Dr. Kathleen C. Bailey
US Senate Armed Services Committee
Subcommittee on Strategic Forces
March 31, 1998
Mr. Chairman, I am pleased to appear before you and the members of the Subcommittee today to address key considerations for shaping the US nuclear deterrent in the post-Cold War world. The views I express are my own, and not necessarily those of any institution.
Because the Cold War has ended, pressures for rapid, irreversible reductions in nuclear weapons have increased. With the negotiation of START II long finished, but not yet ratified by Russia, arms control advocates are unhappy with the slow pace of disarmament. They are pressing not only to make deep cuts beyond 2000-2500 accountable warheads-the limits outlined for START III at Helsinki-but also to de-alert our nuclear forces.
Today I wish to make three points that urge caution in making further deep reductions in our nuclear forces or in making our nuclear deterrent less effective:
· Extant and emerging nuclear, chemical, and biological threats require an effective US nuclear deterrent.
· Russia has a large, functioning nuclear weapons production complex; the United States does not. And, we have no way to verify that there are no undeclared stockpiles in Russia.
· De-alerting would profoundly undermine deterrence, would generate serious instabilities, and, in some cases, introduce safety uncertainties.
Although our relations with Russia are now relaxed, we must continue to take Russia's weapons capabilities, not just its intent, into account. Since the end of the Cold War, there have been important indicators that Russia is increasing its reliance on nuclear deterrence and is improving its nuclear arsenal and delivery systems. In May 1997 Russia announced that it would no longer adhere to a nuclear no-first-use doctrine. It continues to maintain several thousand tactical nuclear weapons at a time when NATO has made dramatic reductions in this class of nuclear forces. Giant, deeply buried, underground nuclear-related complexes in Yamantau and Kos'vinskiy Mountains, as well as shelters, continue to be constructed at great cost. Russia conducted some sort of nuclear or nuclear-related tests at Novaya Zemlya in 1996, and perhaps others since. Russia is not only remanufacturing existing warheads, but also building new designs. Initial units of the SS-X-27, a highly accurate and reliable mobile ICBM, have already been deployed. A new SSBN and SLBM are under development.
China has not been a highly salient threat to the United States, principally because China had few warheads and delivery systems relative to those of the United States. However, the relative threat presented by the Chinese arsenal is increasing. China is estimated to have 300-400 nuclear warheads and is making progress on miniaturization and reliability. The threat from its current arsenal is mostly on missiles with ranges that cannot reach the US mainland, but it also has some silo-based 12,000-km DF-5 missiles that can. China is modernizing its ballistic missiles. In this, it is reported to have received help from Russian scientists as well as from Ukraine, which has SS-25 mobile missile technology. China has been developing the mobile DF-31 ICBM missile with a range of about 8000 km.
Proliferators also could present increasingly significant threats to the United States and its allies in the future. In force planning, we should take their military capabilities into consideration.
India is an emerging secondary nuclear power whose new government advocates making India a declared nuclear weapons state. India has had nuclear explosives capabilities since its 1974 test and has steadily been producing fissile materials since. Some estimates say that India could have more than 200 nuclear warheads. India has also developed impressive ballistic missile delivery capability; its Agni missile can carry a 1000 kg payload to a range of 2500 km. India successfully tested a low-earth and a polar satellite launch vehicle, the latter of which could be used as an ICBM with a 1000 kg payload.
Iraq surprised the United States with the advanced state of its nuclear weapons program and its intermediate range Scud missile derivatives. Iraqi scientists were well on the way to having a workable nuclear weapons design. Although the United Nations has destroyed much of Iraq's infrastructure for making weapons of mass destruction and longer range missiles, Iraq has preserved much of its technology has continued to acquire and stockpile key items.
