- CTBT Ratification: Pros and Cons
- February 1999
Cato Policy Analysis No. 330
January 15, 1999
The Comprehensive Test Ban Treaty
The Costs Outweigh the Benefits
by Kathleen C. Bailey
- Kathleen C. Bailey is an author and defense analyst.
- Executive Summary
- The Comprehensive Test Ban Treaty (CTBT) is now before the U.S. Senate
for its advice and consent. The treaty bans all explosive testing of nuclear
- Advocates of the CTBT make several arguments in support of the treaty.
The reasons reduce to two points: the ban will constrain the modernization
and development of nuclear weapons by the nations that already possess
them, and it will help prevent the spread of nuclear weapons to additional
nations. Both objectives are set out in the CTBT's preamble.
- Opponents of the CTBT are most concerned about one issue: in the absence
of nuclear testing, U.S. nuclear weapons can be neither as safe nor as
reliable as they should be. Those deficiencies will diminish the effectiveness
of the U.S. nuclear deterrent. While the treaty will constrain the United
States from modernizing and developing weapons, it will be possible for
other nations to cheat with little or no risk of being caught because the
CTBT cannot be verified.
- To resolve safety and reliability questions, the Clinton administration
has developed the U.S. Stockpile Stewardship Program (SSP). The SSP is
intended to improve knowledge about nuclear weapons to such an extent that
it will be possible to fix problems and design new weapons without nuclear
testing. The SSP is extremely expensive and technologically very risky.
- Furthermore, it is unclear whether the SSP will accomplish its goal
of attracting, training, and retaining scientists and engineers capable
of fixing future problems with current weapons and designing new weapons.
- Text of Policy Analysis
No. 330 (PDF, 31 pgs, 124 Kb)
- Click here for the Cato Home Page
A Test Ban Will Benefit U.S. and International
A Reply to Kathleen Bailey
By Christopher Paine, Senior Researcher,
NRDC Nuclear Program
- The Cato Institute's new Policy Analysis seeks to weigh the
costs and benefits of the Comprehensive Test Ban Treaty, and finds that
the former outweigh the latter. However, this conclusion is reached through
the application of sweeping and largely unsubstantiated generalizations
about such complex matters as verification capabilities, the role of global
norms in influencing the behavior of nation states, the role of nuclear
explosive testing in maintaining U.S. weapons reliability and safety, and
the role of nuclear explosive tests in the weapon development programs
of prospective proliferant states. Moreover, omitted from the analysis
is any consideration of the consequences for U.S. and international security
that would likely ensue if the United States were now to resume nuclear
explosive testing, as suggested by the paper.
- In the analysis that follows, Dr. Bailey's contentions are highlighted
in italics, followed by my response.
- (The cover page for this report, "Facing Reality," is saved
- Opponents of the CTBT are most concerned about one issue: in the
absence of nuclear testing, U.S. nuclear weapons can be neither as safe
nor as reliable as they should be. Those deficiencies will diminish the
effectiveness of the U.S. nuclear deterrent. (Executive Summary)
- This formulation begs a host of questions which are not answered in
the report, such as "how safe is safe enough," and "how
'reliable' does a weapon have to be to constitute an effective deterrent?"
Moreover, the implied relationship between nuclear deterrent reliability
in the statistical sense and a successful underground nuclear test does
not exist. During the Cold War, neither weapons nor delivery systems were
ever tested sufficiently to provide a valid indicator of their "effectiveness"
in war. It is even more unlikely that the United States would do so now,
even if afforded the opportunity to resume testing.
- In agreeing to the extension of the original 9-month test moratorium
established by the Hatfield-Exon-Mitchell Amendment, the Joint Chiefs of
Staff reviewed the option of conducting an additional 15 tests over four
years for the purpose of incorporating improved plutonium dispersal safety
in 400 W-88 SLBM warheads and fire-resistant pits in two weapon types delivered
by aircraft. These upgrades were rejected by DoD, the Navy, and the Air
Force as not constituting cost effective investments in safety, especially
in view of post Cold War changes in storage and alert procedures that reduced
the likelihood of the very scenarios the proposed upgrades were designed
- While the treaty will constrain the United States from modernizing
and developing weapons, it will be possible for other nations to cheat
with little or no risk of being caught because the CTBT cannot be verified.
- In reality, because of the depth and breadth of the Stockpile Stewardship
Program, and the sheer amount of resources being devoted to it, the United
States is better positioned than any other weapon states to "modernize"
its weapons without resort to nuclear test explosions, and indeed, it is
already doing so.
- Moreover, while some amount of cheating at low yields is "possible"
under a CTBT, as explained in detail below, most scenarios can hardly be
considered probable, and it is utterly preposterous to suggest that such
activity runs "no risk" of detection by the CTBT monitoring regime,
or by U.S. unilateral monitoring capabilities, the intelligence capabilities
of friendly states, and the world scientific community, whose seismic monitoring
capabilities exceed that of the CTB's International Monitoring System in
many areas of the globe.
- Equally preposterous is the implication that nameless "other nations"
would be likely to achieve technical results through clandestine tests
that would somehow nullify America's enormous advantages in resources and
technology and the knowledge that comes from having conducted over 1000
nuclear test explosions.
