Dr. Kathleen C. Bailey
United States Senate Committee on
Subcommittee on International Security,Proliferation, and Federal Services
March 18, 1998
Mr. Chairman, Members of the Committee, I am very pleased to appear before you today to address the relationship between nuclear nonproliferation and the Comprehensive Test Ban Treaty (CTBT). The views I express today are my own and not necessarily those of any institution.
For most of the last 22 years, I have been working on issues related to nuclear nonproliferation and arms control. In 1976, I was the first social scientist to be hired at Lawrence Livermore Laboratory, where I worked closely with nuclear weapons scientists and engineers analyzing nuclear proliferation intelligence. Subsequently, in Washington, I served as Deputy Assistant Secretary for Intelligence & Research in the Department of State and, later, was Assistant Director for nonproliferation at the Arms Control and Disarmament Agency. I served as the US Governments principal policy official responsible for implementation of the Nuclear Nonproliferation Treaty (NPT) and, as one of my duties, headed our nations delegations to the preparatory meetings for the 1990 review of the NPT. In 1992, I returned to Lawrence Livermore, where I have focused on measures to limit proliferation, and have authored several books and many articles on the topics of nonproliferation and arms control.
Table 1: Administrations Goals for CTBT
Constrain nuclear weapons development by non-nuclear-weapons states
Save the Nuclear Non-Proliferation Treaty (NPT)
Improve ability to detect nuclear testing
Establish an international norm against nuclear testing
Constrain development of advanced nuclear weapons by nuclear weapons states
Proponents for the CTBT have said that it will achieve five nonproliferation goals (see Table 1) and will enable America to maintain a safe and reliable nuclear deterrent. Four of the five goals relate to horizontal nuclear nonproliferationthe spread of weapons to additional countries. The fifth goal is to restrict vertical proliferationimprovement of the nuclear weapons capabilities of America and other nuclear weapons states.
The conclusion of my testimony is that the CTBT fails the cost/benefit test. It will not accomplish the nonproliferation goals as claimed, and therefore will have little benefit. At the same time, the treaty will seriously degrade the US nuclear deterrent, and thus the CTBT will have high national security costs.
The Proliferation Arguments
The CTBT Will Not Meaningfully Constrain Nuclear Proliferation
Claims made about the value of a CTBT to nuclear nonproliferation have evolved. Whereas once treaty proponents argued that a CTBT would help prevent or severely restrict nuclear proliferation, they now say that the treaty will constrain nuclear weapons development. In fact, however, the CTBT will not meaningfully constrain nations that seek to acquire a workable nuclear weapons design.
A state that wants to produce a nuclear weapon can do so, given adequate time and resources, without nuclear testing. It is well known that some single-stage fission designs are relatively simple, and nations would not need to test them to have sufficiently high confidence that they will work. Gun-type weapons, using enriched uranium, would not necessitate testing. For example, the bomb dropped on Hiroshima was a design that had never been tested, and South Africa indigenously built six gun-type nuclear weapons without testing.
Furthermore, the CTBT would not confine new proliferators to simple designs. Today, non-boosted, implosion-type weapons may also be designed with high confidence, without testing. The level of complexity of the nuclear design possible without testing is dependent on the technological sophistication of the nation concerned.
Nuclear weapons testing is not essential now for proliferating nations, as it once was, because information related to nuclear weapons is now widespread. The technological hurdles faced by US weapon designers in the 1940s are long gone. Universities teach courses in physics, engineering, metallurgy, and chemistry that can provide a sound basis for a nuclear weapons program. The information superhighway enables researchers in remote locations to access thousands of relevant articles and reports, as well as to seek assistance from experts who, prior to the invention of the Internet, were inaccessible. Advanced computers, although not a prerequisite, are readily available and make weapons design easier. The state of knowledge has also advanced with regard to materials, which makes it easier for a nation to design lighter, less bulky weapons than those built at the outset of the US nuclear weapons program. When combined, these variables make feasible for a nation to design with high confidence a nuclear weapon that, in the not-so-distant past, would have considered relatively sophisticated.
