FRONTLINE

NUCLEAR ISSUES

'The idea of a nuclear deterrent is questionable'

Interview with Dr. Richard L. Garwin.

Dr. Richard L. Garwin, an experimental physicist, has worked in the field of nuclear weapons since 1950. As a consultant to the Los Alamos National Laboratory in the United States, he was responsible for converting into a workable blueprint the first rough outline made by Edward Teller of the hydrogen bomb. Subsequently, while working with IBM at its research division, he continued to be a consultant to Los Alamos on nuclear weapons, focussing on issues relating to the development and testing of diagnostic tools connected with the functioning of new nuclear weapon designs. He was also soon involved in the study of security issues as a consultant to the United States Government.

Dr. Garwin's assignments with the United States Government includes two four-year terms on the President's Science Advisory Committee (PSAC) in the Kennedy, Johnson and Nixon administrations. Apart from being on the Strategic Military Panel of the PSAC, he has been consultant and adviser on a range of subjects including military aircraft, transportation, naval warfare, anti-submarine warfare and aerial and satellite photography and been part of the early negotiations on a ban on nuclear tests.

Dr. Garwin has been actively involved in the debate over the Comprehensive Test Ban Treaty (CTBT), his inputs dealing mainly with the scientific issues involved. Dr. Garwin was a member of the Committee on International Security and Arms Control (CISAC) of the National Academy of Sciences when the committee issued its report titled "The Future of U.S. Nuclear Weapons Policy". The report advocated deep cuts in a very short term in the size of the U.S. nuclear arsenal and a unilateral no-first-strike declaration by the United States. Dr. Garwin served as a member of the Commission to Assess the Ballistic Missile Threat, appointed by the U.S. Congress in 1997, the so-called Rumsfeld Commission.

Dr. Garwin was recently in India as a member of the CISAC delegation that participated in an Indo-U.S. dialogue on nuclear issues organised by the National Institute of Advanced Studies, Bangalore. Dr. Garwin spoke at length to T. Jayaraman on a number of issues related to nuclear weapons in the South Asian and global contexts. Excerpts from the interview:

On Pokhran-II

I find the reason to test very peculiar. That China invaded India in 1962. And China is an aggressive country, that India has had three wars in 50 years with Pakistan. All those are no reason. If that is the best argument that could be made, it is a very poor argument for testing in 1998. You know, we got to test before China actually becomes a friendly capitalist country, then we will not have any excuse at all! Anyhow it does really seem to me, from the information that I have seen here and from what is known in the United States, that (it is) peculiarly, in India, a product of the scientific community. It is just not a demand by the military, but really the scientists doing it on their own. Which I think is actually a disservice to the country. There has always been a great esteem in India for research. But research really ought to advance the state of science far enough to benefit the country. When I was here in 1960 at the invitation of Homi J. Bhabha (the late Chairman of India's Atomic Energy Commission) I thought that really there ought to be much more effort on the technology side in India and perhaps less emphasis on pure scientific research which decoupled from the economy. But of course, I say that in the United States as well in some cases.

On safety of nuclear weapons

When you have nuclear weapons, the most important thing is that, depending on the numbers, you do not have a nuclear explosion when you do not want them. And over the years the United States worked out criteria, that in case of a maximum credible accident the probability of having a nuclear explosion should be less than four parts in a million. Not zero, but four parts in a million. And without accidents any nuclear weapon should have less than one part in a billion chance, in its lifetime, of having a nuclear explosion. And if there is an explosion in those circumstances, how do you find us a fission yield of less than two kilograms of high-explosive equivalent? That is really very small. That criterion was set by the Navy, I believe, so that if you have a nuclear explosion of that magnitude, the range in which people would be killed by radiation would be less than that in which they would be killed by the explosive itself - typically hundreds of kilograms of explosive is required to assemble a nuclear weapon. And in order to prevent accidents resulting in nuclear explosions, which are likely - like dropping the weapon, having a fire, having a short-circuit either in the launch vehicle or in the nuclear weapon itself - one needs to have redundant design, like the ones we use in airplanes to keep them flying, or (make it possible) for them to return home (airplanes crash after all). So redundant design so that you do not have a critical assembly in case, for instance, of detonating the explosive at a single point. In our early explosives, for instance, we kept the fissile material separate from the explosive and inserted it only in flight, of course manually, and then chemical, screw-type devices so that if there were an accident the explosive detonated would have nothing to do with the fissile material, you would not have any kind of nuclear reaction. So that is really important.