- The organization will oversee the treaty's verification regime,
called the International Monitoring System (IMS), and an International
Data Center. (Introduction, p.2)
- Actually, the treaty's "verification regime consists of the IMS
and two other components On-site Inspections and Confidence Building
Measures. The paper does not mention or assess the contributions of these
- Because the CTBT is likely to have a profound impact on the reliability
and future safety of the U.S. nuclear deterrent, the treaty's ratification
is contentious. (p.2)
- In reality, the CTBT is contentious because it (a) tends to downgrade
the importance of nuclear weapons, and hence the role of the U.S. nuclear
weapons establishment, in U.S. foreign and security policy, and (b) it
severely constrains the development of new types of nuclear weapons and
limits the ways in which nuclear weapons can be adapted or redesigned for
changing military missions. As a purely technical matter, it is certainly
possible, and if good judgment and managerial discipline prevail, it is
indeed likely that a CTBT will have either a positive or no impact on weapons
safety, and only a modest or no negative effect on confidence in enduring
stockpile performance. A detailed explanation of why this is true is provided
later in this critique.
- Weapon Modernization
- Constraining modernization is risky because it seriously degrades
the ability of the United States to tailor its arsenal to emerging or as
yet unknown threats or to adapt it to changes in other nations' defensive
- While there is admittedly, a large subjective policy component to the
question of how much importance to assign nuclear weapons modernization
as a means of coping with future security threats, Dr. Bailey overstates
the post Cold War risks in this regard. While compelling national security
justifications are lacking, existing nuclear explosive packages can, as
a technical matter, be integrated into new or modified warhead and bomb
systems, and these systems in turn can be mated to new or modified delivery
systems, without resort to nuclear explosive tests.
- In other words, under a CTBT some of the operational characteristics
of nuclear weapon systems can be adapted to changing military missions.
Improved casings, radars, altimeters, boost-gas delivery systems, neutron
generators, detonators, batteries, integrated circuits, fuzing and arming
systems, permissive action links all can be developed, tested, and
integrated into nuclear bomb and warhead systems without modifying the
primary or secondary components of the nuclear explosive package design.
Indeed, Bailey herself cites the case of the conversion of the B61-7 strategic
bomb into the B61-Mod 11 earth penetrating weapon for destruction of deeply
buried targets, but fails to note that this conversion was carried out
and certified for the stockpile by the Stewardship Program without nuclear
- If the proliferation of missiles armed with chemical or biological
weapons agents becomes a more serious threat to the Untied States and its
allies in the future, it may be prudent to include in the U.S. nuclear
arsenal some warheads designed specifically for the mission of destroying
such agents either in their storage areas or on incoming missiles. (p.5)
- It would not be in the US interest to explicitly adopt a posture of
using nuclear weapons to deter the use of biological/chemical weapons,
since the adoption of this posture would only encourage and legitimize
the proliferation of nuclear weapons by currently non-nuclear weapon states
facing far more imminent threats of such use than the U.S. Moreover, as
a technical matter, it is by no means clear that a new rather than existing
nuclear explosive package would be required to execute the Chem/Bio incineration
mission, and the feasibility of this mission is questionable in light of
the possibilities for delivery via widely dispered submunitions.
- Finally, serious threats from the emergence of hostile ballistic missile
powers armed with WMD is provided for under the treaty's withdrawal clause,
which explicitly allows a right of withdrawal to any party in the event
that "extraodinary events have jeopardized its supreme national interests."
It hardly seems necessary to jettison the near term benefits of a nuclear
test ban merely to massage somewhat tenuous nuclear options for countering
future WMD threats, which can be maintained in any case at some reduced
level under a CTBT. Indeed, the CTBT's positive contribution to international
security might well be such that at least some of Dr. Bailey's future WMD
threats fail to mature as anticipated.
- Preserving the option of modernizing U.S. nuclear weapons is also
important in the context of other nations' emerging defensive technologies.
We cannot know what means opponents may develop to render U.S. warheads
or delivery vehicles obsolete. Such technological breakthroughs could necessitate
a complete overhaul of U.S. delivery systems and nuclear warheads.
- Again, as a technical matter, preserving the "option" of
modernizing U.S. nuclear weapons to counter "emerging defensive technologies"
does not require an ongoing nuclear testing program. Indeed, the most likely
initial countermeasures would involve changes to the missile and reentry
systems, but not necessarily to the nuclear explosive package. On a political
level, Dr. Bailey's concern can only be described as arcane and even bizarre.
Didn't the U.S. government just announce its intention to deploy a missile
defense system, and wasn't it the U.S. secretary of defense who stated
that the U.S. would unilaterally abrogate the ABM Treaty, if necessary,
to achieve this objective? If conservative advocates of missile defenses
are sincere in their desire to shift defense strategy increasingly toward
defenses and away from retaliatory nuclear deterrence of attack, a test
ban should be regarded as a welcome adjunct to such a policy. A CTBT would
severely limit the ability of future adversaries to build the very kind
of sophisticated high yield-to weight MIRVed warheads that could penetrate
such a system.
