Ronald Siegel, Institute for Defense & Disarmament Studies, September 2001.

Introduction

Recent proposals for a National Missile Defense have stressed the emergence of an ICBM threat to the United States from North Korea, Iraq, and Iran. In addition to these countries, China, India, and Pakistan have long-range ballistic missile development programs that bear on the issue. A National Missile Defense (NMD) against North Korea will almost necessarily, perhaps even intentionally, provide a defense against current Chinese strategic forces. North Korea supplies missiles to Iran and Pakistan and has been assisting in their development of longer-range missiles. India has an active ballistic missile development program, and it is engaged in arms competitions with Pakistan and, to a lesser extent, China. While North Korea, Iraq and Iran do not yet have nuclear weapons, they have programs under way to develop nuclear weapons. China, India, and Pakistan already have nuclear weapons.

US government officials have argued that a defense against long-range missiles fielded by countries like North Korea and Iraq is more urgently needed than was, say, a defense against the former Soviet Union. First, these countries have been characterized as having erratic, adventurous, risk-prone leaders who might not be as reliably deterred by the threat of retaliation as other national leaders. Second, these leaders may be capable of using weapons of mass destruction to prevent the collapse of their regimes, and, given their adventurousness, the United States might seek to bring about such a collapse. Finally, defense against them is easier than defense against the former Soviet Union would have been, since the latter possessed thousands more nuclear weapons.

The 1998 Report of the Commission to Assess the Ballistic Missile Threat to the United States (the Rumsfeld Report) and the concurrent National Intelligence Estimate suggest that Iran and North Korea could deploy ICBMs as early as 2005. These estimates have been used as justifications for, among other things, a crash program to develop a National Missile Defense (with reduced testing before a production decision) and early withdrawal from the ABM Treaty. The timetable rests on two assumptions: first, that an ICBM may be considered operational immediately after its first successful test, even though typical development programs involve 20 or more tests over a period of three to five years; and second, that an ICBM capable of delivering a nuclear warhead to the US mainland can be derived from missiles currently under development. Each of these assumptions is questionable.

The missile program of Iraq is currently stalled. That of Iran relies mainly on North Korean technology: Iran tests missiles imported from North Korea. As of 1999–when it suspended missile tests–North Korea had tested at a range of 1600 km a missile which is estimated to have a maximum range of up to 6000 km, that is, long enough to reach Alaska but not Hawaii. The entire threat assessment justifying the crash NMD program seems to be based on the possibility that within a couple of years after ending the moratorium, North Korea will not only successfully test this missile at a range of 6000 km, but develop a version that can reach the continental United States (a distance of 8000-12,000 km). A modest extension to reach Hawaii (7000 km) might be possible within a few years. (The planned Alaska-based NMD would not provide a defense against missiles on a trajectory from North Korea to Hawaii.) Extending the range of the missile to reach the continental United States would reduce the payload well below the minimum needed to carry a first-generation nuclear warhead. Therefore, any threat of a North Korean ICBM attack on the continental United States will require an entire development cycle–for missiles, or for warheads, or for both.

An unimpeded North Korean development program could result in the development of a North Korean ICBM no sooner than 2010. Given their current limited programs, Iran and Iraq are unlikely to be able to develop ICBMs before 2015. The potential dates for all three countries could be further delayed by various contingencies discussed below.

Geography

Table 1 shows flight distances (in kilometers) from various launch sites to various target cities. The launch sites do not represent actual (present or future) bases, but simply the part of a country’s territory that is geographically nearest to the target city. Actual launch sites might well be several hundred kilometers farther away.

North Korea would need an ICBM with a range of at least 8400 km to reach San Francisco. In fact, all targets in the continental United States are at least ICBM range from potential adversaries, while nearly all non-US targets are within IRBM (Intermediate-Range Ballistic Missile) range of all of the countries considered here. India, for example, could reach all of the non-US targets shown in Table 1 with IRBMs; but it would need a missile with a range of at least 11,000 km for continental US targets. Testing an ICBM can mean only one of two things: an intent to launch satellites or an intent to target the United States. Testing an ICBM with a re-entry system can mean only one thing.

