Missile Defense--NMD, TMD, EMD ...

by Richard L. Garwin
Council on Foreign Relations, New York

Aspen Institute, Berlin
March 1, 2001

For more than 40 years, the United States and the Soviet Union (now Russia) have lived with the fact of mutual vulnerability, and this was codified by the U.S.-Soviet ABM Treaty of 1972 that banned any effective national missile defense against strategic ballistic missiles (those with ranges in excess of 5500 km). For years the deterrence of large-scale nuclear war by threat of retaliation had been implemented with forces of bomber aircraft and nuclear and thermonuclear weapons; but soon strategic ballistic missiles became the weapon of choice. There was some air defense of uncertain (but low) effectiveness, but there was no defense against the missile, except for a long-standing ABM system around Moscow that was limited under the ABM Treaty to 100 interceptors. This could be handily overcome by targeting more weapons than that at Moscow, even though relatively few sufficed to overcome the ABM system.


But leaving U.S. citizens "vulnerable" to destruction by Soviet or Russian missiles has been a persistent issue between the political parties in the United States. When Democrats held the White House, they were routinely criticized for the fact of vulnerability, as if it were their fault. Citing vast improvements in technology, ranging from computers to electro optics, to lasers, critics of deterrence argued that it would be not only moral but feasible to replace deterrence by an impenetrable defense. President Reagan's March 23, 1983 "Star Wars" speech, which launched the Strategic Defense Initiative, was typical in its overreaching demand. He asked that the scientists who gave us nuclear weapons give us the means to render those same weapons impotent and obsolete, by perfecting a system to destroy missiles and their nuclear warheads in flight. Of course, this would do nothing to prevent nuclear weapons being delivered by other means. Ultimately, SDI collapsed when Admiral William Crowe, Chairman of the Joint Chiefs of Staff, demanded of the SDI Organization a system that would destroy some fraction of 300 warheads, rather than the full 6,000 Soviet force. The $75 billion cost estimate for this system that would enhance deterrence by improving the survivability of U.S. ICBMs, rather than replace deterrence, rendered it of no interest to the military.

It was clear to President Lyndon B. Johnson that the opposing party in the 1968 election campaign would charge his Administration with laxity in not protecting the American people, so in 1967 then Secretary of Defense Robert McNamara, in a famous speech in San Francisco, gave all the reasons why a missile defense against Soviet missiles made no sense, but closed by stating that the Johnson Administration would deploy a light defense of U.S. territory against Chinese ICBMs. McNamara noted that a Chinese missile was on the launch pad and might be tested within six weeks; in fact, it took China 11 years to acquire ICBMs.

Similarly, the Clinton Administration in 1997, having faced repeated calls for the deployment of missile defenses, adopted a "thee plus three" program, so that within three years (the year 2000) a defense would be developed so that it could be deployed, given a decision in 2000, within three additional years-- so by 2003. And if the threat had not emerged in the year 2000, then development would continue to take place, with deployment three years following a decision, whenever that would be made.

This concept was based on the assumption that the United States would have at least three years warning of the creation of an ICBM force by any threatening state, and the Clinton Administration specifically identified a few "rogue states" such as North Korea, Iraq, and Iran as the targets for such a defense. In July 1998, the report of the 9-person "Rumsfeld Commission"1 was published, with the key judgment that North Korea, Iraq, or Iran, with foreign technology transfer and an urgent and well-funded program, could develop an ICBM within five years, and for several of those years, the United States might be ignorant of that program. The result of the program could be the deployment of a few inaccurate, unreliable ICBMs. The unanimous report of the Rumsfeld Commission also noted that any of those countries could more readily and sooner build more accurate short-range cruise missiles or ballistic missiles, perhaps to be launched from third-party cargo ships against U.S. coastal cities. It is important to note that the Rumsfeld Commission did not discuss or urge deployment of any missile defense; its task was to assess the threat-not what to do about it.

It was a peculiarity of the Clinton program that from the moment of its inception, the Pentagon would describe the purpose of the program as defending against a few ICBMs from a rogue state, but always with a subsidiary capability to destroy missiles launched accidentally or without authorization.