Critics may argue that new proliferators would want to test a designjust as the United States usually doesbefore stockpiling it. However, there are important differences between proliferators needs, perspectives, and targeting requirements, and those of the United States and Russia. During the Cold War, both sides focused on targeting one anothers military sites. A premier objective has been pinpoint strikes against small targets such as silos, rather than cities. This dictated high-performance delivery systems, which, in turn, required tight parameters on the allowable weight, size, shape, safety measures, and yield. In the case of the United States, these requirements contributed to the US reliance on highly complex designs. Additionally, US and USSR interest in the battlefield effects of their warheads was high, and both have had high standards for reliability.
By comparison, proliferators are likely to target cities, not silos. Their delivery vehicles may be ships, barges, trucks, or Scud-type missiles. Proliferators may not care whether the yield they obtain is exact, may not have tight restrictions imposed by advanced delivery systems or safety standards, are unlikely to have highly complex designs, and may not care about battlefield effects. Furthermore, proliferators may have an entirely different standard for reliability. In other words, it is quite feasible for a nation to develop a device that will work, as long as it does not matter if the yield is exactly known and there are no exacting specifications which must be met. Such differences in requirements may help explain why Pakistan, and perhaps others, have concluded that nuclear testing is not a prerequisite to building a nuclear arsenal.
In summary, the CTBT will not create a significant obstacle to nuclear proliferation. The need for nuclear testing increases in proportion to the complexity of and requirements for the nuclear warhead design. However, it is quite feasible for a nation to design, build, and stockpile nuclear weapons without nuclear testing.
The Non-Proliferation Treaty Is At Risk, Regardless of a CTBT
Robert Bell, Special Assistant to the President for National Security Affairs, makes two points about the linkage between the CTBT and the future of the Nuclear Non-Proliferation Treaty (NPT), both of which imply that the CTBT will be necessary to assure the survival of the NPT. The first statement is that the CTBT and NPT are complementary and, One without the other isnt going to be viable. Second, Bell says that the NPT will be at risk when the NPT is next reviewed in 2000 if the United States has not ratified the CTBT.
It is an exaggeration to say that the NPT is not viable without the CTBT. The NPT has been valued by its parties since it came into force in 1970 as a means of increasing their security. Absence of a CTBT has not negated, nor will it negate, that value.
Bell is correct, however, in saying that nonratification of the CTBT could contribute to the NPTs political woes. What is missing from his assessment is the fact that CTBT passage will not save the NPT from those problems. For at least three reasons, discussed below, the NPT will be at increasing risk regardless of whether the United States ratifies the CTBT.
Demand for Timetable for Zero May Unravel the NPT
A contradiction exists which has long been a sore point with many NPT parties and is now building to a crisis which threatens to unravel the NPT. The United States and the other nuclear weapons states pledged in Article VI of the NPT that they would work in good faith toward total nuclear disarmament. Simultaneously they have continued to rely on their nuclear deterrents for security and have said that disarmament is a long-term rather than near-term goal.
Prior to the NPT Review and Extension Conference of 1995, several nations insisted that the nuclear weapons states abandon nuclear deterrence and negotiate a timetable by which zero nuclear weapons would be reached. They used the decision on extension of the NPT as a bargaining tool. Because the nuclear weapons states would not agree either to zero or a timetable, the NPT Conference was very nearly a failure. However, a last-minute compromise was reached which included a commitment to achieve a CTBT and a structural change for future NPT review conferences. Both decisions help set the stage for continuing troubles for the NPT.
The CTBT will be controversial among many NPT parties, who are in the process of discovering that it does not constitute the step toward disarmament that they had thought it was. The CTBT was touted as a disarmament measure because the United States and others said it would help halt vertical proliferationdefined as the modernization and enhancement of the capabilities of the declared nuclear weapons states. The view of most NPT parties has been that the United States and other weapons states would be unable to test, resulting in erosion of confidence in reliability and, ultimately, degradation of the usability of the nuclear weapons. This would set the stage for the United States and others to abandon nuclear deterrence and go to zero nuclear weapons.