North Korea poses another potential threat, particularly since there is a US-South Korea defense treaty and 32,000 Americans are stationed in the South. North Korea secretly separated plutonium for nuclear weapons, and still retains that fissile material, thus remaining in noncompliance with the Nuclear Non-Proliferation Treaty. Furthermore, North Korea may have undeclared plutonium production facilities out of view in one of its myriad tunnels or underground facilities. North Korea also has nuclear-capable Scud missiles that can reach South Korea, and has developed a missile based on the Scud, the Nodong 1, a missile with a range of 1000 to 1300 km. North Korea is developing longer range missiles, the Taepodong 1 and Taepodong 2. The former is estimated to have a range of 2000 km and the latter has been estimated to have a range up to 10,000 km, making it capable of reaching the United States, and could be operational by the year 2000.
In addition to nuclear threats, there are chemical and biological weapons programs worldwide. Moscow has broken its political pledges to discontinue its chemical and biological weapons programs. While the United States ceased its biological weapons program in 1972, Moscow continued its own program unabated, as was admitted by President Yeltsin in 1992. In the early 1990's, Vladimir Pasechnik, a senior biologist who defected to Britain in 1989, provided convincing evidence that extremely potent biological weapons were developed, including a plague bacterium that is impervious to Western drug treatments. Similarly, a Russian dissident revealed that Moscow secretly continued its chemical weapons program, developing nerve agents that are more toxic than any substance known in the West. Against these weapons, the United States has no in-kind retaliatory capability. US biological weapons were destroyed and the program ended; the US chemical weapons modernization program was abandoned and the remaining arsenal is aging, increasingly undeliverable.
Chemical and biological weapons are much less expensive and technologically easier to develop and produce than nuclear weapons. Because they are virtually impossible to detect, we may not know all of the countries that could conceivably deliver such weapons against the United States. However, we do know that North Korea, Iraq, Iran, Libya, and others have chemical and/or biological weapons capabilities.
A key question facing the United States, which has forsworn both biological and chemical weapons, is how we will deter such threats. US policy is ambiguity, fostered by different officials saying different things. Former Secretary of Defense William Perry told Congress, "[I]f some nation were to attack the United States with chemical weapons, then they would have to fear the consequences of a response from any weapon in our inventory... We are forswearing ... chemical weapons ourselves. ... [No] nation should feel that they can use chemical weapons against us without receiving a devastating response. In every situation that I have seen so far, nuclear weapons would not be required for response.
That is, we could make a devastating response without the use of nuclear weapons, but we would not forswear that possibility." (emphasis added) This quotation was read by Undersecretary of Defense Slocombe in testimony before this Committee this month, indicating that we still reserve the nuclear option against chem-bio threats.
Other officials, however, are sending a different signal regarding the role of our nuclear deterrent. For example, Robert Bell, Special Assistant to the President for National Security Affairs, discussed publicly last month whether US policy is to use nuclear deterrence against chemical and biological threats. He said, "You could say, 'if you attack us in any fashion-conventional, chemical, biological-we will use nuclear weapons.' That would be a categorical threat that would maximize your deterrence. Unfortunately, it would derail your non-proliferation policy and your nonproliferation agenda. So we have tried to strike the balance by maintaining long-standing U.S. policy in this area, actually dating back to Secretary of State Vance in 1978." Bell also stated, in the same context, Presidential Decision Directive of November, 1997 reaffirms that "it is the policy of the United States, as restated in this PDD, not to use nuclear weapons first in a conflict unless the state attacking us or our allies or our military forces is nuclear-capable or not in good standing under the NPT or an equivalent regime, or third, is attacking us in alliance with a nuclear capability."
Deterrence works best when the potential aggressor understands that the retaliation will be swift and proportional. Because conventional responses may not be possible or proportional in some scenarios, the option to respond with nuclear weapons must be preserved and clearly communicated. If not, deterrence will be less effective.
