, then have its left-turn signal flash. As nuclear expert Robert Barker explains, the weapon has to be fast enough to enter the ground, but not so fast as to destroy the warhead and the mechanism that triggers it. The warhead design, needless to say, must be very rugged. These difficulties, however, are probably surmountable. We dealt with some of them in creating nuclear artillery shells, which had to withstand enormous G- forces.

But arms-controllers aren't interested in having these problems surmounted. In fact, liberal Dr. Strangeloves want U.S. nuclear weapons to be as indiscriminate as possible. In their 1983 letter on nuclear weapons, the U.S. Catholic bishops opposed making nukes more accurate. This would seem to be in direct contradiction to Just War Theory, which emphasizes "discrimination" in order to minimize civilian casualties. The bishops' spirit lives on in 1994 congressional language prohibiting the U.S. from "research and development which could lead to the production by the United States of a new low-yield nuclear weapon, including a precision low-yield warhead."

There have, nonetheless, been attempts to update the U.S. arsenal. The Clinton administration worked to develop a penetrator without actually creating a "new" weapon. It gave an existing nuke, the B-61, a new needle- shaped hardened case. But the re-jiggered B-61 relies only on its terminal velocity -- it is dropped, unpowered, from the air -- to drive it into the earth. A true earth-penetrator would be powered so that it could hit the ground at much higher speeds. (The B-61 burrows about 20 feet deep into dry earth, whereas a true penetrator might need to go through 100 feet of granite.)

One advantage of the B-61 is that its yield can be adjusted downward from its high of 340 kilotons. A low-yield penetrator would probably be 10 kilotons or less (the Hiroshima bomb was 15 kilotons). The issue of size is so important because the U.S. wants a weapon that can be used in a crisis without the sort of massive collateral damage that would be simply unacceptable. As Stephen Younger, then an associate lab director at Los Alamos, wrote in a controversial June 2000 paper, "A reliance on high-yield strategic weapons could lead to 'self-deterrence'" -- in other words, an unwillingness on the part of the U.S. to use its own weapons.

If the U.S. wants to develop a useful new nuke it will have to cross another arms-control taboo -- against nuclear testing. In theory, it might be possible to jerry-rig a new weapon without testing, but that would be far from ideal as a technical matter. In any case, the "stockpile stewardship program," which was supposed to supplant testing with computer models, was drastically underfunded in the Clinton administration, limiting its usefulness. It is questionable whether we can retain confidence even in our current arsenal without testing, since our warheads were designed to last only 15 to 20 years.

In 1992, the first President Bush signed a bill instituting a voluntary moratorium on testing, although he made clear that he opposed it, and the Senate declined to make it permanent in 1999 by refusing to ratify the Comprehensive Test Ban Treaty. Arms-controllers argue that if the U.S. eschews testing, it will create a new anti-nuclear "norm" around the world. But the Chinese, French, Indians, Pakistanis, and perhaps the Russians have tested subsequent to the U.S. moratorium.

For arms-controllers, the underlying rationale for the taboo against testing seems obvious: to prevent the U.S. from developing a new weapon, and ultimately to force the existing arsenal to die on the vine. It ensures that the U.S. has an aging, less and less reliable arsenal, built for a long-past strategic threat that bears little resemblance to the present one. The last new U.S. nuclear weapon was fielded in the 1980s, which means that it was designed in the 1970s. The longer the U.S. goes without designing and manufacturing a new weapon, the less capable it will be of doing so, as the expertise and manufacturing base wither away.

The thrust of the arms-controllers seems to come down to limiting U.S. power. Consider: Arms-controllers oppose American missile defenses because it is supposedly destabilizing for the U.S. to have sites that can be protected from rogue-state (or Russian or Chinese) attack. On the other hand, arms- controllers apparently don't mind rogue states' having sites that can be protected from U.S. attack. There is no effort to create an international treaty keeping rogues from digging deep bunkers. And arms-controllers oppose a new U.S. weapon that would be capable of holding these sites at risk. Assured destruction apparently looks much better when it applies only to the U.S.

During the debate over missile defense last year, arms-controllers made nice sounds about deterrence (who needs missile defense when you have deterrence?). But deterrence depends on credibility. As long as the U.S. arsenal is chock-full of weapons that can only cause indiscriminate damage -- and mass civilian casualties -- it doesn't seem credible that we will use them, and so their deterrent value is lost. Which is exactly the way arms- controllers like it -- the U.S. arsenal becomes, in effect, irrelevant.

Their complaint about a low-yield nuke is exactly that, in the words of Congress in 1994, it would "blur the distinction between nuclear and conventional war." Or, as a Federation of American Scientists report puts it, "adding low-yield warheads to the world's nuclear inventory simply makes their eventual use more likely." Actually, that's not quite true: It makes their use seem more plausible, which in turn makes their use less likely.

Without a low-yield penetrator, rogue-state leaders are able to have, in effect, a safe haven for themselves and their weapons of mass destruction. If they knew they didn't have any such protection, it might deter them from threatening or attacking the U.S. in the first place. This is how deterrence works. But deterrence also fails, in which case a low-yield penetrator might be necessary to preempt an imminent attack or to retaliate against one -- and keep more from coming -- by, say, hitting all of Saddam Hussein's command bunkers.

The Bush administration is at least making moves toward updating the arsenal. The Minuteman missile is having its 1970s-era guidance system overhauled, so some attention is being paid to applying the benefits of new precision technology to existing nukes. The administration has also undertaken a study of the need for a new earth-penetrating low-yield nuke. Developing one would mean running into the teeth of congressional and international opposition, and it is, of course, extremely unlikely that such a weapon would ever be used. Conventional options would almost always be preferable.

The key word, however, is "almost." Nuclear weapons have always been available as a bad option that might be necessary only if every other option is worse. The world, unfortunately, didn't stop offering us bad options back in 1989. We should stop pretending otherwise.