- Dr. Bailey is also concerned that as the US changes its delivery systems
over time new nuclear explosive packages will be needed for them. After
all, in the past new warheads were designed to fit the delivery system,
not the other way around. But those days are over. If new missile and reentry
systems are required in the future, they can be designed to accommodate
the dimensions and performance envelopes of existing nuclear explosive
packages. While the resulting combinations might represent less than the
optimal yield-to-weight or yield-to-eight ratio achievable with nuclear
testing, should anybody really care? In fact, we have already adopted this
approach. New Trident II missiles are built to carry older W76 warheads
as well as newer W88s, and a modified W87 MX warhead is slated for redeployment
under START II on the refurbished single warhead Minuteman III missile.
- Will a Test Ban Foreclose Needed Safety Improvements?
- "...we can assume that nuclear weapons technology will continue
to advance and that new measures to make nuclear weapons safer will be
discoveredtesting would be required in most cases before such advances
could be integrated into stockpile designs." (p. 6)
- Just because a new technology comes along does not mean that it would
be cost-effective to rebuild the nuclear arsenal to incorporate it. In
theory, with a big enough collector, one could also generate electricity
from the sunlight reflected off the moon. Just because a technological
possibility exists does not make it indispensable or even sensible. As
already noted, the current arsenal is "safe" in that it meets
modern "one point" safety standards against accidental nuclear
detonation. This characteristic is an inherent function of a boosted primary's
nuclear design, and it does not disappear with age. Likewise, the susceptibility
of the chemical explosives in nuclear explosive packages to accidental
detonation does not increase with age, so as the arsenal ages it is not
expected to pose a greater plutonium dispersal risk.
- The only question then is do we want to make nuclear warheads "safer"
than they are today? Dr. Bailey gives the example of high explosive formulations
that are less sensitive to impact. In fact, the Navy was asked in 1992
if it wanted to replace its existing W88 Trident missile warhead with a
new version with less sensitive explosives. The Navy decided it could improve
the safety of the Trident system more cost-effectively by changing the
way it loaded the warhead onto the missile, rather than changing the warhead
itself. Moreover, many warhead parts relevant to safety and/or use control
- such as detonators, fusing and arming systems, and permissive action
links - all can be improved without modifying the nuclear explosive package
design. The lesson: accelerate the retirement of any substandard weapons,
look for approaches to improved safety that do not involve major changes
to the nuclear explosive package design, and minimize the exposure to potential
accident environments of those weapons with the greater plutonium dispersal
- Are Nuclear Test Explosions Required to Ensure the Reliability of
the Enduring Stockpile?
- To use another automobile industry analogy..Ensuring reliability
means that no "recall" will be warranted.(p.6)
- On the contrary, ensuring reliability has always included, and continues
to include, having the capability to both detect the need for and execute
a "recall" should an "actionable defect" recur in a
large sample of warheads. Detecting such defects is a major objective of
the Stockpile Surveillance Program.
- As noted previously, some less advanced types of nuclear weapons
designs do not require testing to ensure reliability. Such designs are
relatively simple, and their performance can be calculated and modeled
with high confidence. Advanced designs, such as those in the U.S. stockpile,
are extremely complicated. They have many variables and several thousand
components.With the technology available today, there is no way to simulate
nuclear detonation of the high-performance complex designs in the U.S.
- Many times in the past, U.S. nuclear weapons designers were surprised
by the results of nuclear tests, which revealed problems the designers
had not imagined. The tests showed them that they had not understood conditions
and technologies as well as they had thought. Thus while some defects have
been discovered through surveillance of the stockpile and non-nuclear testing,
other problems with U.S. nuclear weapon designs have been identified solely
as a result of a nuclear test. (p. 9-10)
- The above statements constitute a compilation of fallacies. First,
they fail to distinguish between the complete bomb or warhead system, which
may indeed have "thousands of parts," and the nuclear explosive
package, which has roughly two orders of magnitude fewer parts.
- Second, they fail to distinguish between the tests that have historically
been required to certify the nuclear explosive performance of a new nuclear
explosive package design, and those intended to confirm, on a random selection
basis, satisfactory performance of stockpiled weapons that had been subjected
to the stresses of a simulated "stockpile-to-target" sequence
and then detonated underground in Nevada. For most of the Cold War, such
"stockpile confidence" tests were not conducted, because until
the advent of the Reagan Administration the laboratory leaderships considered
them a waste of resources. These "reliability" tests merely repeated
the "stockpile certification" tests that all new weapon designs
underwent prior to quantity production for the stockpile, and of the 13
conducted since 1980, only one revealed a significant deviation from expected
performance, which was traced to a tritium maintenance problem and easily
corrected without modifying the nuclear explosive package.
- Finally, the empirical record of computationally predicted versus measured
test yields shows that "with the technology available" the laboratories
actually did an excellent job of predicting the performance of the "complex"
boosted devices now in the enduring stockpile. The mean deviation of predicted
versus radiochemically measured test yields is classified but is reassuringly
small, even for first time tests of new boosted primary designs. Nor does
the test record support the contention that current complex designs are
unpredictably sensitive to small changes in production tolerances and materials,
as the shift from laboratory built R&D test devices to fully engineered
factory built weapons reveals no significant deviations in nuclear explosive
- As discussed in greater detail below, very few problems less
than 1% -- were identified "solely as a result of a nuclear test."