Table 1. Flight Distances (km)

Launch Site Target	N Korea	China (Man- churia)		China (Sink- iang)		India (Kashmir)	Pakistan (N)	Iran (NW)	Iraq (N)
New York	10,400	9,800		9,900		11,100	10,800	9,100	9,200
Chicago	9,900	9,300		—-		11,300	11,000	9,800	9,900
San Francisco	8,400	7,800		—-		11,500	11,500	11,400	11,500
An- chorage	5,400	4,800		—-		8,300	8,400	8.800	9,100
Honolulu	7,000	6,600		—-		11,200	11,500	12,900	13,200
Tokyo	1,200	1,300		—-		5,300	5,800	8,100	8,300
Beijing	—-	—-		—-		3,200	3,700	6,000	6,300
Moscow	—-	—-		3,400		3,900	3,400	1,900	2,100
London	8,500	—-		5,800		6,300	5,800	3,600	3,600

Rocket Science

A ballistic missile works by burning propellant and ejecting the hot gases through a nozzle, typically at a velocity of around 2500 meters per second (m/sec). The thrust from the exhaust causes the missile to accelerate. A given thrust will cause progressively higher accelerations as the missile lightens due to the consumption of its propellant. All of the propellant is consumed in the first few minutes of flight, following which the missile coasts above the atmosphere at a speed of several kilometers per second to its target. In an idealized case, the burnout velocity would be equal to the exhaust velocity times the natural logarithm of the ratio of gross missile weight to the payload. In real life, the missile will need additional impulse to reach a given velocity. Account must be taken of: the structural weight of the missile (typically discarded in several stages during boost phase); air resistance during boost; and gravity during boost. Still, the idealized relationship is useful: it provides an optimistic upper bound on what can be achieved when parameters are varied. At short ranges, the range of the missile will go as the square of its burnout velocity. Due to the curvature of the earth, at longer ranges the range will increase more rapidly. Table 2 shows the burnout velocity needed to reach various ranges, together with the payload fractions associated with missiles that attain any given velocity.

For example, a 6000-km range missile would need a 6200 m/sec burnout velocity and could achieve this while devoting 2-5% of its gross weight to payload; a 10,000-km range missile would need 7200 m/sec and could devote 1.3-3.5% of its weight to payload, about two-thirds as much. Thus, the table can be used to scale payload fractions as the missile range is varied. The scaling indicated in Table 2 is probably a bit optimistic from the missile designer’s perspective. The original missile design is optimized to produce the best possible distribution of propellant and structural weight among the stages. Adding a new upper stage (adapted from a different missile) or offloading payload will not necessarily yield an optimal mix. Thus, the payload penalty for increasing the range could be greater than the table indicates.

The concept of payload merits discussion because definitions can vary widely. Consider the weight remaining when the missile reaches burnout velocity. It includes the empty weight of the burned-out final stage. If this is deducted, what remains is the throw-weight or payload (shown in Table 2 as a fraction of gross weight). This includes the MIRV bus (if any), the decoys (if any), the guidance system, and one or more re-entry vehicles (RVs). Typically, a single RV will account for two-thirds or more of the payload; but because of the need for a bus, multiple RVs will add up to only about half the payload. The RV consists of a nuclear warhead, a fuze, and a heat shield. The heat shield may account for about one-third of this weight. Thus, less than half the payload will commonly be available for the weight of a nuclear warhead.

Table 2. Burnout Velocity and Payload Fraction vs. Missile Range

Range (km)	Burnout Velocity (m/sec)	Payload Gross Weight
300	1700	.150 -.200
600	2400	.080 -.120
1000	3000	.070 -.110
1500	3600	.055 -.095
2000	4100	.040 -.080
3000	4800	.030 -.065
4000	5400	.025 -.060
6000	6200	.020 -.050
8000	6800	.016 -.042
10,000	7200	.013 -.035
12,000	7400	.011 -.030

The first US nuclear warheads weighed 4100-4500 kg. The likely weight of a first warhead produced by a proliferating country has been variously estimated at 450-1000 kg. The lower estimate was for the first effort of an advanced, industrialized country and the higher estimate for a third world country. The United States and the Soviet Union each needed six to eight years to reduce their warhead weights to 1000 kg. Existing North Korean ballistic missiles could carry a 500 kg nuclear warhead. ICBM-range derivatives of these missiles could carry only 200-300 kg. Thus, even assuming that North Korea’s first generation nuclear warhead is at the low end of the estimated range (450 kg), an ICBM derivative of the existing missiles could not lift the warhead.

