No Serious Problems: Reliability Issues & Stockpile Management


Reliability Issues and Stockpile Management

A Review for Tri-Valley CAREs
by Greg Mello
February 6, 1995

  • The U.S. nuclear arsenal has been extensively tested and is reliable. 
  • The arsenal is becoming more reliable each year as older weapon types are retired. 
  • The weapons which are to remain in the stockpile can be expected to last many more years if serviced.  So far, age has tended to make nuclear weapons more, not less, reliable, and no weapon has ever been retired due primarily to aging problems. 
  • Although the nuclear arsenal is increasing in reliability, confidence in nuclear weapons is declining as nuclear weapons lose political importance.  The two issues -- reliability and confidence -- should not be confused. 
  • U.S. nuclear weapons are adequately "robust" with respect to the minor variations in manufacture that have occurred throughout the prototyping and production processes. 
  • There is no serious technical impediment to remanufacturing existing kinds of weapons and weapon parts.  For example, Los Alamos is now gearing up to remanufacture plutonium pits. 
  • Changing the design of nuclear weapons without nuclear proof-testing has been the source of serious reliability problems in the past and can be expected to cause more problems if it is done in the future. 
  • The present, effective process for monitoring the reliability of nuclear weapons places only small demands on weapons research and development personnel. 
  • A high level of weapons reliability can be assured with modest levels of funding and with current diagnostic facilities.  New facilities, especially the $2 billion National Ignition Facility, are not needed.  Engineering and production skills, rather than scientific research, are most essential to maintaining and remanufacturing nuclear weapons. 
  • The only purpose of reliable nuclear weapons is to deter a nuclear attack.  Efforts to continually refine nuclear weapons conflict with U.S. nonproliferation goals and could well degrade confidence in the overall efficacy of our deterrent. 


In the debate over a nuclear test ban in the late 1980's, advocates of continued testing, led by the two weapons physics laboratories, took the position that the reliability of U.S. nuclear weapons would decline unacceptably under any kind of testing ban.  Specifically, weapons labs managers argued that changes in manufacturing techniques, materials, and personnel would seriously degrade confidence in the U.S. nuclear deterrent.

These concerns were not shared by everyone in the weapons design community, however, and the lab managers' view, when weighed against contrary technical evidence from senior independent and retired experts, failed in the end to convince even the Pentagon.  All relevant agencies of the government -- including the managers of the labs -- now agree that it is feasible to maintain a nuclear deterrent without nuclear testing.  While not everyone in government wants a test ban -- John Deutch, Undersecretary of Defense does not, for one (1) -- everyone now finally admits, either implicitly or explicitly, that the U.S. nuclear stockpile can be maintained under a comprehensive test ban treaty (CTBT).

Indeed, the current leadership at the Pentagon "requires" that the nation retain the capacity -- even under a test ban -- to design, fabricate, and certify new weapons, a much more difficult task than simply refabricating existing weapon types. (2)   With a CTBT now an official goal, reliability issues, like safety concerns, are again being raised, this time to promote the new facilities and large appropriations proposed for the Department of Energy's (DOE's) expansive "science-based stockpile stewardship" (SBSS) program.  Are these concerns real?  Is it necessary to keep research, development, and testing (RD&T) spending at Cold War levels to maintain reliable weapons?  Are major new facilities needed?  This paper, part of a series on these and related questions, suggests that reliability concerns can be addressed in a much simpler and relatively cost-effective manner through more effective management of the nuclear weapons complex. (3)

Congressional testimony shows that the U.S. arsenal has no significant current reliability problems.