- In 1997 the [then] director of Los Alamos National Laboratory, Sig
Hecker, wrote to Sen. Jon Kyl (R-Ariz.), stating that confidence in the
U.S. stockpile had decreased since the last U.S. test in 1992. Hecker also
said that several problems, some of them age related, had developed which
previously "we would have turned to a nuclear test in the kiloton
range to resolve." (p. 7)
- This is a rather incomplete synopsis of Hecker's views, which were
provided in response to a list of questions sent to him by Senator Kyl,
a test ban opponent. It condenses the response to two separate questions.
In response to the question, "will confidence in the safety and reliability
of U.S. nuclear weapons decline without nuclear testing?" Hecker actually
replied as follows:
- "The stockpile stewardship and management program,
designed jointly by the Department of Energy's Defense Programs and the
weapons laboratories, has allowed us to continue to certify the safety
and reliability of the stockpile although it has been five years since
we last conducted a nuclear test. As anticipated, our confidence in the
nuclear stockpile has decreased somewhat during that time frame. This decline
in confidence is an inevitable consequence of lack of testing. To date,
we have found the decline to be manageable because we have not introduced
any new weapons into the stockpile and we still have on hand a cadre of
experienced nuclear weapons designers and engineers. Moreover, we have
an adequate nuclear test history for the weapons in the stockpile. I have
just sent my second annual letter to the Secretaries of Energy and Defense
certifying the nuclear weapons we designed to be safe and reliable without
nuclear testing. For the longer term, science based stockpile stewardship
is designed to develop new tools to better understand the fundamental science
and technology of nuclear weapons that will help us shift to basing our
confidence in the nuclear stockpile on SBSS, and away from our historic
reliance on nuclear testing (emphasis added)."(2)
- In the second question, Hecker was asked: "Since the last nuclear
test, have there been age-related or other changes in the stockpile that
previously would have been addressed by conducting at least one nuclear
test? If so how certain are you of the fixes? If your level of confidence
in the fixes is not extremely high, how has this affected your view of
stockpile reliability?" Hecker replied:
- "Yes, there have been several instances since the
cessation of nuclear testing in September 1992, where we have found problems,
either age related or otherwise, for which in the past we would have turned
to a nuclear test in the kiloton range to resolve. In the absence of testing,
we have used the methodology of SSMP to evaluate the problem and suggest
fixes if required. This has included more extensive calculations, non-nuclear
laboratory experiments, comparison to previous nuclear test data, and the
extensive experience of our designers and engineers. Moreover, our assessment
has been checked against the rigors of peer review by the Lawrence Livermore
National Laboratory. We have examined several problems of this nature during
this year's certification cycle. At this time, we have sufficient confidence
in our solutions to certify the stockpile without a resumption of nuclear
testing. If our confidence in the fixes were not sufficiently high,
we would not certify the stockpile. Our experience to date in resolving
suspected problems has increased our confidence in SSMP and in the process
of annual certification (emphasis added)." (3)
- Some additional observations by Hecker, not cited in the Cato Policy
Analysis, are pertinent to the subject at hand:
- "I believe that the SSMP as currently configured
and fully funded provides the best approach to keeping the confidence level
in our nuclear stockpile as high as possible for the foreseeable future.
We recognize that there is no substitute for full-systems testing in any
complex technological enterprise. This is certainly true for nuclear weapons.
A robust nuclear testing program would undoubtedly increase our confidence.
However, our long-term confidence in the stockpile would suffer if we substituted
a program consisting of an occasional nuclear test for a robust stewardship
program because it would lock us into an empirical approach tied to limited
testing data without the benefit of the flexibility and resiliency provided
by better scientific understanding (emphasis added."(4)
- Hecker certainly realizes, even if Dr. Bailey does not, that in the
post-Cold War era "a robust nuclear testing program" cannot be
justified by DOD's current or reasonably foreseeable nuclear weapon requirements,
and could not be justified politically to the American public and the international
community, which overwhelmingly support an end to nuclear explosive testing.
- In his responses to Kyl, Director Hecker returns twice more to the
theme of the tradeoff between continuation of a modest nuclear test program
without the CTBT, and a robust stewardship program with the CTBT, and he
repeatedly chooses the latter:
- "Again, I would like to add the caution that conducting
an occasional nuclear test in lieu of a fully-funded SSMP will jeopardize
our long-term confidence in the stockpile. The SSMP is designed to predict
and correct problems in the stockpile, whereas an occasional nuclear test
would focus primarily on existing problems. It is critical at this time
that we focus the attention of our people on being able to do the best
possible job without nuclear testing."(5)
- "I should also add that in August 1995, when the
President made his [zero yield CTBT]decision, we had already not conducted
a nuclear test for almost three years. Our budgets had decreased precipitously
over the previous six years. Our people were looking to get out of the
nuclear weapons program. The production complex appeared hopelessly broken.
The prospects of doing an occasional nuclear test was proving to be a barrier
to adopt[tion of] a new approach to nuclear stewardship. This situation
has turned around dramatically in the past two years with the emphasis
on science-based stockpile stewardship. Our people have a renewed commitment
to stockpile stewardship and an enthusiasm for the development of a new
methodology, based on rigorous science and engineering, to ensure the safety
and reliability of the stockpile." (6)
- "The 1958-61 test moratorium provides a relevant comparison.