The objective for the CTBT, from the perspective of many NPT parties, will not be met because the United States and other weapons states are not abandoning nuclear deterrence, but are taking steps to assure that their stockpiles will remain safe and reliableand therefore usabledespite the testing ban. The US stockpile stewardship program (SSP) is designed to enhance understanding of the physics of nuclear weapons, to assure the ability to remanufacture warheads, to maintain the ability to test if needed, and to certify credibly the safety and reliability of US nuclear weapons. Thus, nuclear erosion, the goal set for a CTBT by many nations around the world, is effectively undermined by a successful SSP. As a result, many nations and non-governmental groups already have declared that the CTBT does little or nothing to fulfill the NPT Article VI obligation to abandon nuclear deterrence and move toward zero.
Another part of the compromise reached at the 1995 NPT conference established an enhanced review process for the NPT. This process locks in treaty review conferences every five years, as well as preparatory meetings in the three years preceding the review conferences to evaluate progress toward meeting the objectiveswhich refers primarily to the objective of total disarmament as contained in the NPTs Article VI. Whereas it was previously possible for the United States and other weapons states to say that nuclear disarmament was a goal whose achievement was at a time unknown in the future, there now is expectation that concrete steps be defined and fulfilled. We can therefore expect that each NPT review will be acrimonious and that the risk to the NPT will grow.
Dissatisfaction with US Technology Transfer Restrictions Is Growing
There also has been a long-standing unhappiness among many parties that the United States and other nuclear suppliers are not fulfilling the NPTs Article IV requirements to provide technology for peaceful uses. That dissatisfaction is growing significantly due to the US pressures to stop civil nuclear exports to Iran and other countries, particularly from Russia and China. Although Iran is a party to the treaty and has volunteered to allow visits to its nuclear sites to bolster its claim that there are no illicit activities, the United States is leading a campaign to prevent suppliers from selling nuclear technology to Iran. US officials openly voice suspicions that Iran is working on nuclear weapons, but has not convinced the international community that these claims are based on fact. The US actions are made more disturbing to many NPT parties by the US-led effort to provide nuclear technology and aid to North Korea, which remains in noncompliance with the treaty. It appears to many nations that the United States has rewarded a nation that cheated on the NPT and is punishing one that is seemingly in compliance.
Erosion of NPTs Contribution to Security
Violations of the NPT by North Korea and Iraq, as well as Irans unconfirmed activities, have contributed to the third major problem that plagues the NPTerosion of the NPTs security value. The fact that nations can design nuclear weapons and build weapons production facilities without timely detection, and that the international community does not deal effectively with such cheating when it is found, degrades the value of the NPT. The NPT has long been marketed as a means of helping assure security of nations, but these violations and lack of response belie that assurance.
The value of the NPT to its parties security is further eroded by a phenomenon unrelated to nuclear proliferationthe spread of chemical and biological weapons (CBW). Increasingly, nations are recognizing that their perceived adversaries are turning to CBW because they are technologically much easier, less expensive, and less observable than nuclear weapons. While NPT parties remain concerned about some nations nuclear activities, nuclear proliferation is not as prominent a security concern as it once was.
In summary, the NPT is a troubled treaty, regardless of whether the CTBT is ratified or not. Some nations will attempt to hold the NPT hostage to demands for total disarmament on a timetable and/or freer nuclear technology transfers. Willingness to abandon the NPT will increase as nations come to view the NPT as unable to contribute meaningfully to their security. The United States should not ratify the CTBT, and thus compromise the safety and reliability of its own nuclear arsenal, in a futile attempt to ameliorate the NPTs problems.
The International Norm Argument is Meaningless
CTBT proponents contend that the test ban will constrain even those who are not party to the agreement from conducting nuclear tests because an international norm will have been created. History is replete, however, with examples when normsand even legally binding treaties, which are a stronger constraintfail to inhibit nations. Usually this occurs when a nation views breaking the norm or the treaty to be in its security interests.
The NPT norm against the pursuit of nuclear weapons has been broken repeatedly, both by the treatys parties and by non-parties. The norm was established when the treaty went into effect in 1970. The list of states which broke or are thought to have broken the norm include: Argentina, Brazil, India, Iran, Iraq, Israel, North Korea, Pakistan, South Africa, South Korea, and Taiwan.