Low Numbers, Abolition are Unverifiable
When planning force structure for the future, we must consider not only the likely threats that must be deterred, but also the break-out potential of other nations. Arms control that takes our nuclear arsenal to low numbers or to zero greatly increases the incentives for others to cheat because opponents can garner tremendous advantage over the United States with only a small nuclear force. This applies not only to Russia and China, but to India, North Korea-any nation with nuclear capabilities. Thus, before further steps are taken to cut US and Russian arsenals, we should consider the fact that verifying low levels, or zero, is impossible in Russia, and perhaps elsewhere.
The United States does not know how many nuclear weapons or warheads the Soviets built, nor the size of Russia's current stockpile. If Russia were to hide some of its stockpile and declare a lesser number, there are no existing technical means that would enable us to detect the discrepancy or locate the hidden nuclear warheads. The wide range of error possible in estimating Russian warhead inventories was highlighted in 1993, when Minatom director Viktor Mikhailov stated that the Russian arsenal peaked at 45,000 warheads in the mid 1980s. This was 12,000 more than generally believed in the West.
The problem of undeclared weapons stockpiles is complicated by the possibility that Russia claims to have been destroying thousands of nuclear warheads per year since the end of the Cold War. We cannot verify that these claims are true. Even if Russia did dismantle weapons, we cannot verify that it did not manufacture new ones using either recycled or new special nuclear materials.
Detection is also problematic with undeclared special nuclear materials. Current technology does not allow us to determine with confidence how much material the Soviet Union produced during the Cold War, or how much Russia has produced since. Even with highly intrusive inspections, it could be impossible to find materials not only because there is no way to pinpoint where to look, but also because materials could readily be transported secretly.
Estimates of materials stockpiles could be based on plutonium or highly enriched uranium production capability and operation records, but discrepancies would be difficult to resolve and uncertainties could be significant. For example, Russian plutonium production up to 1996 has been estimated to be 145 tonnes. A 20% error-25 tonnes-could reasonably be expected. This could correspond to primary fuel for as many as 5000 warheads.
There is tremendous asymmetry between US and Russian nuclear materials and warhead production infrastructures. Russia has a large, functional infrastructure with extensive redundancies. It is maintaining this complex, despite high financial cost. There are multiple plants in Russia-four of which are major plants the size of the US Pantex facility-that can produce thousands of warheads per year. This includes not only facilities for manufacturing fissile materials components and other weapons parts, but also facilities for final assembly. In short, Russia has a functioning capability that would allow it to rapidly reconstitute a nuclear force in a break-out scenario.
By comparison, the United States has no facilities or infrastructure to manufacture many key materials and components for weapons. For example, there presently is no US facility for producing plutonium pits. The US plan is to begin pit production at Los Alamos National Laboratory by 2003, but the capability will only about 20 pits per year compared to the Russian capacity of thousands per year. The US objective is to maintain the enduring stockpile, thus it will not have the capacity to undertake large-scale production should there be a Russian break out some years hence.
Undeclared Production Facilities
Secret plutonium production reactors and reprocessing facilities can be constructed underground or in a mountainside, with emissions eliminated or significantly minimized, and with no observable features to attract attention. Uranium enrichment plants can be hidden even more easily. The ease of hiding varies with the type of technology used. A 20,000 kg-separative-work-unit/year centrifuge plant would fit within a typical factory building and would consume a relatively small amount of electrical power (600 kW). The power consumption of a plant using laser isotope separation would be a factor of three smaller. Laser as well as chemical isotope enrichment processes can also be used to separate plutonium-239 from reprocessed spent reactor fuel. The technologies to produce fissile materials other than plutonium-239 and uranium-235 are even easier to hide.
The International Atomic Energy Agency (IAEA), which is responsible for assuring the use of nuclear technology for peaceful purposes only, has acknowledged that there currently are no technical tools enabling detection of clandestine weapons activities when they take place at undeclared facilities. The IAEA has noted that the problems of finding hidden plutonium reprocessing are greatly complicated in countries where openly acknowledged reprocessing has already occurred. This conclusion is echoed in the JASON Report of 1993, which stated that a determined and highly disciplined evader could undertake clandestine production of weapons or special nuclear materials without being detected by national technical means. Only real world lapses of discipline would leave traces of sizable activity that would be detectable.