At that time, some stockpile problems were fixed, and there was confidence
that the solutions worked. When the moratorium ended and testing resumed,
the "fixes" were found to be inadequate." (p. 7)
- WRONG. The 1958-61 moratorium does NOT provide a relevant comparison.
Immediately preceding or during the 1958-61 moratorium, nine warheads were
developed and deployed in "crash" programs, when the United States
imprudently rushed some 6500 inadequately tested warheads per year into
the stockpile. Seven of these new warhead types were never tested before
deployment with aged end-of-life components, and developed problems related
to a common cause aging of their tritium reservoirs. This failure
was particularly acute given the lack of experience at the time with the
behavior of boosted warhead designs.
- Today we have an additional 30 years of test experience with such
boosted weapons. Defects in the other two warheads were also directly
traceable to the crash nature of the program in the moratorium era. One
underwent a change in high explosive during the moratorium and was stockpiled
without a nuclear proof test. The other was designed with a known inattention
to the problem of neutron radiation external to the warhead, and then deployed
without a neutron vulnerability test. The resumption of testing uncovered
a severe vulnerability problem. Finally, all nine of these problem warheads
have long since been retired from the stockpile.
- The experience and understanding of the experts who designed the
current U.S. nuclear weapons have not been well documented because the
entire U.S. nuclear weapons program was predicated on the absolute need
for an ability to conduct testing throughout the life of the design. (p.
- This is an excessively dire and therefore misleading view of the state
of U.S. nuclear weapons knowledge. While the specific technical rationales
for certain design decisions lack extensive contemporaneous documentation,
DOE is seeking to fill in these gaps through improved archiving and interviews.
Much of the accumulated U.S. nuclear weapons knowledge is embodied in the
nuclear test calibrated nuclear design codes, which in the hands of skillful
users yielded remarkably accurate performance predictions when new designs
represented iterations or modest extrapolations from previous test experience.
- Moreover, it is emphatically NOT TRUE that "the entire U.S. nuclear
weapons program was predicated on the absolute need for an ability to conduct
testing throughout the life of the design. For example, from 1965 to 1980,
no underground tests were conducted for the primary purpose of identifying
or developing corrections to stockpile problems. According to Sandia National
Laboratory Director Paul Robinson:
- "Historically, only a small fraction of our nuclear
tests were for the purpose of evaluating the stockpile's health, because
we could depend on a variety of other evaluation techniques. The introduction
of new and improved non-destructive surveillance techniques should provide
us the ability to detect problems in the stockpile and to maintain or even
improve the safety of our systems.
- "A principal concern for the future is that, without
carrying out nuclear testing, nuclear design engineers (at Los Alamos and
Lawrence Livermore labs) may not be able to confirm that a fix proposed
to solve a stockpile problem does, in fact, work without introducing some
other unintended and unforeseen problems. This concern has more to do with
whether we will end up unable to correct problems, once they are detected,
than with a general loss of confidence that might result from the elimination
of nuclear testing (emphasis added)." (7)
- In fact, the extent of U.S. dependence on nuclear explosive tests to
maintain stockpile weapon "reliability" is known rather precisely
known. A 1996 tri-lab study of the Stockpile Surveillance Program reveals
that, of some 830 specific recording "findings" of defects in
stockpile weapons from 1958 to 1993, less than 1% were "discovered"
in nuclear tests, and all but one of these tests involved weapons that
entered the stockpile before 1970 and are no longer in the U.S. nuclear
- After 1970, one warhead maintenance problem, related to the effect
of tritium decay on the design yield, was discovered in a Stockpile Confidence
Test (SCT) of the W84 warhead (now in the "reserve" stockpile)
for the GLCM missile eliminated under the INF Treaty, but the problem was
easily rectified without modification of the nuclear assembly system. Another
three underground tests confirmed the existence of problems in the high
explosive of the W68 SLBM warhead (fully retired) and in the cold temperature
performance of the (then new) Insensitive High Explosive (IHE) used in
the W80 ALCM, and the "Mod 4" version of the B61 bomb. But only
four out of 141 (i.e. about 3%) of "Product Change Proposals"
to war reserve weapons specifically required underground nuclear explosive
tests to develop or confirm the corrective actions. In addition, three
routine SCT's reportedly served the dual function of confirming fixes to
already identified problems. Hence a total of 11 tests, or less than 3%
the 387 tests conducted after 1970, were directly related to maintaining
the reliability of the existing stockpile.(8)
- When asked to quantify the reduction in stockpile confidence resulting
from the tran-sition from underground testing to the stockpile stewardship
and management program (SSMP), Sandia Director Robinson noted:
- "Because we cannot guarantee how successful SSMP
will be, it follows that I cannot quantify any reduction in stockpile confidence
that might eventually result. What I can say, however, is that the kinds
of stockpile problems that would lead to such an erosion in confidence
seem unlikely during the next 5 or 10 years. Beyond that time frame, however,
as the base of experienced individuals disappears, it becomes much more
difficult to make any predictions." (9)
- In the light of such testimony by highly qualified individuals with
detailed nuclear weapons knowledge, one wonders why Dr. Bailey does not
feel similarly constrained from making dire predictions that the SSMP will
inevitably prove unable to maintain a safe and reliable stockpile.