In 1999, Congress passed legislation, signed by President Clinton, which clearly stated that it was the policy of the United States to deploy, as soon as technologically feasible, a defense against limited ballistic missile attack, whether intentional, accidental, or unauthorized. No mention of rogue states; a limited ballistic missile attack from China or Russia (or France or Britain, for that matter) was to be intercepted.

Until October, 1999, the proposal was to deploy 20 ground-based interceptors either in Grand Forks, North Dakota, or at a new site in Alaska, to counter a few ballistic missiles launched initially from North Korea, or perhaps ultimately from Iraq or Iran. But that month, in testimony to the U.S. Congress, Under Secretary of Defense Walter B. Slocombe indicated that the program would henceforth be planned to deploy initially 100 interceptors, with potential growth (in the program phase C-3) to 200 or 250 interceptors. No reason was given, but one can infer that because the ABM Treaty permitted 100 interceptors at a single site, the Clinton Administration feared ridicule for a proposal to deploy only 20 interceptors.

As you all know, in September, 2000, President Clinton decided that the NMD program had not met his criteria for technical feasibility, with the emergence of a threat, at reasonable cost, and with allowable impact on the possibility of major reductions in nuclear weaponry. In fact, only one of three NMD interceptor tests has succeeded, and that first test probably only because a large balloon attracted the attention of the interceptor, which then saw the mock warhead in its vicinity.

The Clinton system was to be a mid-course intercept, in the vacuum of space, by large ground-based interceptors. The launch of the ICBM with its nuclear warhead would be observed first by satellites the U.S. has had operating since 1970. These so-called DSP satellites in geosynchronous orbit have seen every ballistic missile launched for many years, and would surely see an ICBM from any of these three states. A rough trajectory would then be obtained, and some minutes later, the reentry vehicle ("RV") in its ballistic arc through space would be detected by powerful early warning radars, such as those in Fylingdales, Britain, or Clear, Alaska, or Thule, Greenland. Unfortunately, there is no radar in an appropriate position to detect the RV from a North Korean ICBM launched north to northeast at the United States, and the proposal was to build an advanced X-band radar at Shemya, Alaska. In fact, a much simpler radar would have done the job.

To complete the intercept, several of the ground-based interceptors would have been launched from North Dakota or Alaska, aiming at the position predicted for the RV in space at the time of intercept. As the interceptor came within perhaps 1000 km of the RV, its own homing system would detect the RV by the thermal radiation it emits (assumed to be about as much thermal radiation as one of us emits). The interceptor employs the marvels of modern control technology and computation to steer itself to collide with the RV. In fact, such hit-to-kill intercepts have been demonstrated, and not only in the first NMD flight test, but in other programs. They are a marvelous technical accomplishment.

However, the system, as defined, has fundamental vulnerabilities, as described in my article with Hans Bethe in the March, 1968, Scientific American. This vulnerability is due to the prospect that the nation launching the ICBM will employ "penetration aids" or countermeasures, in order to reduce the prospect that the ICBM could be intercepted before it reaches its target.

Two such countermeasures are particularly appropriate to the mid-course hit-to-kill national missile defense program defined by the Clinton Administration. The first is the use of bomblets for delivering biological warfare agent ("BW"-- perhaps anthrax) instead of a single unitary charge of anthrax. Rather than a payload of 500 kg of anthrax with a reentry vehicle, which would be disseminated at ground level after impact, the military effectiveness would be increased by delivering even half as much agent in 100-200 "bomblets", each equipped with its own reentry heat shield. These would be distributed from the missile just at the end of powered flight-- perhaps 250 seconds into the 30-minute trajectory of a strategic ballistic missile. As the engine burns out, the bomblets are given velocities ranging up to some three meters per second, so that in the remaining 1500 seconds or so they will go as far as 5 km from the impact point of a single RV, thus delivering a lethal dose more uniformly throughout a large metropolitan area. Calculations done for the 11-author group of a Countermeasures report2 indicate that on the order of 100,000 people would die from such an attack on Washington, DC, in comparison with an estimated 50,000 or so from a first-generation nuclear weapon. A detailed assessment of the design of bomblets, using the M-143 BW delivery bomblets developed by the United States in the mid-1960s (before President Nixon's Executive Order banning BW and the ensuing Biological Warfare Convention, which was also the initiative of the United States) is found in the Countermeasures report.