Downsides of De-Alerting
De-alerting (and, in some cases, precipitous de-activation) has several serious ramifications for our nuclear deterrent and national security. Before outlining the reasons why, let me define de-alerting.
De-alerting measures are those which, if taken, would increase the amount of time between a decision to launch a nuclear attack and the actual launch. Some examples of proposed de-alerting measures are listed in Table 1. Dealerting should be distinguished from de-activation, which is defined as removal from the alert force of those weapons which have been slated for destruction under arms control agreements. For example, in 1991, President Bush removed from alert 450 Minuteman II ICBMs that were due to be eliminated under START I.
Table 1: De-Alerting
Remove and store warheads
Safing of missiles
Disabling silo covers
Keep submarines on modified alert
Remove launch codes from submarines
Proposed de-alerting measures have a number of significant detrimental effects, one of the most important being their impact on deterrence. Deterrence occurs when a potential aggressor is dissuaded by being convinced that his potential victim has both the capability and the will to retaliate with consequences that the aggressor views as unacceptable. The capability must be ready; if not, the opponent may believe that the retaliatory capability can be destroyed before it can be used. Thus, nuclear deterrent must be able to survive an attack and/or be launched quickly enough to avoid destruction in a first strike. De-alerting undermines deterrence by reducing both survivability and the ability to respond in a timely manner. Let me elaborate with six points.
First, a serious downside common to most, if not all, dealerting proposals is that they generate instabilities by making a first strike more attractive to an aggressor. Because de-alerting reduces readiness, by definition, any de-alerted forces that are not survivable make more inviting targets. The aggressor could undertake a first strike with forces that have not been de-alerted-either known, declared systems or clandestine ones-or could regenerate quickly in a break-out scenario. The incentive for an opponent to make such a pre-emptive strike would be particularly pronounced in times of tension, when he might observe US regeneration of de-alerted forces, or may simply fear that US regeneration may occur. The United States, knowing that its own regeneration of forces might be destabilizing, could hesitate to re-alert, thus undermining the nuclear deterrent.
De-alerting could also lead to a destabilizing "regeneration-of-arms race." Nations, worried about whether their adversary's forces can be regenerated more quickly and effectively than their own, would work to streamline the realert process. In turn, this could lead to cutting corners with safety and security procedures, and could create or escalate tensions. It could even make the risk of nuclear war more likely because once the regeneration race begins, neither side would know whether the other was likely to stop at the brink. The incentive to be the first to strike would be high.
A second, paradoxical problem is that, if one did not exercise to assure regeneration, deterrence and survivability would be decreased. The reason is obvious-the level of preparedness would be too low to regenerate swiftly and decisively in a crisis. Such operational practice would require extensive investment of resources. To have an idea of the extent of the effort, imagine what would be required to rearm SSNs with the nuclear SLCMs that were removed in 1991 and placed in central storage.
Third, de-alerting could adversely affect safety. Procedures to assure safety would need to be reassessed and perhaps redesigned. Also, the question of whether removal of warheads or parts affects the safety of the whole system must be addressed. Nuclear weapons systems are designed with extremely high standards for safety within specified parameters for operation. If those parameters are altered in any fashion, safety could be compromised. Certification for systems whose parts are separately stored could require expensive, time-consuming evaluation and may even require redesign.
Fourth, security also could be compromised by some dealerting measures. Having warheads separate from the missiles could make them more vulnerable to theft or sabotage. Also, in crises requiring rapid regeneration, hasty efforts to make weapons operable, in absence of having had extensive training and exercises in regeneration, could lead to errors resulting either in an accident or a situation in which the regenerated systems do not work properly.