- Can't Remanufacture of Warhead Components Maintain Reliability?
- Why can't measures other than nuclear testing surveillance
of the stockpiled weapons, non-nuclear testing of materials and components,
and rebuilding of aging weapons --reveal problems and provide high
confidence solutions. To some extent, they can and already have. However,
we have learned from experience that weapons in the U.S. stockpile can
have design flaws or problems that are introduced as a result of field
handling. A particularly difficult problem to address is what nuclear testing
experts call the "unknown unknown" the unanticipated problem
that is exposed only by the extreme stresses encountered in the environment
of a full-scale nuclear test. (p.9)
- It would seem that rebuilding warheads regularly to replace their
parts and materials would correct age-related problems that develop in
nuclear warheads. Indeed, Russia's approach to ensuring reliability depends
on rebuilding; it produces thousands of weapons per year to replace aging
warheads in its inventory." (p. 10)
- Dr. Bailey seems to take the odd (and indefensible) position that warhead
surveillance and component remanufacture will work in a technologically
lagging and nearly bankrupt Russia, but cannot be made to work in the U.S.,
the richest and most technologically advanced nation in the world. Moreover,
Russia is no longer producing "thousands" of warheads per year
to replace aging warheads in its inventory. The actual number is closer
to a few hundred, and this number will further diminish with time as Russia's
budget crisis continues.
- In the case of the U.S. arsenal, rebuilding warheads to ensure reliability
is not currently an option. Some components and materials are no longer
available, and there is no way to duplicate them. It may not be possible
to determine, in the absence of nuclear testing, what the functional equivalent
of a particular component or material is. And there is a Catch-22: Even
when new U.S. production capabilities are built, it will be impossible,
absent nuclear testing, to validate the new plants, processes, and people.
Nuclear tests are the only known means of demonstrating that new production
lines produce functionally identical products. (p.10)
- The above statements fail to distinguish between non-nuclear component
production capabilities, which are currently functioning and producing
replacement and new components for the stockpile, and nuclear component
fabrication capabilities, which have not been fully reestablished. This
is hardly cause for alarm, however. While limited life components, such
as neutron generators, tritium reservoirs, and batteries, must be replaced
periodically, the plutonium pits and uranium-lithium secondary components
have as yet indeterminate life spans, conservatively estimated to be on
the order of 50 years or more.
- There is ample time, therefore, to reestablish plutonium pit recycling
and fabrication capabilities, and in light of ongoing arms reductions and
the likelihood of further reductions, there is an advantage to proceeding,
to avoid hasty, very expensive, but potentially redundant investments in
plutonium fabrication capabilities. A limited capability for pit fabrication
has been reestablished at Los Alamos, and current efforts concentrate on
precisely the issue of concern to Dr, Bailey "certification"
of a new process for casting plutonium pits through carefully diagnosed
hydrodynamic experiments at Los Alamos and NTS. Likewise, facilities for
modification of secondary components are now operating again at Oak Ridge.
- As for the role of nuclear tests, while undoubtedly adding to confidence
in the remanufactured product, even if successful a nuclear test does not
in itself guarantee that subsequently manufactured devices will perform
as intended. They are in fact one of the least viable means for "demonstrating
that new production lines produce functionally identical products."
Statistically valid sampling, detailed inspection, and non-nuclear testing
of warhead components, including full scale and intensively diagnosed integrated
testing of the nuclear assembly system, are the best way of verifying the
"functional identity" of the reestablished product lines. Underground
nuclear explosions are too expensive, and often provide too highly integrated
a result, to be of much use in identifying the root cause of any system
- Is the CTBT verifiable?
- ...any adversary that covertly tests while the United States
forgoes testing can gain significant military advantage.
- ...two key questions must therefore be addressed: what is the minimum
yield of a nuclear test that can provide militarily significant information,
and can the CTBT verification system detect tests at that level.
- ...Testing at any yield, regardless of how low it is, may provide
militarily significant information to a proliferator and, perhaps, to and
advanced nuclear weapons state. (p.11)
- Clearly, this must be regarded as an extreme position. Bailey glides
easily between "significant military advantage" and "militarily
significant information," but they are not the same concept. Given
the advanced state of U.S. nuclear weapon capabilities, it would take much
more than a few covert tests for any state to gain "significant military
advantage" over the United States. Did India or Pakistan gain significant
military advantage over the U.S. or any other nuclear weapon state as a
result of their recent tests? Can anyone honestly say that a covert nuclear
test by Russia today would confer upon it a significant military advantage
over the United States? It is one thing to acquire nuclear weapons information
through one or a few covert low yield tests. It is another matter entirely
to exploit that information in a manner that preempts timely detection
of production and deployment and a deterrent response, and therefore confers
a meaningful military advantage.
- The International Monitoring System (IMS) will not be able to detect,
with any significant degree of confidence, nuclear testing below one kiloton.
If the test is evasively conducted, the system will not detect a test of
- These misleading assertions deserve a closer look. First, there is
recent demonstration that the IMS will be able to detect and identify non-evasive
explosions of less than 1 kiloton in some strategically important areas.