A guidance system is needed to hit a predictable target. This functions only during the boost phase, correcting the flight path to adjust for various deviations. After burnout, the missile is unguided and any further deviations from course, such as those caused by winds during the re-entry phase, are not corrected by the guidance system. The ballistic missiles now in use by North Korea, Iraq, and Iran achieve CEPs (Circular Error, Probable, the radius of a circle within which one-half of the warheads can be expected to fall) of several kilometers. This corresponds to velocity errors on the order of 0.1%, which in turn could lead to CEPs as great as 40 kilometers at ICBM range. With such accuracy, a missile aimed at Los Angeles would run a significant risk of missing the entire metropolitan area. This would seem to preclude any very near-term threat. However, on the extended timetables suggested above (2010 for North Korea, 2015 for Iran and Iraq), guidance should not be a problem. There will be time to develop a new system. If adapted to an ICBM, strapdown systems coming into use in civil aviation could yield CEPs of a few kilometers at full range. Global Positioning System updates (even on the clear channel) could reduce guidance system errors to a level that is small compared to the re-entry error discussed below. Thus, while North Korea, Iraq, and Iran are not remotely prepared for ICBM guidance now, they should not be expected to have difficulty hitting large cities at ICBM range after 2010.

A final problem is re-entry. As the RV re-enters the atmosphere at a velocity of more than 7000 m/sec, it encounters tremendous drag and slows down, eventually striking the ground at somewhere between 200-3000 m/sec. While slowing down, the RV generates tremendous heat that must be removed or else the RV will burn up. The ICBMs that could be developed by North Korea, Iraq, or Iran would use a blunt, high-drag heat sink–essentially a dome of copper armor. As the RV decelerates, the heat sink warms up and transfers most of the heat to the air rushing past. Most

Table 3. Missile Development Programs

Sources: Duncan Lennox, Jane’s Strategic Weapon Systems (JSWS), Issue 35, London: Jane’s Information Group, August 2001; IISS, The Military Balance 2000-2001, Oxford: Oxford University Press, 2000 (for force levels) and The Military Balance 1999-2000, 1999, Table 53 (for missile characteristics); Rodney Jones and Mark McDonough, Tracking Nuclear Proliferation 1998 (TNP), Washington DC: Carnegie Endowment for International Peace, 1998.

of the deceleration occurs at high altitude where the air is thin. By the time the RV reaches the denser lower atmosphere, it is no longer traveling at tremendous speed. The heat-sink approach was used on all of the early US and Soviet ICBMs. It has three drawbacks: it is heavy; its low final velocity gives the wind a longer time to alter the RV’s course, degrading its accuracy; and its low final velocity makes it vulnerable to terminal interceptors. One consequence of relying on heat sinks is that no matter how good the guidance system is, the ICBM will be limited to a CEP of at least 500 meters at full range. Shorter range missiles often use pointed, low-drag heat sinks, but this approach is not available at ICBM re-entry velocities: the RV would burn up.

The Missile Programs of North Korea, Iraq, and Iran

North Korea, Iraq, and Iran have had ballistic missile programs underway since the 1980s. Table 3 summarizes the characteristics of all but their shortest-range ballistic missiles. All missiles from these countries rely on liquid-fuel rockets based on Scuds (developed in the USSR in the 1950s) or on North Korean derivatives of Scuds. By contrast, Pakistan, India, and China all have active solid-fuel programs underway. Table 4 summarizes the flow of technology among missile derivatives. For example, the North Korean NoDong 2 has given rise to the Iranian Shahab 3 and the Pakistani Ghauri. There are reports that the Iranians are financially supporting the North Korean missile program in return for technology transfer.

Table 4. Missile Derivatives

Iraq	Russia	Iran	N. Korea	Pakistan	China
	®	®	®¯
AlHussein ¬ ¯	Scud	® Shahab 2	HwaSong 6 ¯	Hatf 3 ¬	M-11
AlAbbas ¯			NoDong 1 ¯
AlAabed		Shahab 3 ¬	NoDong 2 ¯	® Ghauri 2 ¯
			TaepoDong 1 ¯	Ghauri 2
	SS-4	® Shahab 4	TaepoDong 2

Note: The table shows only international transfers of missile technology and derivative missiles. Russian and North Korean Scuds have been exported to nearly thirty countries.