The best starting point for this discussion is the most authoritative testimony available.  Speaking before Congress on March 15, 1994, Dr. Harold Smith, Assistant Secretary of Defense for Atomic Energy, said
I am pleased to report the stockpile today is safe, secure, reliable, and meets current military requirements.  We make that statement with confidence today and for the immediate future....Our stockpile is becoming safer and more reliable simply because we are retiring older weapons...Thus, we should enter the 21st century with a modern, safe, and reliable stockpile consistent with the demands of START I and with anticipated military requirements. (4)
So not only are U.S. nuclear weapons reliable, but they are, at least for the "immediate future," becoming even more reliable as older weapons are retired.  This statement was made in the presence of, and with the approval of, Dr. Victor Reis, Assistant Secretary of Energy for Defense Programs, who added,
Right now, as Dr. Smith said, that stockpile is safe and reliable. (5)
Make no mistake:  nuclear weapons in the stockpile are certified to explode at nominal yield after experiencing all the environments that a deployed weapon could reasonably be expected to weather in its life, and after the severe accelerations, cold, heat, shock and radiation (e.g. from the explosions of other warheads) that are expected to occur during the stockpile-to-target sequence.  This performance is exhaustively analyzed and tested in a diverse array of demanding tests coordinated primarily by Sandia, whose mission is to weaponize the physics packages (6)   developed at the other two laboratories.  It appears that most, or perhaps even all, of the weapons in the current stockpile have been tested in underground nuclear tests using not only prototypes but also actual production weapons, with simulated end-of-life and stockpile-to-target conditions.(7) So the reliability concerns that have been raised are concerns over the long term. What, in this context, is the "long term?"  And just how serious are these concerns?

No serious reliability problems are expected in the arsenal for many years, even without remanufacture.

Dr. Dan Kerlinsky investigated in detail the subject of reliability in relation to stockpile stewardship throughout the summer and fall of 1994 for the Secretary of Energy's Panel on the Future of the DOE Labs (the Galvin Panel).  In the process, he attended and initiated numerous classified discussions and interviews with senior scientists and managers at all three weapons laboratories.

Kerlinsky found that data on historical reliability and repair had been systematically collected by Sandia National Laboratory in a two-year study commissioned by Secretary Watkins, called the Stockpile Life Study, which was completed in 1994.  This study was designed to answer two basic questions:  (1) "How long do nuclear weapons last?" and (2) "What programs and activities are required to keep them in the stockpile?"  To answer these questions, the detailed history of the entire stockpile was examined over the past thirty years. (8)

Sandia was in an excellent position to gather this data, as it is the laboratory that coordinates the stockpile surveillance program for the Department.  But in order to be sure that the Sandia study did not leave out any problems that might be known to Los Alamos and Livermore but not to Sandia, Kerlinsky interviewed weapons program personnel at these two laboratories as well.  He found that the Sandia data was indeed comprehensive. 

The thirty years of experience summarized in this study revealed that there is not known to be any upper limit on weapon life, given appropriate maintenance and renewal of perishable materials and parts (e.g. tritium).  No U.S. weapon has ever been retired due primarily to aging problems, even though some weapons have, in the past, been in the active stockpile for more than 30 years before being superseded by new designs.  (The weapons in the present stockpile are, in contrast, believed to be all between about 6 and 15 years of age.)  Aggregate data show that the rate of required modifications and repairs of stockpiled weapons decreases as the years go by, reflecting continually increased reliability as the "bugs" are gradually worked out of weapons systems.

Fully half of the 300-odd problems encountered in the stockpile over the years were judged to have no effect at all on the reliability of the weapon.  At least another 40% of the problems investigated would have had only a slight impact on reliability.  Only a small minority -- less than 10% -- of the problems found would have led to a decrease of 10% or more in the probability of achieving the certified yield.  All of these problems have been fixed.  Many of them were caused by rushing inadequately-tested weapons into the stockpile, as will be discussed below.

Overall, among the defects found, very few have been in the physics packages, which typically are the simplest part of the weapon, comprising less than 5% of the overall number of parts in the B83, to pick one example. (9)

What is most important to realize is that the weapons in the enduring stockpile have been designed and built using all this and more accumulated experience.  Design and production mistakes that occurred with earlier weapons have not been repeated.  The stockpile is more reliable than ever before, and is increasing in reliability as older systems are retired.  Before examining this subject in greater technical detail, it is important to step back and look at the difference between reliability of, and confidence in, the stockpile.  Reliability is sought only for the purpose of creating a convincing deterrent.

It is important to distinguish between the reliability of a warhead and confidence in a nuclear deterrent.  Reliability is a technical problem.  Confidence is based on reliability, but it is also based on psychological, social, and political realities and perceptions.  Confidence is not at all the same to every observer.  For most decision makers, confidence in a nuclear deterrent is the genuine objective of nuclear weapons expenditures, not simply reliability.  They seek this confidence in an attempt to allay fears of a nuclear attack, and they attempt to project this same confidence to would-be aggressors and potential enemies in our foreign wars.