A fifth problem is that most proposed de-alerting measures are either unverifiable or only verifiable with low confidence. Efforts to verify some of the measures would entail full-time presence of inspectors or observers. In a time of crisis or tension, these people would likely be viewed as "enemy agents" and their access would be curtailed. In turn, this would probably escalate the crisis.
Sixth, de-alerting would circumvent the arms control process. This is, of course, what some advocates want because they believe that disarmament is not moving quickly in the post-Cold War era. For example, some advocates have suggested that START III is likely to eliminate 4 more US submarines and reduce the number of warheads on each missile to 4. They argue that these steps should be taken immediately as a de-alerting measure. Such steps are dangerous because they are unilateral, unverifiable actions that affect force structure. No such measures should be undertaken without extensive, thorough review of their impact, and without the negotiation processes with Russia, and perhaps others, to assure that the United States' security is not compromised.
Below is a brief assessment of some of the problems associated with the de-alerting measures most commonly mentioned by proponents.
· Remove and store warheads
Removing warheads from ballistic missiles and storing them separately-de-mating-would have significant adverse impact on survivability and stability. If the warheads were clustered in a limited number of storage locations, they would be high-value targets for pre-emption, thus having a negative impact on stability. Regeneration would be observable and would take as long to complete as the downloading required, which would mean that no timely use could be made of the weapons.
Safety and security are also adversely affected by demating. Warheads are smaller, easier targets for theft or sabotage when separated from their missiles. If warheads were dispersed in many storage facilities to make them less inviting targets, the task of protecting them would be extraordinarily demanding compared to the security required for either warheads on missiles, or warheads in few storage areas.
Verifying the presence of removed warheads in storage would require a complete chain-of-custody from the time they are removed from the missiles. It would be impossible, however, to assure that no duplicate warheads were available for uploading. Verifying that no warhead is secretly replaced on a missile would require continuous intrusive monitoring and, even then, cheating may be feasible.
The financial cost of de-mating would be high, not only because warhead storage areas would need to be constructed, but also because there would necessarily be significant investment in new procedures, logistics for downloading and uploading, and operational training.
· Safing of missiles
Safing of missiles by disabling rocket ignition switches or removing batteries is readily reversible within hours.
Survivability is adversely affected in the case of a sudden first strike, which therefore increases the value of preemption.
Reliability of the missiles probably would suffer from safing measures that start and stop power. The danger would be that, upon restarting power, the missile would not work properly. The problem is akin to a lightbulb or appliances; failure comes when you start them up, not when at a steadily powered state.
Verification would be virtually impossible, even if highly intrusive, constant surveillance were allowed.
· Disabling silo covers
Proposals on disabling silo covers vary, but the general idea is to make it necessary to use heavy equipment to allow the covers to open. It almost certainly be easy to defeat such disabling measures and to open the silos quickly. If a way could be found to make disablement work, the principal downside would the effect on survivability; silo disablement would make quick launch in event of sudden first strike impossible. Verification of most immobilization options would be low-cost, but probably not effective.
· Keep submarines on modified alert
Keeping the submarine force on modified alert (not allowing any portion of the force to be on full alert) delays the capability to retaliate in event of a nuclear strike. As discussed above, this would undermine deterrence. But, it would also would require a complete redesign of the highly complex US C3 system and how it is implemented. The C3 system is designed to support a force that is on alert; procedures and practices are based on this. Any change would require development of new measures to assure that the national command authority would be able to provide timely instructions to the submarine force.
Safety may also be adversely affected by remaining on modified alert because operators will not routinely practice going to alert. They will be more prone to mistakes. Verifiability of this measure is nil.
· Remove launch codes from submarines
Removal of launch codes from submarines would completely defeat the objective of having a stealthy, survivable leg of the triad. Submarines would have to expose themselves in order to receive the launch codes. Verification is not possible.
The "Hair-Trigger" Problem?