In the following country evaluations, it should be remembered that the projections are based on unimpeded development programs. The United States and North Korea have been negotiating to end North Korea’s missile program and missile exports. An agreed ban on all long-range missile missile-related technology and missiles over the limit prescribed by the Missile Technology Control Regime (MTCR) would set back Iran’s program. In the absence of such an agreement, many things could impede these programs before an operational capability to launch a nuclear attack on the United States is achieved:

1. The sponsoring countries could reassess their priorities.
2. Their current governments could collapse.
3. Their economies could collapse, leading to budgetary shortfalls.
4. They could lose sources of external financing.
5. They could lose sources of external technology transfer.
6. They could lose sources of external technology transfer.
7. Their programs could hit insuperable technical snags.

North Korea

During the 1980s, North Korea began both nuclear weapon and ballistic missile development programs. During the 1990s, the North Korean economy collapsed and its weapon programs embroiled the region in crisis. The North Korean nuclear program began with research reactors provided by the former Soviet Union, in return for which North Korea acceded to the Nuclear Non-Proliferation Treaty. However, when it came time for verification in the early 1990s, North Korea refused to submit to the full panoply of International Atomic Energy Agency (IAEA) safeguards. At the time it was estimated that there might have been as much as 12 kilograms of plutonium distributed among various stages of its fuel cycle–enough for one or two first-generation weapons. North Korea allowed IAEA inspectors into part of a large, underground nuclear complex near Yongbyon, but kept the bulk of the site off limits. This triggered an extended crisis during 1992-1994. In the Framework Agreement that resolved the crisis, North Korea agreed to end construction of two small power reactors and to accept a limited number of full inspections. In return, the United States, South Korea, and Japan agreed to build two somewhat larger Light Water Reactors (LWRs), subject to IAEA inspection, for electricity production. Some of the agreed-upon inspections were to be deferred until the non-nuclear components of the LWRs were fully installed. Originally scheduled for 2003, the deadline has now slipped to at least 2005, perhaps 2008.

The LWR has been described as not suitable for producing weapon-grade plutonium, but this is not precisely correct. If operated in the best way for power production (fuel rods left in the reactor for three to five years), LWRs would produce a distinctly inferior (high Pu-240), though still useable, bomb ingredient. However, if operated in the best way for plutonium production (fuel rods left in the reactor for less than one year), the two reactors could produce enough high-quality plutonium for tens of nuclear weapons per year. Given reasonable allowances for startup, reactor residence time, plutonium reprocessing, and weapon fabrication, these weapons would not become available until three years after unsafeguarded operation began. With the currently expected delay, that would be 2008-2011.

The North Korean missile program also began in the early 1980s. After having gained experience with local production of the HwaSong 5 and 6 (based on the Scuds B and C), the North Koreans began development of the NoDong, a scaled-up Scud. With a 1300 km range, it can reach targets throughout South Korea and Japan. The NoDong has been exported (together with production technology) to Iran and Pakistan, providing the basis for Iran’s Shahab 3 and Pakistan’s Ghauri. The next North Korean development was the TaepoDong 1: this placed a Scud atop a NoDong, yielding a two-stage missile with a range of about 2000 km. This caused a sensation in August 1998 when a three-stage version was fired through Japanese airspace during a failed attempt to launch a satellite. The TaepoDong 2 consists of a new, larger first stage (employing 3 NoDong engines) and a NoDong used as the second stage. Its range has been estimated at 4000-6000 km. This is a theoretical estimate; the TaepoDong 2 has never been tested, and no TaepoDong has been tested beyond 1600 km. Since there are few significant targets between 1300 km (the NoDong 2 range) and 6000 km from North Korea, there has been speculation about North Korea’s intent in developing the TaepoDong: could it be an export product intended for Iran and Pakistan (which are developing the Shahab 6 and Ghauri 3, respectively), a diplomatic bargaining chip, or a technical stepping-stone to a longer range missile?

The authors of the US National Intelligence Estimate claimed that a three-stage variant of the TaepoDong 2 could deliver a payload of “several hundred kilograms” to the United States. What this means depends on what is meant by “payload” and what is meant by “several hundred.” The TaepoDong 2 is estimated to have a payload of 1000 kg (a fuzzy estimate for a missile that has never been tested). Of this, 750 kg is thought to be available for the RV, with perhaps 500 kg for the nuclear device itself–the ‘physics package.’ If the missile has a range of 6000 km, then increasing the range to 10,000 km would require a velocity increment of 800 m/sec (see Table 2): this would allow the TaepoDong 2 to carry a 300 kg nuclear device. If the missile has a 4000 km range, extending this to 10,000 km would require an 1800 m/sec velocity increment, which would allow it to carry only a 200 kg device. These payloads are both too small to accommodate first-generation nuclear warheads.