And then in the late 1950s we had the one-point safe criterion. Typically an implosion weapon has many points of detonation, and (you) must have at least two mechanically separate explosive points of detonation. Otherwise if you detonate at that one point you will get the full yield. So even though you can have an explosive distribution system, it will have to be at least two, and maybe, in some cases, dozens of points of detonation. And the system has to be such, in our case, that if you detonate the explosive at the worst point, again the fission yield should be less than two kilograms of high-explosive equivalent. So we actually tested all of our weapons to make sure that they are one-point safe. And if you have a lot of weapons it is desirable. "Weaponeers" want to do that. But it is an expensive system; it is a discipline that has to be maintained. One needs a good deal of independence and openness within the community in order to ensure that the weapons are safe.

The so-called "hydronuclear tests", conducted between 1959 and 1961, were important for the route that we chose. Because, in those days one could not calculate well enough the behaviour of the plutonium driven by high-explosive especially when it was not symmetrically imploded, or that was imploded by a one-point detonation. And so, that is at least a two-dimensional problem rather than a one-dimensional radius versus time and so although people designed these things to be one-point safe they were not sure. So we did a number of tests. It takes something like 60 such tests, because we do not just take a nuclear weapon and put it in a shallow well and ignite it at one point and say "good, it did not give a fission yield, it is one-point safe," because it might give a fission yield of kilotons and then it would not be properly contained. And so we made these tests with smaller amounts of plutonium and strong neutron sources to make sure that it did not have a yield and then gradually increased the amount of plutonium. But that whole approach is unnecessary if one instead says, "I will take the approach of separating the explosive from the fissile core," that would be perfectly valid, in which case there would be no relevance of one-point safety. It would be one-point safe because it would be separate. But we did use explosive testing and we had tiny yields because of this "creep up" process but we used a lot of plutonium and we used a lot of tests for that purpose. But they are not necessary any more because you could take the other approach, and the United States of course does not need to do it because all of our weapons in the enduring stockpile, about eight types, have been thoroughly tested. So we have no need, we signed the Comprehensive Test Ban Treaty. We do non-nuclear tests, there is no hydronuclear (test), our criterion is that we do not approach criticality, that we will fire no system that exceeds a reproduction factor of 0.8 or something like that. One is critical, and two, probably, is the condition that is obtained in a real nuclear explosion. So no explosions will take place at all.

I should also say that we do have hydrodynamic tests. These take place underground. I have been critical of the government for doing this. I have also said that we must be transparent about this and in fact the Department of Energy has been. Although they have gotten no publicity at all. Most of these tests, so far, have been little coins of plutonium driven by high-explosive. You are just looking at the surface behaviour of plutonium, and the scientists would like to understand in particular how, when they re-manufacture a nuclear weapon after 30 years or 50 years, they have to be very careful about the machining of the plutonium because they never studied that in the past. In my opinion that is justified, if the responsibility of these people is to maintain reliable nuclear weapons. But all these details have to do only with reliability, not safety.