Confidence is eroded whenever the experts at the nuclear weapons labs, who have a strong financial and institutional interest in the matter, say there is a reliability problem of any kind, no matter how trivial or even false. So the scientists at the weapons labs, along with their sponsors at the DOE and DOD and the consultants who serve them, are presently in a position of considerable influence over confidence in the U.S. stockpile, quite independent of actual weapon reliability.

In a different and perhaps more fundamental sense, confidence in the stockpile is also eroded whenever the utility of warheads to accomplish national goals -- including deterring non-nuclear aggression -- is questioned.  This is now occurring with greater and greater regularity.  Many people believe that the 49-year de facto norm against the use of nuclear weapons, together with the repugnance with which the nations of the world would regard their use, make that use quite unrealistic and counterproductive for any purpose other than the narrowest one of deterrence of nuclear attack. (10) In a very real sense, nuclear weapons are becoming less reliable as instruments of policy and power.  They are becoming politically unreliable--which makes everyone more secure.

In this context, senior bureaucrats, members of Congress, defense officials in the executive branch, and others who have focused years of their life on the production, maintenance, and deployment of nuclear weapons sympathize readily with the fretting about long-term reliability now being done by the laboratories and their sponsors.  Confidence in the ability of U.S. nuclear weapons to explode is not in any serious doubt, but the funding, purpose, direction, and meaning of the U.S. nuclear weapons program is very much in doubt, and the two kinds of doubt are very easily confused.  In any case, perfect reliability is not essential to deterrence.  The technical -- as opposed to the political and psychological -- requirements of deterrence consist only of providing any aggressor with enough confidence that U.S. weapons might explode to deter his attack. (11)   A more detailed technical analysis of the reliability question does not reveal any serious problems.

The "reliability" question was authoritatively examined in a 1987 report to Congress by Dr. Ray Kidder, a senior Livermore weapons physicist, commissioned by former Congressman (and later, Secretary of Defense) Les Aspin. Kidder's overall conclusion was,
It is found that a high degree of confidence in the reliability of the existing stockpile is justified, and that it is sufficiently robust to permit confidence in the reliability of remanufactured warheads in the absence of nuclear explosive proof-tests.(12)
Note that this confidence was warranted in 1987, without the new facilities that are proposed for the SBSS program.

In the course of his study, Kidder found that in the entire history of the U.S. stockpile, only one weapon (the W68) experienced reliability problems after being in the stockpile more than four years. (13)   The Stockpile Life Study's conclusions, as we have seen, confirm this vote of confidence and bring it up to date.

In his paper, Kidder analyzed the fourteen often-cited cases of stockpiled weapons that required nuclear tests to evaluate problems. Nine of these occurred in the 1960's -- Kidder calls these the "Sixties Nine" -- as a direct result of the rush to build and stockpile weapons during the 1958-1961 test moratorium.  That rush led "to a stockpile that was very poorly tested by today's standards," he notes.  Kidder continues,
Our understanding of how nuclear weapons work, our experience with nuclear tests, and our computational capabilities were all significantly inferior to that which exists today [i.e., in 1987].  There has been no rush to build the present stockpile, and it has benefitted from a quarter-century of additional nuclear and nonnuclear tests since the hectic days of the Moratorium.  For these reasons, it is concluded that experience with the Sixties Nine, long ago, has little or nothing to say about the reliability of the stockpile of nuclear weapons that exists today. (14)

Five other actual or potential reliability problems occurred in the 1980's, likewise as a result of inadequate testing.  In two cases, the production weapons were different than the tested prototypes and had never actually been tested at all (one worked; one didn't).  In two cases, weapons had never been tested after the rugged stockpile-to-target sequence of events (one worked; one didn't).  And in one case, the high explosive in a weapon was changed; the explosive -- unique to that failed weapon -- had never been adequately tested in the first place and didn't work properly.  All these problems, and the lack of testing that allowed them to go unnoticed prior to production, are irrelevant to the current stockpile.

Carson Mark, Director of the Los Alamos Theoretical Division for 26 years, summed up the situation in 1993 by saying that the reliability argument, as it was then being used against a test moratorium, was "utter rubbish." (15)   Mark links reliability and stockpile confidence to preservation of the original warhead design, without change or "improvement."