A second penetration aid appropriate for a nuclear weapon is to use appropriate decoys to far exceed the available number of interceptors. The United States and other countries with ballistic missiles have long had decoys, but they are usually costly replicas that look as much like the RV as possible, both to radar and in the visible and infrared. Much simpler would be an "antisimulation decoy"3. Rather than face the difficult task of preparing replica decoys, this would use the simplest possible decoy in the form of a roughly spherical inflated mylar balloon, with an aluminum foil skin; the actual RV would be enclosed by a similar inflated balloon.

Neither radar nor optical or infrared observation could see the interior, and the appearance of the balloon and its motion can be arranged to be entirely similar to radar observation. The infrared homing device of the hit-to-kill seeker (or any other infrared viewer) would see exactly the same signal from the decoy balloons if each contained a 0.5 kg battery that would radiate as much heat to the balloon, and hence into space, as does the warm RV.

Our Countermeasures Report did not say that any kind of countermeasures could overcome every kind of missile defense. We simply pointed out that the particular missile defense under development was peculiarly susceptible and would have zero effectiveness, according to our analysis, to these countermeasures. In the Fall of 2000, an Independent Review Team under retired General Larry D. Welch submitted its report to the Ballistic Missile Defense Organization (BMDO) apparently identifying the need for a realistic study of countermeasures and the response to them.

All involved agree that the multiple BW bomblets could not be intercepted by the system. However, the BMDO head, LGen Ronald T. Kadish, testifying to Congress in September, 2000, discussed at length bomblets containing chemical agents, which indeed would be ineffective if delivered by an ICBM. But the real threat, as briefed to LGen Kadish, is an attack with bomblets containing BW, not CW, and it does not help the ability of the U.S. Congress to assess the effectiveness and need for a missile defense system if the principal threat is ignored.

Despite its ineffectiveness, I believe that the Clinton Administration defense would provoke substantial increases in nuclear weaponry on the part of China and would stall arms reductions by Russia. Why is this so, if those countries could easily defeat the system by countermeasures? They would not believe that the United States would spend such funds (on the order of $60 billion) to deploy an ineffective system, and of course the government would not publicize its ineffectiveness. We have some experience in this regard, in the use of the Patriot anti-aircraft system against extended-range Scud missiles launched by Iraq against Israel and Saudia Arabia during the 1991 Gulf War. President Bush, at that time, declared that Patriot had defeated 41 Scud warheads out of 41 launches, and the contractor (Raytheon) and the U.S. Army long maintained that the Patriot was highly effective in those engagements as a missile defense.

A brilliant analysis by Professor Ted Postol of MIT, with his colleagues, showed that commercial broadcast videotapes could be interpreted to indicate that none of those intercept attempts resulted in the destruction of the warhead. Nevertheless, a bitter attack was maintained against Postol. In January, 2001, Secretary of Defense William Cohen finally stated flatly about Patriot's performance during the Gulf War, "It didn't work". And responsible Israeli officials agree that either zero warheads were intercepted or possibly a single one. So this is a case in which everything possible was done to delude the American public for years after the Gulf War to believe that the Patriot was effective, when it was actually not. And it would take a strong leader of another country to count on the ineffectiveness of Patriot under these circumstances.

Of course, the primary protection of the American people against North Korean ICBMs armed with nuclear weapons or BW is deterrence, as it is against Iraqi and Iranian missiles. There is also the possibility of preemptive strike, to destroy the missiles before they can be launched.