On the dangers of unauthorised and accidental launch of nuclear weapons

But then beyond that is the question of the safety of the overall system. You may have a weapon that is perfectly safe, but it goes off when it is commanded to go off. And so the question is not just of the accidental explosion but of inadvertent or unauthorised launch. And specially where there are ethnic animosities, and especially when the weapons are ready to use, the whole system could be misused even by those people who have physical possession of the weapons or because of an invalid assessment of the situation. So a lot more effort has to go into ensuring that the system cannot be launched unless two people, who independently turn their keys, receive coded orders so (that) they cannot just collude to launch the weapon but they actually must enter the numbers that they receive by encrypted channels into the system. And furthermore the mechanical assurance that is in the United States, introduced in 1962, the Permissive Action Link, which goes to the warhead itself. Even though the aircraft may drop the bomb, the missile may be launched carrying the warhead to the terminal area, the bomb will not explode unless it has set into it the coded signal which is a separate authorisation. It is highly desirable to have such a thing. But those are routine but costly design features and there is no doubt that anybody who puts his or her mind to it and invests sufficient resources can accomplish that.

T. Jayaraman: But costly, you would say?

Dr. Garwin: Yeah, costly. According to The Atomic Audit, 86 per cent of the cost went to command and control, including the safety features. Then, of course, you have to modify the system as a result. You need to have more people involved, you need to have preventive means, and then at the same time authorising means. So there is this big tension between having a system that will actually launch the missiles when there is a valid national command authority decision to do so and a system that will not launch unless there is that authority. It is costly, it takes time. But it can be done.

If it is not done then, of course the weapons are a tremendous danger to the country themselves. The biggest danger is from unauthorised or accidental launch. The other side, in the case of the reasons for these nuclear weapons, are nuclear arms themselves. Even though your weapon which is launched by mistake may be out in the field someplace, the weapons that will deter you will hit your cities. It is not simply a nuclear explosion from the deployment side of the weapon. And in order to minimise all these problems, which are very serious problems, it is better not to have a doctrine or a system that needs to be launched rapidly. So not to have a doctrine to try to destroy something that is time urgent on the other side, like the nuclear force of the other side before it can be readied and used, and a system which is not vulnerable so that there is no question of it having to be launched before it is destroyed.

So that has been a terrible problem for the United States, has forced the evolution of our systems, from aircraft which had a reasonable probability of delivering the weapons but were very vulnerable to Soviet attack, especially when they got missiles (fortunately we never had an unauthorised nuclear explosion or an accidental nuclear explosion), to silos of missiles that could be launched rapidly but were vulnerable because they were outside. The silos became increasingly hard, because the accuracy on the other side improves.

So the best thing for a deterrent would be to have the actual warheads in caves, which could be gotten out through one or another entrance after hours or days and put them on remaining delivery means, but could be not days but a few weeks, and that way you could have mated only a small number of warheads to aircraft or to missiles. Then they are more vulnerable because there are fewer of them than if you have the warheads themselves properly protected and then you use whatever delivery vehicle was available afterwards.

But the whole idea of nuclear retaliatory forces, or a deterrent, is really quite questionable and what we need are not more independent nuclear forces, we need more people who will take a leadership role, in the United Nations for instance, and the other nations of the world, to respond to aggression, and especially to nuclear aggression. So I have advocated great reductions in U.S. nuclear forces. Instead of the 30,000 that we had in the peak of 1967 and perhaps something like 15,000 nuclear weapons now, we should have a total of 1,000 nuclear weapons as soon as possible. Within just a year or two, we could, not necessarily disassemble them all, but de-militarise them. And then try to transfer the responsibility for these nuclear weapons to the United Nations so that they would be used not for the protection of the United States but for the protection of all peaceful nations against attack. And especially if a nation espouses no-first-use which I advocate. In fact this is the one thing I say that speaks for the Committee on International Security and Arms Control (CISAC), in that blue volume (refers to the CISAC report, The Future of U.S. Nuclear Weapons Policy), is that we explicitly advocate a unilateral no-first-use posture on the part of the United States. Not a treaty, it does not matter whether other people have no-first-use, we should have no-first-use of nuclear weapons. And that means no use of nuclear weapons even in response to a biological or chemical attack. So if you have a no-first-use doctrine, and you really mean it, then there is no hurry to respond through nuclear weapons.

The first-use and the vulnerability and the need to launch before they are destroyed is what has driven our nuclear force to great numbers and to very dangerous postures.