Most important, no changes or improvements in matters which could affect the behavior of the nuclear system in an existing, certified weapon design would be acceptable [during a test ban]. (16)
Mark makes the further point that the real threat to confidence and reliability of the arsenal is the historical insistence that, to quote an old policy of the DOE, "the nuclear weapons stockpile must dynamically evolve to satisfy changing threats and deterrent requirements." (17)    Reliable warhead remanufacture is feasible.

The concept of maintaining confidence by preservation of design is at the heart of "warhead remanufacture," to which we now turn.  Kidder tabulated the results of U.S. nuclear tests, dividing them into four groups, and found that the agreement between predicted and actual yield was "remarkably" good.  He concludes that
this impressive record would not have been possible if U.S. nuclear weapons were not comfortably tolerant of the small variations in materials and manufacturing that accompany any practical production process...The test record indicates that the nuclear weapons in the existing U.S. stockpile are sufficiently robust to allow for future replication...This conclusion is in agreement with earlier statements by nuclear weapons authorities Dr. Hans A. Bethe, Norris E. Bradbury, Richard L. Garwin, J. Carson Mark, and Andrei Sakharov affirming the possibility of reliable remanufacture without nuclear explosive proof-tests. (18)
This conclusion is reinforced by an experiment at Los Alamos in which an intentionally off-spec pit was made and successfully tested in Nevada. (19)

Kidder appends the convincing statements of these authorities to his paper.  Needless to say, when he wrote these words in 1987 Kidder did not, nor could he have, made them contingent upon the existence of new science-based stewardship facilities.

Kidder added, however, that for later remanufacture to be feasible, original materials would have to be used.  If changes were made, confidence would be undermined.  When I spoke to him last year about this, he did not think that material substitution would be a problem in each and every case.  Some changes could, in fact, be tolerated:  parts which are supposed to be transparent to radiation, for example, could be replaced with parts which are at least as transparent, and parts which are supposed to be opaque, with parts at least as opaque.  He noted that since his 1987 study, there have been further classified studies of this issue, including a "Stockpile Remanufacture Study" in FY91.  At the present time, he is "not aware of any problems" with warhead remanufacture, provided adequate funding is available. (20)

Mark agrees.  On the materials question, he says
it is ridiculous to suppose that substitutes would ever be needed.  Nuclear weapons production has never been dependent on commercial supplies or suppliers to meet its needs for obtaining and handling separated isotopes, polonium, plutonium, or explosives.  The production system can certainly be set up, or arrange, to acquire safely any needed supplies of beryllium, plastics, asbestos, or whatever else may have served effectively in past production even if in the future it should be dropped from commercial use because of hazards or lack of demand.  (21)
As for the level of expertise needed to implement such a remanufacture program, Mark points out that the only real need for experienced weapons designers would arise in the determination of "whether a particular problem found in the surveillance program did or did not require replacement of the stockpiled weapons with new ones built to the original, certified, and tested specification."  Mark, for one, has no doubt that if the problem were considered important enough, a body of experienced experts could be convened regardless of the level of staffing at the weapons laboratories. Mark goes on to say,
Past experience casts doubt on our supposed utter dependence on maintaining a corps of scientific and engineering veterans of the Nevada test site.  There were no such persons anywhere in 1943 when the effort to build an atomic bomb was begun; and in 1950 when official hydrogen bomb efforts were announced, the corps of seasoned experts was very small and had very meager experience in testing - now said to be the sine qua non of capability.  Yet in each case, the United States succeeded in developing and testing a weapon based on an entirely new physical principle within two to three years. (22)
Kidder points out that the institutional arrangements needed to maintain reliability are "nothing new."  It would, in his opinion, take only relatively small teams, with focused responsibility for each weapon system, to resolve stockpile questions.  At present, he said, it is rare for weapons designers to receive questions regarding changes in weapons specifications.  Before this happens, several levels of expertise -- the inspectors and production engineers at Pantex, then their supervisors, then the weapons engineering teams from the weapons labs which regularly travel to the production plants -- must be exhausted first.  He believes there is no reason to change this system, which works.  The highest priorities for a strong remanufacture program would probably be product and production engineering, test equipment, and so on, rather than new science per se, although a scientific base must also be maintained, including hydrotesting facilities. (23)