But an effective system not vulnerable to BW bomblets (or to midcourse decoys and penetration aids) can be mounted by ground-based interceptors deployed close enough to the launching sites (within about 1000 km) so that the large interceptor can strike the ICBM while it is still burning. If a 250-sec burn-time is interrupted at 240 sec, the warhead is likely to fall 5000 km short of its target. It will not reach the United States or even Canada. The missile booster is the size of an automobile rather than the size of a person. It is highly visible and the intercept is probably considerably simpler despite the fact that the missile is accelerating, while the mid-course intercept would be against an RV coasting in space. I urge the full analysis of this system, which would use the existing DSP satellites and large 15-ton boosters of technology like those under development for the mid-course NMD system. In my writings I have urged the deployment of such a system jointly with the Russians at a site south of Vladivostok, as well as on U.S. military cargo ships stationed in the Japan Basin hundreds of km off North Korea.

An eventual ICBM from Iraq-another state the size of North Korea, and each of them the size of the U.S. state of Mississippi-could be countered from a single interceptor farm in southeast Turkey. Iran, four times the area of Iraq, is more difficult and would require interceptors based in the Caspian Sea and the Gulf of Oman, or on other land north of Iran. I urge that the mid-course system be cancelled, and the program redirected to early deployment of a boost-phase system against North Korea. I judge that if BMDO focused on this, they could have a BPI system operational in 2005 or 2006 at a fraction of the cost of the mid-course NMD system. The spur to increasing missile numbers in China and Russia would be avoided by a boost-phase intercept system which would not be able to reach ICBMs launched from Chinese or Russian land-based launchers, from the deployment area against North Korea.

The ABM Treaty bans even boost-phase intercept of any strategic missile in flight trajectory (except in test, at declared test sites), but President Putin in June, 2000, offered to help the United States counter such long-range missiles launched from North Korea. Since the purpose of the ABM Treaty was to prevent the intercept of the strategic missiles of the two parties to the treaty, President Putin is presumably willing to modify or to interpret the ABM Treaty to permit intercept of strategic missiles of another state, so long as the defense is not effective against Russia.


Theater missile defense has been receiving the bulk of American investments. The Patriot interceptor of the Gulf War is being replaced by the Patriot Advanced Capability (PAC-3) interceptor. The pellet warhead of the Patriot is replaced by a hit-to-kill interceptor which intercepts within the atmosphere, has a range of some tens of km, and is to be supplemented by theater-wide defenses expected to be available in 2008 or so. There is an Army system, THAAD, and a Navy Theaterwide system, both using hit-to-kill interceptors, and are to make their intercepts in the descent phase of the attacking missile-but outside the atmosphere.

There is no reason not to develop TMD systems to reduce the threat to deployed forces and to allies. Until recently, the accuracy of theater ballistic missiles was so poor that they could not be directed to strike a specific building or aimpoint. However, that can change. For instance, it is reported by India that the recent test of the Agni-2 missile was "guided to impact," presumably using the U.S. Global Positioning System-GPS-used by the United States to guide its bombs and cruise missiles to their targets. In protecting specified targets, however, very short range interceptors do a good job, and missiles not headed toward valuable targets can be ignored.

There is a big difference between the damage caused by a nuclear or BW warhead in a city, and the damage which would be caused by a warhead landing in the countryside. In the United States the average population density is about 100 times smaller than that in the cities, so that protection of population centers would devalue an offensive force by a factor 100, even if there were no defense outside the population centers.


There was criticism of the Clinton Administration NMD project, even if it were effective, because it would not defend Europe. With the NATO charter stating that an attack on one member is to be regarded as an attack on all, it might be tempting to a rogue nation unable to strike the United State because of the defense, to launch a strategic missile against one of the European NATO allies. In this way, deployment of an effective NMD would increase the threat to Europe.

Furthermore, it does not take a missile of strategic range to threaten Europe from Iraq or Iran. And Europe is vulnerable to missiles of much shorter range launched from Libya, which has already launched two missiles against Italy. Europe has been comfortable with deterrence, perhaps as a way not to look at the real threat. In any case, there were also suggestions that U.S. NMD be extended to defend Europe, by basing interceptors closer to the European continent.

Last week, Russia presented to NATO a proposal to jointly investigate European defense against non-strategic missiles. Of course, a missile launched against Europe from North Korea would be a missile of strategic range, but it is still worth asking what would be involved in such an investigation and in the kind of defense that might emerge from it.