In August 1997, a small seismic event was detected in the area of Novaya
Zemlya, Russia's primary nuclear test site. At first, ambiguous preliminary
U.S. seismic intelligence data, along with simultaneous preparations for
underground "subcritical" experiments at the Russian site, suggested
that this event was a secret nuclear test. Additional data from stations
in the still incomplete IMS quickly served to more precisely locate the
event well off-shore in the Kara Sea, and revealed seismic signatures characteristic
of an earthquake, not an explosion. Had this been an underground nuclear
test, its magnitude (3.3) would have corresponded to a yield of less than
100 tons (0.1 kilotons) in the absence of evasive measures. A nearby event
identified as an earthquake in January 1996 was a factor of ten smaller
(2.4), corresponding to a yield of about 10 tons. This is much better than
the minimum expected global seismic event detection and identification
capability that Bailey and others have often mistakenly characterized as
the best that can be achieved (i.e.1 kiloton non-evasively tested, magnitude
4). It is simply not cost effective or sensible to attempt to achieve a
uniformly low seismic threshold for all areas of the globe and against
all "possible" evasion scenarios, no matter how implausible these
may turn out to be in practice. Such a capability would result in a huge
increase in the number of seismic events detected in low threat regions,
covering most of the globe, that would require further processing and identification.
- Moreover, the seismic capabilities to monitor the CTBT go beyond the
IMS. In addition to the IMS's 170 seismic stations, for example, there
are more than 10,000 other seismic stations providing dense regional coverage
-- in many cases providing a better detection capability than the IMS.
Also, the United States is deploying enhanced capabilities as part of its
own national intelligence means that will exceed those of the IMS in important
- Finally, the capabilities of the verification regime go beyond seismic
monitoring to include hydrophone, infrasound, and radioactive debris detection
capabilities, voluntary confidence building measures (such as voluntary
site visits and cooperative transparency measures) on-site inspections
at the request of 3/5 of the treaty's Executive Council, and the ever present
possibility of human intelligence and leaks regarding clandestine testing
- The most likely cheating scenario may be an underground nuclear
explosion in a cavity. That would muffle the energy, reducing the blast
signal by as much as a factor of 70. Thus, a 1 kiloton explosion could
be made to look seismically like a 14 ton explosion; a 5 kiloton explosion
could look like a 70 ton explosion. (p. 12)
- The effects of "decoupling" are well documented. For 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. The apparent seismic yield was only 5.3 tons,
a reduction by a factor of 71.7. (Footnote 12)
- How Probable Is Successful CTBT Evasion via Cavity "Decoupling"
- This synopsis is seriously misleading. Much more can be said about
the prospects for successful evasion via "decoupling." In reality,
the prospective evader would be very far from certain that cavity-decoupled
explosions as large as several kilotons would escape detection and subsequent
identification as nuclear tests. This is particularly true for states with
little or no nuclear test experience. Cavity decoupling is a major technical
undertaking, requiring specialized knowledge and equipment and hundreds
of skilled personnel. Decoupling even low-yield explosions would a difficult
and uncertain task.
- To begin with, even so called "full decoupling" means not
the absence of a seismic signature, but a difficult-to-predict reduction
in the seismic signature, ranging from the experimentally observed maximum
of a 70 -fold reduction at the low frequencies (< 6 hz) that propagate
long distances, down to only a factor of seven reduction (at 20
hz) for higher frequency waves propagating over regional distances (this
significant detail is omitted from Bailey's analysis).
- Far from being "well documented," as Bailey asserts, these
"full decoupling" factors are based on analysis of the 380 ton
Sterling event and the even smaller Diamond Beech/Mill Yard tests in volcanic
tuff at the Nevada Test Site. Whether these results hold for larger decoupled
explosions on the scale of kilotons in correspondingly larger cavities
and different geologic media is simply not known, and will be obviously
very difficult to establish once the test ban goes into effect. Data from
a 10 kiloton partially decoupled explosion carried out by the former Soviet
Union in a salt dome, now located in Western Kazahkstan, indicate that
teleseismic signal amplitudes were reduced only by a factor of ten relative
to a tamped (i.e. well coupled) explosion of similar yield, suggesting
that the decoupling factor drops off rapidly if the explosion is too large
for full decoupling within the cavity.
- Hence, the postulated "full decoupling" factors of 7-70 (depending
on the frequency) are obtainable only if the yield of the device and the
size and shape of the cavity are appropriately matched. In the proliferation
evasion scenario, the yield of the explosive device will not be known in
advance with any degree of precision, and thus the appropriate match with
cavity size cannot be assured, making it likely that only partial or possibly
no decoupling will be achieved.
- Indeed, based on Russian and U.S. test experience, such a scenario
involves a substantial risk of containment failure, leading to discovery
of the test through large-scale venting of radioactive gases into the atmosphere.
The United States and Russia developed their containment technology largely
through trial and error based on repeated experiences at well-studied test
sites. A prospective evader without extensive nuclear test experience would
have to consider that the probability of getting it right the first time,
at an untried location, is at least highly uncertain and probably quite
low. This risk was demonstrated recently when one the recent low yield
Pakistani tests vented radioactive debris into the atmosphere which was
subsequently collected by U.S. monitoring aircraft. Indeed, the use of
a new site raises the risk of containment failure even when the decoupling
is attempted by an experienced nuclear weapon state.