In 1987 Kidder concluded, long before SBSS, with its proposed panoply of new design and diagnostic facilities, appeared, that
The robust character of the nuclear weapons in the present stockpile, together with the ample time available to accomplish the task, suggests that it will eventually be possible to be confident of the reliability of remanufactured nuclear weapons without requiring the services of nuclear weapon design engineers and scientists that have themselves benefitted from direct experience with nuclear explosive tests. (p. 9)

At the same time -- 1987 -- as Kidder conducted his study, managers at Lawrence Livermore National Laboratory (LLNL) were writing their own report. (24)   This report -- a cri de coeur for nuclear testing -- is very useful as a compendium of problems that had been experienced in weapon design and manufacture.  Based on the historical record, these managers believed three kinds of problems would occur in a future without nuclear testing:  the loss of both scientific and production experience, unavailability of materials (and/or subtle variations in them) with no way to test how substitutions and changes affected performance, and inadequate documentation and specifications for many materials and processes, leading to inadvertent production changes.

Yet the official LLNL conclusions of 1987 -- that exact remanufacture is impossible and that nuclear testing is necessary to retain the reliability of the arsenal -- have both now been widely rejected.  In the end, data like that of Ray Kidder have been strongly persuasive over the attempt to extrapolate future problems from problems discovered -- and corrected -- in the past.  For if minor variations in manufacture, materials, or specifications were indeed a serious problem, we just would not see what Kidder called the "impressive" testing record in Nevada, either for new primaries or stockpiled weapons.  Especially given the near-perfect record for new primaries, why should it be so difficult to merely maintain existing designs? (25)

The JASONs believe the best approach to reliability is exact remanufacture.

In their recent report on SBSS, the JASONs -- an elite group of academic defense consultants convened by the MITRE corporation -- endorse exact replication of weapons, especially plutonium pits, as the best way to ensure confidence in the arsenal.
...the primary -- if not the sole -- nuclear weapons manufacturing capacity that must be provided for in an era of no nuclear testing is the remanufacture of copies of existing (tested) stockpile weapons...the ultimate goal should be to retain the capability of remanufacturing SNM [special nuclear materials] components that are as identical as possible to those of the original manufacturing process and not to "improve" those components.  This is especially important for pits... (26)

The plutonium pit forms the core of the fission primary, and its function is critical to the reliability of the entire weapon.  Pits are composed of various materials which must be alloyed, shaped, and joined with great precision.  Even so, the JASONs point out that it is the finished pit that must be the same as the proof-tested model, and not every manufacturing detail or process along the way.  To ensure this, they call for "a narrowly defined, sharply focused engineering and manufacturing curatorship program." (27)   This is an excellent approach to stockpile management overall, but the point here is that they found no serious obstacles to their exact remanufacturing proposal.  That is, the inevitable inexactnesses of "exact" remanufacture evidently appeared quite tolerable to the JASONs, contrary to what is often said by the laboratories.   Remanufacturing is not only practical -- it is about to start.

In addition to alleged reliability problems caused by declining expertise, unavailable materials, and inadequate documentation, it has more recently been suggested that there will be reliability problems associated with the adoption of novel production methods.  For example, plutonium metallurgy is complex, and the processes formerly used at Rocky Flats to make pits are different in some respects from those now being set up at Los Alamos to do the same job.  But Los Alamos is remanufacturing pits.
Associated with pit surveillance activities is the Pit Rebuild Program, which will demonstrate the capability at Los Alamos to fabricate pits of war reserve quality.  Specific technology areas that must be developed or enhanced at Los Alamos include certification of the beryllium machining capability, certification of the tubulation capability, development of the capability to interface pit materials, and development and certification of joining processes.  The capability for manufacturing a pit for a W88 weapon will be in place in FY 1996...
Los Alamos is currently the only DOE site capable of fabricating a plutonium pit and, as such, may be the only practical near-term upgrade-in-place option for plutonium processing and fabrication.  In conjunction with its plutonium research and development responsibilities, Los Alamos will maintain those technologies and capabilities required to build plutonium pits for development and demonstration purposes...(28)

Thus the capability to produce W88 pits for the stockpile will be put in place long before any of the so-called "necessary" new diagnostic facilities, such as the Dual-Axis Radiographic Hydrotest (DARHT) facility, are available. (29)   While both DOE and the laboratories are trumpeting exotic solutions to alleged reliability "problems," a fully-certified production line is being set up this year for the most critical and difficult-to-manufacture component in a weapon: the plutonium pit.