More realistically, if the TaepoDong 2 has a 4000 km range, it can reach Guam. If the full range is 6000 km, it can reach Anchorage, and the range could reasonably be extended to Honolulu as well. In contrast, an ICBM threat to the US mainland will have to await one of three developments: a larger missile, a missile with more efficient stages, or a second-generation (maybe even third-generation) nuclear warhead. Any of these would require a new development cycle lasting some years. For these reasons, we estimate that at worst the TaepoDong 2 could pose a threat to Alaska and Hawaii by 2005 and a follow-on ICBM could threaten the continental United States by 2010, assuming the availability of additional plutonium for use in ICBM warheads.
North Korean missile testing is currently subject to a moratorium, which runs until 2003. However, there is an option to circumvent this moratorium by surrogate testing. Test launches of a Shahab 6 from Iran or a Ghauri 3 from Pakistan (with North Korean technicians in attendance) could substitute for explicit tests of a TaepoDong. In the past, there has been only one North Korean test of the NoDong, but if Iranian and Pakistani tests are counted, there have been six.

Iraq

Iraq has made persistent efforts to conceal from UN inspectors (at great cost to Iraq) ballistic missile and chemical warfare inventories as well as missile, nuclear, chemical, and biological development programs. These programs have been suspended due to the combined effects of bombing, decommissioning by inspectors, and sanctions. However, the lengths to which Saddam has gone and the costs he has been willing to incur (including a decade of sanctions) to avoid disposing of Iraq’s last kernel of development capability suggest ominous future plans for these weapons. The Rumsfeld Report estimates that Iraq could pose an ICBM threat to the US within ten years of reviving its nuclear and missile programs. There is no reason to question this estimate.

Iraq’s missile program was based on variants of the Scud, which does not have much long-range potential. The Al Hussein and Al Abbas increased the Scud’s range to 600 km and 900 km, respectively. This was done simply by lengthening the Scud’s fuel tank and reducing its payload. The work was completed before the Gulf War when Iraq still had access to many foreign experts. The Al Hussein was the principal Iraqi missile in that war, in which its reliability and accuracy were shown to be awful. Finally, Iraq has the Al Aabed, a failed satellite launcher comprising five Scuds strapped together as a first stage, and one more Scud on top as a second stage.
Iraq has had an active nuclear weapon program that has been destroyed (twice). It has the potential to pose regional missile threats once its program emerges from suspended animation.

Iran

Iran has had a faltering nuclear development program going back to the days of the Shah. It is currently developing a complete nuclear fuel cycle at many locations throughout the country (this in a country which produces more natural gas than it can export).

Iran’s ballistic missile program has included the Shahab 2, a Scud copy, and the Shahab 3, a NoDong copy. North Korea has sold Iran missiles, technical information, and production equipment. Iranian missile developments have been tracking North Korean developments, about five years behind. The Shahab 4 was to have been a copy (or import) of the Russian SS-4, a 40-year old missile. This program may have been stillborn, but (if nothing else) the RD-214 rocket engine technology seems to have been transferred. Iran has also imported Chinese M-7 solid-fuel missiles, with a 150 km range. These could form the starting point for an eventual solid-fuel program. The Shahab 5 and 6 are developmental programs based on the TaepoDong. Since Iran is about 1000 km farther from US targets than North Korea, a TaepoDong-based ICBM would be even more of a stretch for the Iranians than for the North Koreans. Therefore, an Iranian ICBM threat to the United States is estimated no earlier than 2015.

Other Missile Programs

Pakistan

Though not historically hostile to the United States, Pakistan’s nuclear weapons, its internal instability, its chronic confrontations with India, and its cooperation with North Korea raise concerns. Pakistan has had a nuclear weapon program under way since the late 1970s, and it may have had a workable bomb prototype by the late 1980s. In May 1998, it conducted six tests, one of which may have been a boosted weapon (in which a small fusion reaction is induced to provide a source of neutrons to enhance the fission yield). Unlike the nuclear weapons that might be developed by North Korea, Iraq or Iran, Pakistan’s nuclear weapons cannot be assumed to be first-generation designs.