Another example of the failure of an international norm is provided by the history of non-adherence of some parties to the Biological and Toxin Weapons Convention (BTWC), which outlawed the possession of biological weapons. Iraq had signed but not acceded to the BTWC, but proceeded secretly to produce massive quantities of biological agents. The Soviet Union, and later Russia, violated not only the norm, but the treaty. As a party to the BTWC, Russia was obligated not to produce biological or toxin agents, yet there is convincing evidence provided by defectors that Russia continued to do so. Given the non-verifiability of the BTWC, one could argue that the propensity of nations to respect a norm is directly proportional to the probability that violations would be detected. In the case of the CTBT, the lack of verifiability is a serious drawback.
Downsides of the CTBT
The CTBT Is Unverifiable
In his letter of transmittal to the Senate, President Clinton states that the CTBT is effectively verifiable. The is correct only if you define verifiability by weak standards. Let me discuss the definition of effective verification and give you the facts that support my conclusion that the CTBT is neither effectively nor meaningfully verifiable.
When I served in the Reagan and Bush Administrations, effective verification was accepted to mean high confidence that militarily significant cheating will be detected in a timely manner. In the case of the CTBT, this would mean that you are highly confident that you will be able to detect, within hours or a few days of the event, any nuclear testing which will provide the tester with militarily significant weapons information. There are at least two key questions that therefore must be addressed: What yield nuclear test can provide militarily significant information? and, Can the CTBT verification system detect to that level?
Experts will disagree on the usefulness of different yields to different testers, but I think they will all concur that less-than-full yield testing can be enormously important, and militarily significant. For example, many in our own nuclear weapons design community wanted to retain the ability to test at some level. Attached is a table, agreed to by all three US weapons laboratories, which makes the point that testing at 500 tons is a very useful testing level, although not sufficient to gain full confidence in all aspects of an existing weapons performance or to develop sophisticated new nuclear weapons
The lowest possible yield to accomplish new designs as well as safety and reliability depends upon warhead requirements, but most designs could be adequately tested at yields between one and 10 kilotons. 500 tons would be sufficient for reliability testing, but a higher yield would be needed to certify any new design that was a major departure from already-tested designs. Therefore, it is reasonable to assume that 10 kiloton tests would be militarily significant, and tests down to a level of 500 tons may also be.
The International Monitoring System (IMS) of the CTBT is expected to provide the ability to detect, locate, and identify non-evasive nuclear testing of 1 kiloton or greater. However, a nation may conduct nuclear tests evasively, which would allow several kilotons to be tested with little or no risk of detection. One method by which this might be done is through energy decouplingdetonation of the device in a cavitythat can reduce the signal by as much as a factor of 70 are thought to be possible. Thus, a kiloton explosion could be made to look seismically like a 14 ton explosion fully coupled. A 10 kt explosion could look like a .14 kt explosion.
Let me give you an example. The United States conducted two nuclear tests in the Tatum salt dome located at Chilton, Mississippi. Sterling, the test conducted on December 3, 1966, had a yield of 380 tons, but the apparent seismic yield was only 5.3 tons, a reduction by a factor of 71.7.
It is clear that the IMS will not be able to detect nuclear testing below 1 kt and, if the test is evasively conducted, will not detect several kilotons. CTBT proponents say that supplemental data from US national technical means will fill the gap. This is not entirely accurate. The United States has stated that its objective is to have the capability of identifying and attributing with high confidence evasively conducted nuclear explosions of about a few kilotons yield in broad areas of the globe. The US Intelligence Community acknowledges that this is a complex task that will require a lot of work, time, and the necessary resources to achieve. For the present, even with a fully functional IMS supplemented with data from US national technical means, it is possible that a militarily significant test could be evasively conducted without detection.
Another problem with detection and identifying low-yield events is the large number of signals in these ranges. At lower yields, the number of non-nuclear events of similar size increases (e.g., mining explosions and earthquakes on land, explosions for geophysical exploration, volcanoes at sea, meteorite impacts in the atmosphere). These non-nuclear events increase the total number of events to be processed by a verification system, and a small percentage of them generate signals similar to those expected from nuclear explosions. This increases the difficulty of identification.