What is more, much of the slow start-up in the Pit Rebuild Program results not so much from the difficulty of being sure the pits produced will be good enough per se, but from the painstaking work of matching or modifying the specifications and procedures -- the official cookbook -- that LANL inherited from Rocky Flats.  It is the quality assurance aspect that takes the most time.  Asked how long it would take to produce a working pit for the stockpile in an emergency, independent of the formality of the certification process, LANL managers told Kerlinsky, "Three days." (30)

But this is not all --
As part of the Stockpile Support Program, Los Alamos will maintain a capability to make the components for a complete nuclear physics package, thereby ensuring that the requisite technologies and expertise are retained in the DOE complex and that weapons RD&T requirements can be met in the future.  Because upgrades in canned subassembly (CSA), assembly, and radiation-case capabilities will be required, Los Alamos and Livermore are preparing a study for DOE/AL to evaluate joint capabilities in both pit and CSA manufacturing. (31)
Thus LANL at least believes that it can manufacture not only pits but entire physics packages for RD&T purposes, if not also for the stockpile.  And the two labs together not only have confidence in, but are actively seeking, manufacturing capability for complete physics packages for the stockpile(32)   Thus, what the labs said just a few years ago couldn't be done for all the reasons they trumpeted, they are now actively promoting -- and at their own facilities, no less.  To the extent that nuclear weapons are to be retained, reliability and quality assurance issues will always require attention.  But the inescapable conclusion from the labs' current programs and proposals is that they do not believe that reliability issues stand in the way of remanufacture of nuclear weapons.

There is no need to retain weapons in the stockpile long after their expected life.

It is often mistakenly asserted that, with an end to nuclear testing, weapons will have to remain in the stockpile for unprecedented periods of time.  There is no reason why this need be the case.  About six weapon types could remain in the "enduring stockpile" for a long time, should further disarmament measures fail, but individual weapons can be remanufactured if it is decided to do so.  Therefore there is no need to endlessly study how problems which develop only after decades in the stockpile might affect reliability -- if reliable nuclear weapons are desired, new parts, or new weapons, of existing types can be made.  This logic applies to the aging of insensitive high explosive (IHE), to metallurgical changes in plutonium pits, and in fact to essentially all potential long-term stockpile problems.

Stockpile problems are now detected by a variety of non-exotic means, such as physical inspection, static radiography, chemical analysis, and testing of subsystems.  These problems can be simply remedied, if necessary, by replacement of the affected part from inventory or from remanufacture -- or by retiring the weapon.


Retention of a nuclear deterrent is not a demanding technical requirement.  Such a task is not a major scientific undertaking at all.  It is rather an engineering and organizational problem with scientific components.

In particular, there is no need to construct advanced new diagnostic facilities.  While these facilities could, in some cases, marginally advance the state of U.S. nuclear weapons science, to the extent they do this they will inevitably undermine the global nonproliferation regime. (33)

Thus, while contributing little or nothing to the reliability of U.S. nuclear weapons, these new facilities could well decrease confidence in the overall efficacy of our nuclear deterrent.

What is more, a DOE budget centered around SBSS, with funding included for large, unnecessary, and distractive gadgets like the proposed National Ignition Facility, could overshadow and weaken those programs most useful for actually maintaining the existing stockpile.  This fact has not escaped the notice of many analysts, including the Congressional Budget Office and the Natural Resources Defense Council. (34)

Fundamentally, the scale and direction of the science-based stockpile stewardship program envisioned by the DOE and the weapons labs is not so much driven by concerns about the reliability of weapons as it is driven by concerns regarding the reliability of funding.  Creating new "products" -- i.e. new weapons designs -- in that program, an outcome which is probably necessary to maintain internal interest in the program in the long run if not also in the short run, may be the best way to create reliability problems in an arsenal which now has none. (35)

1.  The following exchange occurred between Chairman Hamilton and Dr. Deutch during the latter's presentation of the Nuclear Posture Review to the House Foreign Affairs Committee, October 5, 1994:
Rep. Hamilton:  "Do we need more [nuclear] testing?"