Pakistan’s missile development has followed two parallel tracks. Like Iran, Pakistan has been importing liquid-fuel technology from North Korea. At the same time, it has been importing solid-fuel technology from China. From North Korea, it obtained the Ghauri 1 (Hatf 5), a copy of the NoDong 2. With extended fuel tanks, this became the Ghauri 2 (Hatf 6). The Ghauri 3, under development, is based on the TaepoDong. With Chinese aid, Pakistan developed the Hatf 3, a longer-range, lighter-warhead variant of the M-11. Chinese solid-fuel technology has also figured in the Shaheen 1 and 2 (Hatf 4 and 7), MRBMs that may have a greater role in Pakistani force planning than the Ghauri.

Its ballistic missiles, with ranges up to 2500 or 3000 km, can reach any target in India, but they do not come close to being able to reach the United States. Since Pakistan is 2000 km farther from US targets than North Korea, even a three-stage TaepoDong 2 would not have the range to reach the continental United States. Even apart from the historical lack of animosity, there is no Pakistani ICBM threat to the United States on the horizon.

India

India tested a nuclear device in 1974. It then conducted five more tests in May 1998, including a boosted weapon. There is no history of confrontation with the United States, but the Indians have sometimes felt threatened, as when they attached great significance to the movements of a US aircraft carrier during the 1971 Indo-Pakistan War.

Indian missile development began with liquid-fuel rockets derived from the Russian SA-2. Recently India has favored indigenously-developed solid-fuel rockets, often using a liquid-fuel rocket for the final stage, because liquid-fuel permits more precise thrust termination and, thus, greater accuracy. The main Indian weapon is the Agni series of Medium-Range Ballistic Missiles (MRBMs), which will be able to reach all potential regional adversaries. The Agni 1 has been cancelled, the Agni 2 (tested in 1999) is in service, and the Agni 3 and 4 are under development. Beyond the 5000 km range of the Agni 4, there are no likely targets for Indian ballistic missiles outside the United States. Nevertheless, India has the Surya ICBM, with a 12,000 km range, under development. While this could be used to launch satellites (India has had an active space program employing other boosters), any testing of the Surya with a re-entry system would suggest the goal of posing an ICBM threat to the United States.

China

China became a nuclear power in 1964 and has deployed strategic nuclear forces for several decades. Without treating China’s overall nuclear posture, this section considers two issues: China’s role as a supplier of missile technology, and the possibility that a National Missile Defense aimed at North Korea will inadvertently, or perhaps deliberately, make China’s strategic deterrent vulnerable.

China developed the M-series of solid-fuel short-range ballistic missiles for export in addition to domestic use. These are the M-7 with a 150 km range, the M-11 (DF-11) with a 300 km range, and the M-9 (DF-15) with a 600 km range. Recently, Global Positioning System (GPS, a satellite-based navigation aid) receivers or radar terminal guidance have been added to some versions of these missiles, offering the prospect of greatly improved accuracy (not reflected in the CEP figures shown in Table 3). With China’s adherence to the Missile Technology Control Regime, it has stopped marketing the 600 km range M-9; but it might have already transferred some technical information regarding the M-9 (and even some missiles). M-series missiles have been supplied to Iran and Pakistan (and possibly Syria, although that deal may have been cancelled). Earlier, China supplied the DF-3, an older liquid-fuel MRBM, to Saudi Arabia; and China has helped Pakistan develop a plutonium production reactor.

The Chinese deterrent against the United States consists of 18 liquid-fuel DF-5 ICBMs deployed in the early 1980s. Since these would approach the United States on approximately the same flight path as would North Korean missiles, a workable National Missile Defense with 100 ground-based interceptors in Alaska could engage all of the Chinese strategic missiles. This has been a source of Chinese concern.

There are several reasons why China’s concern may be overstated. First, a workable NMD, particularly one that seeks non-nuclear intercepts of high mid-course targets, is an impossibly daunting technical challenge for reasons discussed below. Second, most of the Chinese missiles will launch from too far back from the Pacific coast to allow radar tracking by the United States during boost phase, depriving the defense of one discrimination tool. Third, by the time the United States can deploy a functioning missile defense system, the DF-5s will be approaching the end of their service lives and will be being replaced by DF-41s. The DF-41, a solid-fuel ICBM, will complete its boost phase at a lower altitude, making it harder to track by radar. It may be MIRVed (at some point, even if not initially), it may carry more decoys than the DF-5, and it may be deployed in larger numbers than the DF-5. Moreover, it will be supplemented by the JL-2 SLBM (derived from the DF-31), which could approach the US on a different flight path, avoiding Alaska-based interceptors. For these reasons, a modest NMD program directed against countries with first-generation ICBMs need not pose much of a strategic challenge to China.