At least three aspects of joint study would seem to be in order. First, there is the analysis of the threat. Presumably Russia would add its intelligence and assessment to that of NATO, and one would in this way try to estimate the threat from these nations. This assessment would be assisted by the unclassified and perhaps classified U.S. National Intelligence Estimates (NIE) which despite their orientation toward threats to the United States, could be helpful. Sharing of intelligence would pose severe security problems and might be a stumbling block.

A second phase of the study might be done concurrently, and that could be a review of the types of defense that might be available. This could include the S-300 Russian system, the PAC-3, and other candidates. The detailed assessment of the performance of the systems against countermeasures would be a matter of great sensitivity.

Finally, there would need to be investigated how such a system would be implemented-- the command and control, the number of interceptors that would need to be deployed at which sites, and whether there would be total coverage of the European area or only of population centers.

I think that it would be useful for "Europe" to undertake such an assessment, and presumably within NATO. At present, arguments in Europe for and against U.S. NMD seem to have less substantive basis than does the choice of a breakfast cereal. But that differs little from the situation in the United States.

In fact, the current U.S. Administration has a similar problem in deciding what to do. In a perceptive article by The Wall Street Journal's chief diplomatic correspondent4, Carla Anne Robbins (The Wall Street Journal Europe, February 9-10, 2001), her first question is "Who is the enemy?" And her second question is "Who is to be protected?" And one might add, "How?" The Bush Administration will need to decide whether to deploy the Clinton Administration defense in Alaska, which will probably require until the year 2006 or 2007, or whether it should return that to a lower level of investigation, and instead structure and deploy a Boost-Phase Intercept system to counter strategic ballistic missiles from North Korea.

Another option is to understand that accidental or unauthorized launch from China or Russia is no more desired by those countries than it is by the United States, and one ought to determine the cost and the feasibility of cooperative methods to enable those countries to enhance their control over their nuclear weaponry. This would benefit all possible targets of unintended or accidental launch, as well as those countries themselves.

Then if the remaining targets of the U.S. NMD systems are the three rogue nations, an adaptive, progressive deployment of Boost-Phase Intercept would seem to be appropriate.

The defense of Europe should be a judgment of Europe writ large, and of individual nations. It may be that a terminal defense deployed across the continent would not be desired in every country, considering its cost. Some countries might opt to protect their capitals and population centers, while others will want to protect every square kilometer of territory.

The important thing is that such studies be done promptly and not used as an excuse to delay indefinitely an understanding and a decision on defense against ballistic missiles. This decision could perfectly well be negative-- the hazard is not worth the cost. But it might be positive. The Bush Administration might choose first to deploy 20 interceptors in North Dakota, to give a capability to defend almost the entire United States against ICBMs that might emerge from North Korea, Iraq, or Iran. This could be done relatively promptly, although the long pole in the tent is the perfection of the three-stage booster that would be used for the interceptor. And this would be a defense in name only, if the missiles were equipped with BW bomblets or effective mid-course countermeasures.

As Niels Bohr is reputed to have said, "It is very difficult to predict-especially the future." But it will be interesting.



  1. "Report of the Commission to Assess the Long Range Missile Threat to the United States," Donald H. Rumsfeld, chairman. R.L. Garwin was one of the 9 commission members.
  2. "Countermeasures," A Technical Evaluation of the Operational Effectiveness of the Planned U.S. National Missile Defense System, UCS-MIT Study, A.M. Sessler (Chair of the Study Group), J.M. Cornwall, R. Dietz, S.A. Fetter, S. Frankel, R.L. Garwin, K. Gottfried, L. Gronlund, G.N. Lewis, T.A. Postol, and D.C. Wright, April 2000.
  3. "Midgetman Needs Anti-Simulation Decoys," paper by E. Teller on p. 44 of Armed Forces Journal International, March 1987.
  4. "In Missile Defense, Question No. 1 Is Whose Missile?", by Carla Anne Robbins, THE WALL STREET JOURNAL EUROPE, February 9-10, 2001