Dr. Deutch:  "In the near term?  How long do you have in mind, sir?"

Rep. Hamilton:  "You pick the time frame.  It doesn't make any difference."

Dr. Deutch:  "The position of the administration, I believe, is that we don't need more testing..."
After three more attempts by Rep. Hamilton to elicit candor from Deutch, the latter finally admitted:
"Mr. Chairman, you catch me personally here at a rather awkward position, and let me explain to you straight why, and that is that I have written widely on this subject before I entered government, and so what I'm trying to answer to you now is the position of the administration."
Although Hamilton questioned Deutch closely as to reasons to continue nuclear testing, Deutch provided none, even though he was personally in favor of testing.  Hamilton even suggested that safety as a reason to continue testing, but Deutch demurred, saying "It is my judgment that all the nuclear weapons that we have are adequately safe."  At no point did Deutch bring up any reliability issues -- either past, present, or possible future.

2.  See the viewgraphs used to present the Nuclear Posture Review in late September and early October, 1994, available from the Pentagon; the one quoted here is titled "DOD requirements to DOE."  The "requirements" terminology is both problematic and revealing.

3.  Greg Mello, "Redefining Stockpile Stewardship," Tri-Valley CAREs, Livermore, CA, 1995.

4.  Testimony before the House Appropriations Committee, Subcommittee on Energy and Water Development, Energy and Water Development Appropriations for 1995, Part 6, pp. 413-414.

5.  Ibid., p. 419.

6.  Physics packages: the nuclear "core" of a weapon, typically and schematically comprised of the fission primary, the thermonuclear secondary if present, related initiating and boosting devices, and the case.
7.  Ray Kidder, "Maintaining the U.S. Stockpile of Nuclear Weapons During a Low-Threshold or Comprehensive Test Ban," October 1987, UCRL-53820, LLNL, pp. 4-5.  His exact word or adjectival phrase has been omitted by the classifiers but the import of his sentence is clear.  Kidder discusses stockpile confidence tests (SCTs) more extensively on p. 3 and in Appendix B of his report, both of which have been substantially classified as well.  I do not know how many SCTs were done in the years between 1987 and 1992, when testing ceased, or whether there remain, at this point, any weapons which have not been so tested, or whether any systems without SCTs are to be retained in the START II arsenal.  It is quite doubtful that inadequately tested weapon systems would be chosen for retention in the so-called "enduring arsenal."  Even if so, the "remarkably accurate" predictability of SCTs (Kidder, p. 15) supports a high level of confidence in the reliability of any weapon (if there is one) which has not had this final, post-production, test.  The large number of nuclear tests that were available for such tests argues that if any SCTs were not done, it was because they were not considered very important.  The Nuclear Testing Moratorium Act specifically allowed testing for reliability purposes, but none has been requested.

8.  Classified and unclassified results of this study were made available to the Galvin Panel.  The unclassified summary is currently being suppressed by Assistant Secretary Reis and Undersecretary Curtis.

9.  This information is found in viewgraphs prepared by Sandia National Laboratory.

10.  Seth Cropsey of the Heritage Foundation is one such person (see "The Only Credible Deterrent," Foreign Affairs, March/April 1994, pp. 14-20.  General Horner, head of the Air Force Space Command, is another ("U.S. Should Trash Nukes, Top Air Force General Says," Albuquerque Journal, July 16, 1994).  Tom Thompson, dean of current weapons designers at Livermore, is another.  As he recently put it, "I can't think of any target for anything in our stockpile" ("Science Comes in from the Cold," Los Angeles Times, 12/22/94).

Many senior military commanders have held and/or hold the view that nuclear weapons are useful only to deter other nuclear weapons, and not for any military use; Robert McNamara listed some he has known to hold this view in the United States and Britain in a speech to the Economists Allied For Arms Reductions in New York on May 19, 1994.  Still other civilian and military defense leaders cited by McNamara have stronger views -- that nuclear weapons should never, under any circumstances, be used at all.  McNamara says he conveyed this opinion, which he has held since the early 1960's, to Presidents Kennedy and Johnson.