Interactions with National Missile Defense

This is not the place for a thorough treatment of national missile defense. However, a few points of interaction with potential first-generation ICBM programs should be noted. The Clinton program would have deployed interceptors to Alaska, with the later addition of interceptors in North Dakota. The Alaska defense would have been useful only against North Korea and the older Chinese forces; any defense against Iran or Iraq would have to await the North Dakota deployment (and even that would have been pressed to defend the east coast). Neither of these defense sites would have protected Hawaii. Trajectories originating in North Korea and aimed at 49 of the states would all pass within 1000 km of the Alaska site: those aimed at Hawaii would not pass any closer than 3000 km.

Since the proposed NMD is a midcourse defense, it suffers from the fundamental weakness of all such systems–decoy discrimination. An attacker could provide dozens of decoys for each warhead, exhausting the defender’s interceptors unless there were an effective means of telling them apart. Given the possibilities for making decoys that look like warheads (and warheads that look like decoys), this problem has been viewed as virtually insurmountable. Two aspects of the proposed system make the problem even worse. First, the Alaska system will make its intercepts near the mid-point of the ICBM’s flight, when the warheads are at an altitude of 1000 km or more. Some discrimination techniques (meant for late midcourse use) that may be marginally useful at 200-300 km, when the warheads and decoys are beginning to encounter traces of atmosphere, will be hopeless at 1000 km. Second, the interceptors are non-nuclear and have a lethal radius of no more than a few meters. This means that even if two objects are only 50 meters apart, the system must examine and classify both of them, for a non-nuclear intercept of the decoy will not damage the warhead. If the target is a city, any decoy within a few kilometers of the warhead will appear to be on a credible trajectory. Therefore the number of decoys that can be usefully deployed is limited only by the throw-weight devoted to them (a decoy can be 100-1000 times lighter than a first-generation RV). With spacing of a few tens of meters, various objects needing classification may be closer together than the range resolution of the radar, aggravating the tracking problem immensely. For these reasons, a high midcourse defense using non-nuclear interceptors must be viewed as technically very improbable.

Bush has proposed two changes to the Clinton plan. First, the administration is planning to compress the testing schedule and prematurely start construction at prohibited sites, requiring near-term withdrawal from the ABM Treaty. Article XV of the ABM Treaty allows any party to withdraw on six months notice if “extraordinary events” jeopardize its “supreme interests.” It seems unlikely that the nascent programs of North Korea, Iraq, and Iran would rise to the standard of the treaty language. Notwithstanding this, recent administration commentary suggests that, failing Russian acceptance of an amendment to the treaty, the United States could withdraw soon.

Second, the administration may add a boost-phase component to the NMD program. This could be based on high-speed interceptors, becoming operational perhaps three years later than the midcourse system or, further in the future, on airborne lasers. Either way, the boost-phase system would have an effective range of no more than a few hundred kilometers and would operate on a stringent timeline. If any of the missile developers replaced their liquid-fuel rockets with solid-fuel rockets (not a near-term prospect), then due to the shorter burn times of solid-fuel rockets and the stringent timeline, a previously workable system might no longer be able to meet its timeline. Finally, a boost phase intercept of a North Korean ICBM would probably take place in Russian airspace.

There is a disturbing possibility behind all of these manifestly unworkable NMD concepts. There is a workable NMD scheme, not against strategic peers, but against the countries considered here. If the ICBMs can be tracked by radar during boost phase, if the attacker does not have MIRVs, and if the defender uses nuclear interceptors, then a defense is possible. This is because the combination of radar tracking and no MIRVs (more precisely, no post-boost vehicle) will allow the velocity of the warhead to be measured to less than a meter per second. This will permit the warhead to be localized to within a few kilometers during midcourse. With a high-yield nuclear warhead, this will be good enough for a kill, notwithstanding any decoys. Most of the components of the proposed non-nuclear NMD would find a place in a hypothetical nuclear system. Even the hit-to-kill interceptors could be recycled as antisatellite weapons. (No hard evidence for this yet. Just something to ponder.)