Before closing, let me comment on a principal reason given by proponents for de-alerting. They believe that de-alerting is necessary to address the problem, as they view it, of unreliable C3 for nuclear forces in Russia. They are fearful that Russian nuclear weapons can be launched more readily than can ours, and that this might lead to an unwarranted, unauthorized, or accidental launch. An incident in January 1995, involving a Russian alert response to a research rocket fired from Norway, sparked strong concerns that nuclear war could result from a misinterpretation or misunderstanding in Moscow. Then, on 6 February 1997, then-
Defense Minister Igor Rodionov asserted, "if the shortage of funds persists...Russia may soon approach a threshold beyond which its missiles and nuclear systems will become uncontrollable."
It is important to consider all evidence regarding Russian C3 before reaching the conclusion that there is significant danger of an unwarranted, unauthorized, or accidental launch. There are on-going efforts in Russia to modernize nuclear C3. For example, Russia submarines reportedly no longer have the ability to launch without receipt of enabling information from the General Staff.
Additionally, we should look at what Russian leaders say about C3. Rodionov's statement-which appears to have been motivated, at least in part, by a desire to increase the amount of US funding to Russia-was immediately repudiated by high-level Russian officials. The then-Strategic Rocket Forces Commander and now Defense Minister Igor Sergeyev said the C3 system was not on the verge of failure. His remarks were supported by President Yeltsin and Prime Minister Chernomyrdin. Rodionov, in March, declared that the Russian nuclear forces are "reliable and stable" and excluded "the possibility of unusual situations." Later, during a trip to the United States in May, Rodionov reassured the United States that the C3 system was not a problem.
In August 1997, Major General Vladimir Dvorkin, Chief of the Fourth Central Research Institute (for the Strategic Rocket Forces) of the Russian Ministry of Defense, made a trip to the United States to address a group of defense specialists at the US Naval Postgraduate School. His speech dealt extensively with Russian C3. Points he made included: Russia does not rely principally on launch on warning, but rather on the survivability of its mobile forces; the US view of Russian C3 vulnerabilities is erroneous; C3 is very centralized in Russia and there is no possibility that "underlings" can gain control; Russian warning systems are multi-layered; there are positive and negative hardware controls on tactical nuclear systems that prevent their misuse. He closed with the remark that the United States should do a better job of understanding the complexity and competence of the Russian C3 system.
Efforts to do just that began in earnest in late 1997, with an exchange of visits involving the head of US Strategic Command, General Habiger. His impressions of the Russian C3 system were quite positive. He was able to view first-hand the extensive degree of control over Russian nuclear weapons. While there may always be room for improvement in nations' C3 capabilities, care must be taken to avoid concluding erroneously that there is "hair-trigger" danger.
If indeed there is a Russian C3 problem, or if we want to assume that there is in absence of sufficient evidence to the contrary, there is a better way to address the concern than undertaking steps that would be detrimental to our own security. The dialog and exchange of information on C3 should be continued and enhanced, involving multiple levels of the US and Russian nuclear force structure. This information should include procedures for control, doctrine, and technologies, and the full panoply of related issues. Ideas such as a database to provide advance notice of missile firings worldwide might be considered.
The US nuclear deterrent is necessary to address a host of threats-nuclear, chemical, and biological-posed by an increasing number of nations. To assure that the deterrent is as effective as possible, the US should clearly enunciate a policy of reserving the option to retaliate with nuclear weapons against the use of any weapon of mass destruction.
The actual number of warheads or amount of materials for warheads possessed by Russia, and perhaps by others, is not verifiable with current technologies. Great caution therefore should be exercised in any arms control efforts to lower the numbers of nuclear warheads beyond the levels of 2000-2500. If the United States were to reduce to low numbers of warheads while other nations obtained or retained larger numbers, US national security would be gravely at risk.
Additionally, efforts to de-alert our nuclear forces should be strongly resisted. De-alerting has a severe impact on force readiness and stability, as well as a host of other problems. If we have concerns about C3 problems, we shoud address them through other means, not by reducing nuclear readiness, survivability, and safety.