In addition to its technical limitations, the IMS has other problems. A number of the countries in which the facilities are to be located will be unable or unwilling to pay for them, and may not have the technological wherewithal to properly manage them. Upkeep and protection of the facilities is also an issue, as is the actual location of some facilities. One difficulty, for example, is that some of the stations monitoring China will be within China. Hypothetically, if China wanted to test, it could assure that the station(s) would not be working during the time of the test. Another scenario is that a nation might test a nuclear weapon in the ocean, where identifying the origin of the device might be impossible. The perpetrator could then depend on the IMS to analyze the yield of its device and there would be no way to attribute the test.
In summary, the verification regime of the CTBT increases our capabilities to detect nuclear tests at yields higher than 1 kt non-evasively conducted, and up to 70 kt evasively conducted. This means that militarily significant testing can be conducted with little or no risk of detection by either the IMS system or the supplemental capabilities of US technical means.
The CTBT Will Constrain Nuclear Modernization
The Clinton Administration asserts that a purpose of the CTBT is to prevent the United States and others from being able to modernize nuclear weapons. There is no question that the inability to test will limit the abilities of nuclear weapons statesassuming that they do not cheat under the CTBTto modernize their nuclear forces. The directors of US and Russian nuclear weapons laboratories have stated that they would not choose to introduce new, modern, high yield-to-weight ratio warhead designs into their countries stockpiles without nuclear testing. It is reasonable to expect that decision-makers in other nuclear weapons states would have similar views.
The important question to ask is whether constraining nuclear modernization is desirable. Some argue that constraint is a good outcome because it will undermine the confidence that nuclear weapons states have in their arsenals, making them less usable. However, there are also serious downsides, as outlined below.
Lack of Modernization Forecloses Safety Improvements
Think of the safety improvements for automobiles that have resulted from evolving technology over the past decade. Advances in materials science, electronics, and concept innovation have led to better crash-proofing, airbags, design principles, etc. These could not have been introduced without actual testing.
It would be foolish to assume that similar advances in technology will not produce measures which could make US and other nations nuclear weapons safer. For example, it is possible that in the future, researchers will discover higher energy, insensitive explosives which will make it possible to make warheads safer, without compromising weight limitations. Because nuclear weapons are extraordinarily complex, testing would be required before such advances could be integrated into stockpile designs.
Modernization May Be Needed for New Requirements
The only claim by treaty proponents that rings true is that the CTBT will constrain not preventmodernization. Constraining modernization is risky, however, because it seriously degrades our ability to tailor the arsenal to emerging or as-yet-unknown threats, or to adapt it to changes in defensive technologies.
At present, the United States has no specific requirement for new nuclear weapons designs, but we cannot be certain that this always will be the case. Saddam Hussein and Desert Storm have taught us that we need to be able to strike and destroy deeply buried targets such as underground bunkers. Similarly, the proliferation of chemical and biological weapons worldwide raises the question of whether the current US stockpile, which was designed primarily to destroy Soviet nuclear weapons silos, is tailored for the missions of the future. For example, much has been made of the inability of the Patriot missile system to destroy incoming SCUD missiles during the Persian Gulf War. Fortunately, none of these warheads contained anthrax, ricin, or even a nuclear warhead. If they had, the casualties in Saudi Arabia and Israel could have been staggering. Patriot missiles tipped with low-yield nuclear warheads are capable of engaging and vaporizing any incoming Scud missile and rendering its lethal cargo useless.
Preserving the option of modernizing US nuclear weapons is important also in the context of emerging defensive technologies. In the future, adversaries may discover new means to render US warheads or delivery vehicles obsolete. Such a technological breakthrough could necessitate a complete overhaul of US warheads and delivery systems.
Old Weapon Designs Decline in Deliverability
Nuclear warheads are designed to be mated with specific delivery systems. Delivery systemsaircraft, boats, ships, etc.age and must be replaced. Furthermore, as adversaries develop countermeasures, our systems must be improved. These newer systems have improved electronics, materials, performance criteria, and other attributes, all of which affect the optimal design of weapons they deliver. (It is, of course, possible to reverse the process and begin to design delivery vehicles to the parameters of existing weapons, but this has its own set of downsides and should be carefully considered.)
It is possible to make some changes to warhead design without testing. However, for nuclear weapons to continue to be deliverable by newer systems, it may be necessary to change the weight, size, and shape of the warhead. Without testing, the ability to significantly change the parameters of US nuclear warheads will be limited.