- In the event the seismic signal from a cavity explosion of a few kilotons
is only partially decoupled, it would readily be detected and discriminated
from earthquakes of comparable magnitude, so the penalty for not doing
the job exactly right is severe. Human intelligence, data from national
technical means, and on-site inspections, would pose a substantial risk
of further identifying this "point-source" explosion as being
nuclear rather than chemical in origin.
- Even if containment failure can be avoided, there is still the problem
of preventing the delayed leakage or "seep" of radio-isotopes
from the cavity due to barometric pressure. Radioactive contamination from
the cavity may also finds its way into the water table, where it can be
detected through drilling conducted during an on-site inspection, and the
cavity must also be protected against subsequent collapse to avoid creating
a telltale subsidence crater at the surface.
- Due in part to its lack of joints and fractures which might allow seepage
of radioactivity from the secret blast to the surface, domed or thickly
bedded salt is the preferred medium for decoupling. However, the locations
of major salt deposits capable of sustaining a cavity explosion on the
order of one kiloton or more are known.
- Creating a decoupling cavity is a delicate balance between the minimum
depth required to fully contain the cavity-generating explosion, and the
maximum depth at which a stable cavity of the desired size can be sustained
without being crushed by the pressure of the overburden. If solution mining
techniques rather than an explosion were employed to create the cavity,
tens or hundreds of millions of gallons of brine would have to be disposed
of without drawing attention to the project. And given that the hydrostatic
pressure of the brine in the cavity helps to support the weight of the
overlying rock, evacuation of the brine from the cavity in preparation
for the test would make the stability of the cavity highly uncertain. Existing
solution mined cavities are elongated and irregular in shape, creating
the likelihood of an unpredictable and possibly sharp reduction from the
nominal "full decoupling" values assumed in Bailey's paper.
- As a consequence of these and other difficulties, such as concealing
or masking such preparations from the prying eyes of U.S. intelligence,
the CTB inspection system, and one's own citizens, most experts agree that
attempts at full decoupling are completely impractical for yields above
a few kilotons and highly uncertain at any yield for nations with no experience
with conducting underground nuclear detonations. For example, the full
decoupling of a 5 kt explosion in a salt dome would require a deeply buried
cavity with a minimum diameter of 86 meters (282 ft), big enough to contain
the Statue of Liberty on its pedestal with room to spare. And yet Bailey
tosses off the possibility of decoupling a 5 kiloton explosion as though
it were just another walk in the park.
- Bailey raises the possibility that a nation might conduct a test in
the open ocean, where it would be relatively easy to detect but difficult
to determine who did it. She is right about that; it is quite possible
that the international community would be unable at first to attribute
an open ocean explosion.
- But this does not mean that there would be no risk to the violator.
Once detected, the US and other nations would go to great lengths to figure
out who did it. Secret operations are always vulnerable to leaks, both
before and after the fact. The movements of naval vessels and even commercial
shipping around the world are extensively monitored by the U.S. and other
Naval intelligence services, and by ocean reconnaissance satellites, and
all their communications are likewise subject to monitoring. Hence the
support vessels involved in such an operation are potentially subject to
detection and surveillance. Any debris from the bomb's platform (boat,
submarine, etc) not vaporized in the blast might be traceable back to the
point of origin. There is also some possibility that the characteristics
of the radioactive debris from the nuclear explosive could be traced back
to the maker.
- There is only one reported case of a publicly unattributed open ocean
explosion, the still-unresolved 1979 event in the Indian Ocean. If this
method of evasion is so easy, why haven't we seen more of it? And if nations
have not had reason to use this method in the past, why would we start
to see more of it under the CTBT? We will not. And even if we did, the
US could reconsider its adherence to the Treaty, as could other nations
that have greater reason to feel threatened by such developments than the
- This paper largely misses the point on verification. One hundred percent
certainty is not the goal. The goal of any verification system is to deter
all violations of the treaty, while assuring detection of violations that
would deprive a party of the security benefits it derives from the compliance
of the other parties. Thus the probability of detecting violations must
be high enough so that potential violators will believe that the risks
of being found in violation outweigh the expected benefits of the illegal
act. Conversely, the law-abiding nations must be convinced that the security
risks posed any by undetected violations are substantially less than the
security benefits of the treaty.
- The CTBT verification system meets these conditions. In general, the
higher the yield of a nuclear test, the greater the chances that it will
be detected and identified as a nuclear explosion. Above a few kilotons,
these probabilities are high irrespective of location and the mode of emplacement.
It is true the treaty's verification system will not have a uniformly high-confidence
capability worldwide to detect and identify very low-yield nuclear tests
(i.e. those ranging from a few pounds to a few hundred tons). However,
at the higher end of this spectrum, events with seismic yields of tens
to hundreds of tons can be readily detected in many areas of the globe,
but discrimination of potential nuclear events from large point-source
chemical explosions remains a problem. The treaty addresses this issue
by providing for voluntary notifications and exchanges of data regarding
the conduct of large chemical explosions, and by mandating procedures for
the conduct of on-site inspections of the vicinities of suspect events.
Tests below a few hundred tons do not permit an adequate assessment of
deuterium/tritium boosting, a major performance indicator in advanced nuclear
weapons. A full yield proof test of a boosted primary for a two-stage thermonuclear
device requires a test in the range of several kilotons or higher.