Effects of nuclear weapons and deterrence

The effects of nuclear weapons on cities ought to be more thoroughly discussed in the press. We have a lot of studies in the 1960s made in the United States by the Office of Technology Assessment of the effects of nuclear weapons detonated on Detroit, or other places, Birmingham, England, some Soviet cities. And if we get back even to the 10-20 kiloton devastation of Hiroshima and Nagasaki. 100,000 people killed in each case, the cities levelled. But that is still what will happen. Some people say, well, Pakistan would never use a nuclear weapon against India, and vice versa because we are so close and the winds will carry the debris and thousands would be killed and people will not stay at home. But that is not true. The devastation from an air-burst is local. Now if it is not raining, and basically you have no-first-use and so whatever you have no particular urgency of making a first-strike, there is no particular urgency. But if it is not raining, then the air-burst carries all of the fission products and radioactive materials up and it comes down all over the world in a couple of years. The fact is that the United States had many dozens of detonations, air-bursts, in its continental test site in Nevada and we did have fallout across the country. But overall, these magnitudes are really important, the 300 or so megatons of fission in the atmosphere involved in the nuclear tests would probably have killed, let us say 300,000 people worldwide, and that is over generations, due to fallout. Now that is a thousand people per megaton for worldwide fallout. So for 10 kilotons what would be the worldwide fallout contribution? It is about 10 people. So you kill 100,000 (or) 200,000 people locally from the intended explosion, you kill 10 people over many generations in the whole world. That is not enough to deter the use of nuclear weapons. It was not even enough to deter the testing of nuclear weapons. So people really ought to understand what would be the effect of nuclear weapons, the reason why if somebody has 100 nuclear weapons we should not feel secure. If you have a thousand nuclear weapons or a million nuclear weapons it really does not change your security at all to be able to wreak much more devastation on them. They are adequately deterred from the intention of the use of nuclear weapons if you had 10 probably; but much better the deterrence because there is nothing today and because the nations of the world would respond rather than having every individual nation have a nuclear system and threaten devastation on its neighbours.

On the anti-ballistic missile defence debate in the U.S.

McNamara in his famous San Francisco speech, in 1967 I guess, said that we were going to deploy a limited defence system against the Chinese. This was an excuse. McNamara said that the Chinese had on their launch pad an ICBM which might be fired, tested, within six weeks. But it took 11 years before it came. So it was an excuse for deploying a national missile defence system because there were the elections coming up and President Johnson thought that he was vulnerable to charges from the Republicans that he was not protecting against the nuclear threat.

And that is where we are now. Again the Republicans in Congress have been badgering the Clinton administration, saying that in 1983 Reagan announced the Strategic Defence Initiative. And they say what everybody knows really that, if you stop to think about it, you do not really have a defence against nuclear-armed ballistic missiles. We could not defend against even one nuclear weapon, coming maybe from North Korea. But the fact is, I was on the Rumsfeld Commission last year and I have been on television with Don Rumsfeld and others, but the fact is we are totally vulnerable as a nation, not only to nuclear weapons on ballistic missiles, but to nuclear weapons that are detonated in harbours, to nuclear weapons on short-range ballistic missiles on ships, to cruise missiles, with nuclear weapons that are available commercially that can be launched from ships 20 km, 100 km offshore. And it would be a folly, you notice the word folly, to defend against the nuclear-armed ballistic missile from so-called rogue nations. Again it would be much easier for them, and more reliable if they do not have a much bigger force, if they would use these other delivery means. And so it is totally unbalanced. There is a very careful choreography on the Republican side of how they present this, this threat and a make-believe, without this, create this total vacuum of the discussion of the other threats. And the fact that this is a dangerous world. And the first thing that one should do is, first understand that any nation that uses nuclear weapons against any other nation is going to be wiped out. At least if it uses (nuclear weapons) against nuclear powers. But if it uses them against non-nuclear powers it is going to have terrible problems as a result. Because of the response.