In summary, the CTBT will constrain nuclear weapons modernization, which may be a negative outcome. Any advantages of constraint must be weighed against the disadvantages of freezing our weapons technology. Evolution in technologies for safety, defenses, and delivery systems may render the now-modern US nuclear arsenal less-than-optimal.
The CTBT Erodes Confidence in the US Nuclear Stockpile
The United States continues to need a strong nuclear deterrent to respond to existing and emerging threats. Both Russia and China have significant nuclear arsenals, and perhaps undeclared chemical and biological arsenals as well. Meanwhile, chemical and biological weapons capabilities are proliferating, a threat against which our nuclear deterrent may be the only workable response. We should not inhibit our capabilities to keep that deterrent strong, safe, and reliable.
Nuclear testing has helped assure the reliability of US nuclear weapons. According to former Los Alamos National Laboratory Director Sig Hecker, confidence in the stockpile has decreased since testing ceased in 1992. Nevertheless, the decline in confidence is manageable because no new designs have been introduced, many experienced designers and engineers are still at hand, and there is an extensive test history related to the weapons in the stockpile. This will change, however, over time. Experts will retire and the weapons will age, possibly causing deterioration that could affect the workability of the weapons.
The weapons laboratories hope to assure the continuing confidence in reliability through the stockpile stewardship program (SSP). However, the SSP faces challenges which raise serious concerns, including:
The SSP facilities will not be completed for some years, perhaps more than a decade. There is no certainty that the technologies of SSP will work as intended or that SSP will enable scientists to understand weapons physics well enough to replace the knowledge previously gained through testing.
The funding for the program is promised, but the required $4.5 billion/year for 10 years must be annually agreed by Congress and defended in light of other pressing priorities. Already, there is a budding campaign in the House of Representatives against full funding which, if successful, could at best push completion of SSP too far into the future and, at worst, assure its failure.
Support from the arms control community is lukewarm at best. Many advocates of CTBT say that they will have to reconsider their support for SSP in the future, after the CTBT is ratified.
SSP managers may limit the types of experiments they are willing to do because of fear of adverse reaction from anti-nuclear activists. This could make the SSP less relevant to nuclear weapons design.
In summary, we still face threats for which we need a strong nuclear deterrent. That deterrent is eroded by the cessation of nuclear testing. SSP, the stopgap measure, faces serious challenges and may fail.
The CTBT May Promote Nuclear Proliferation
Nuclear testing not only enabled the United States to keep its nuclear weapons design skills at a high level, it demonstrated to allies and potential adversaries alike that the US arsenal is reliable and US commitment to nuclear deterrence was strong. Any decline in US confidence in or commitment to its nuclear deterrent will increase the likelihood of proliferation by nations currently under the US nuclear umbrella. Japan and South Korea worry about the plutonium that North Korea still possesses and about the fact that Pyongyang may have secret nuclear weapons and/or a clandestine nuclear weapons program. European nations may not trust Russia to disarm at the same pace as the United States, or to cease its nuclear testing at yields below detectability. Although nonproliferation commitments by all three advanced nations are strong, these may change if they perceive that they are threatened by nuclear weapons, but have little trust in the viability of the US nuclear deterrent.
For the sake of nuclear nonproliferation, we are considering ratification of the CTBT. Yet, the CTBT will not meaningfully accomplish the nonproliferation goals set out for it. Nations will be able to design and deploy nuclear weapons without testing. The NPT will be at risk regardless of whether the CTBT is ratified by the Senate. The CTBT will not be able to detect militarily significant cheating, even if supplemented by US national technical means. The objective of creating an international norm against testing is, as history has demonstrated with other arms control norms and agreements, not meaningful. Thus, the potential benefits of the CTBT to nuclear nonproliferation are meager.
On the other hand, the CTBT will have a profound impact on the ability of the United States to assure that its nuclear weapons continue to be as reliable, safe, and effective as can be. Ratifying the CTBT will foreclose the ability of the United States to modernize its nuclear forces because US compliance is certain. However, given that the CTBT is not effectively verifiable and that other nations have a history of noncompliance with arms control treaties, militarily significant cheating may occurto the disadvantage of US security. Thus, the limited political benefits of the CTBT are not worth the high cost to our national security.