Alternative Means of Delivery

A state that wants to attack the United States with only a few nuclear weapons need not do so with ICBMs. There are various stealthy ways in which a nuclear weapon could be placed on or near US territory. A short-range missile could be fired from a ‘merchant’ ship offshore. This possibility was considered in the Rumsfeld Report and has been brought up in discussions to emphasize how soon the threat could arrive. A surprisingly large fraction of US cities are within range of such an attack (see Table 5) and the current NMD proposals would offer no defense against it. Potential adversaries would need to obtain single-stage solid-fuel missiles because stealthy deployment of a liquid-fuel missile and its supporting paraphernalia on a disguised merchant ship is not practical. Neither North Korea nor Iraq has solid-fuel missiles.

Table 5. US Urbanized Areas with Populations over One Million

< 50 km from coast (11 cities)	50-200 km from coast (10 cities)	200-500 km from coast (5 cities)	> 500 km from coast (10 cities
Boston	Philadelphia	Pittsburgh	Cleveland
New York	Baltimore	Atlanta	Columbus
Norfolk	Washington	Dallas	Cincinnati
Orlando	New Orleans	Phoenix	Detroit
Ft Lauderdale	Houston	Las Vegas	Chicago
Miami	San Antonio		Milwaukee
Tampa	San Bernardino		Minneapolis
San Diego	Sacramento		St Louis
Los Angeles	Portland		Kansas City
San Jose	Seattle		Denver
San Francisco

Also, a cruise missile could be used. Since commercially available cruise missiles cannot lift a first-generation warhead (Shaddock-type missiles are not being exported), an adversary would have to build one, perhaps by adapting a general aviation aircraft (at least the plane might come supplied with a first-class navigation system).

If a potential adversary simply wants a revenge weapon and is willing to forego the general deterrence that derives from a recognized capability in the standing armed forces, a few more options open up. Cities could be mined–nuclear devices could be emplaced in a warehouse in peacetime and detonated by remote control in wartime. Nuclear devices could be delivered overland by truck (probably not by car). Or, returning to cargo ships, one could dispense with the missile entirely: the nuclear device could be in the bottom-most freight container; the ship could dock, the crew could set a timer, and take a taxi to the airport. Alternatively, an airliner could be used, perhaps marked with someone else’s insignia. A suicide crew could be used or the crew could set the autopilot and bail out an hour before the target. All of these options are practical only for very small numbers of weapons; unlike standing forces, none would provide any political benefit prior to their use. These options would, however, hold at least the possibility of an anonymous attack and, therefore, some chance of escaping retaliation.

Conclusions

From the foregoing considerations, it appears that the forecasts of an imminent ICBM threat from North Korea, Iraq, or Iran directed at the continental United States are unfounded. At the earliest, North Korea could pose a minimal threat to Alaska and Hawaii by 2005 and a more substantial threat to the United States by 2010. Iraq and Iran could pose substantial threats by 2015. These timetables assume that nothing happens along the way to stop or delay ICBM development. In order to pose even a minimal threat, a government must be willing to risk national annihilation in order to deliver a few low yield nuclear warheads to US cities, killing perhaps a few tens of thousands of civilians. In order to pose a more substantial threat, a state would have to undertake and complete a great deal more development than any has done so far.

The lack of an impending threat undermines the justification both for early withdrawal from the ABM Treaty and for committing to the acquisition of a troubled system before there has been thorough and realistic testing. If the more substantial threats eventually materialize, the currently proposed NMD systems will be less than completely relevant to the problem.

Acronyms

ABM Anti-Ballistic Missile
CEP Circular Error, Probable
GPS Global Positioning System
IAEA International Atomic Energy Agency
ICBM Intercontinental Ballistic Missile (range: > 5500 km)
IOC Initial Operational Capability
IRBM Intermediate Range Ballistic Missile (range: 3000-5500 km)
LWR Light Water Reactor
MIRV Multiple Independent Re-entry Vehicle
MRBM Medium Range Ballistic Missile (range: 1000-3000 km)
MTCR Missile Technology Control Regime
NMD National Missile Defense
RV Re-entry Vehicle
SLBM Submarine Launched Ballistic Missile
SRBM Short Range Ballistic Missile (range: < 1000 km)

Top