11.  One senior weapons manager at LANL put the whole reliability question in perspective with a pointed question, "Would an aggressor gamble that Israeli nuclear weapons [which are far less tested than U.S. weapons] won't explode?"  To be undeterred, any aggressor must have confidence that the U.S. deterrent as a whole won't work, i.e. that all, or essentially all, of the weapons launched by the U.S. will be complete duds.  It is this aggregate reliability that counts for deterrence, and the burden of proof is solidly on those who would doubt present writer included -- believe it would be a mistake to retaliate against an attack, even a nuclear attack, with nuclear weapons.  As General Horner put it,

"I just don't think nuclear weapons are usable.  I'm not saying that we militarily disarm, I'm saying that I have a nuclear weapon, and you're North Korea and you have a nuclear weapon.  You can use yours.  I can't use mine.  What am I going to use it on?  What are nuclear weapons good for?  Busting cities.  What President of the United States is going to take out Pyongyang?"

If, however, nuclear weapons are to be maintained and deployed as just another weapon of war, ready for first use against an enemy in times of crisis, a higher standard of reliability is required.  Yet even this dangerous and counterproductive posture can be readily and reliably supported by maintenance and remanufacturing of existing weapon types.

There is yet another reason why confidence in deterrence is robust relative to reliability of a weapon.  In a nuclear explosion, the maximum radius at which a given overpressure is experienced is proportional to the cube root of the explosive yield.  A five-fold decrease in explosive power thus causes only a 1.7-fold decrease in destructive radius and a 2.9-fold decrease in destructive area. (See Samuel Glasstone, The Effects of Nuclear Weapons, U.S.A.E.C., 1964, p. 127.)  Even significant declines in expected yield are unimportant outside a warfighting context.

12.  Kidder, op. cit., p. 1.

13.  This problem, arising from chemical instability in a type of HE that is no longer used, is not relevant to the current arsenal.

14.  Kidder, op. cit., p. 4.

15.  Personal communication.

16.  J. Carson Mark, "Do We Need Nuclear Testing?", Arms Control Today, November 1990, pp. 12-17; p. 14.

17.  Mark, op. cit., cites DOE, Program Status of Preparations for Further Limitations on Nuclear Testing, February 1990.  Far from being dead, this anachronism lives on, most conspicuously in the recent Nuclear Posture Review quoted above.

18.  Kidder, op. cit., p. 6.

19.  Daniel Kerlinsky, personal communication.

20.  Ray Kidder, personal communication, 11/17/94.

21.  Mark, op. cit.

22.  Ibid.

23.  Ray Kidder, personal communication.

24.  George Miller, Paul Brown, and Carol Alonso, "Report to Congress on Stockpile Reliability, Weapon Remanufacture, and the Role of Nuclear Testing, October 1987, UCRL-53822, LLNL.

25.  "During the past decade [1977-1986], new boosted primaries have been designed and developed by the weapons laboratories...performed satisfactorily the very first time they were tested, the observed yield in no case falling short of that expected by more than...The one new primary that failed was of a more complex, less predictable design than the others.   This primary was subsequently redesigned, tested, and failed again. None of the primaries in the existing stockpile employ...

"This experience demonstrates that the ability of the weapons labs to predict the performance of newly designed, as yet untested, boosted primaries of the kind currently in stockpile is indeed impressive -- there were no significant surprises.  This could hardly have been the case had these primaries been sensitive to differences that inevitably exist between the weapon configuration calculated and the weapon tested."  Kidder, op. cit., p. 25.

26.  "Science Based Stockpile Stewardship," Sidney Drell et al, November 1994, MITRE, McLean, VA, p. 81.  The chapter quoted proceeds from different assumptions than most of the rest of the report.  That report is criticized in a paper related to the present study, available from Tri-Valley CAREs, "Ask Few Questions, Get Few Answers:  A Critique of the JASONs' Stockpile Stewardship Study," Greg Mello, 1995.

27.  Ibid., p. 85.

28.  Los Alamos National Laboratory, Institutional Plan, FY1995-2000, pp. 50-51.

29.  DARHT, like the existing Flash X-Ray (FXR) and Phermex facilities at Livermore and Los Alamos, respectively, would evaluate mock pits during implosion.  DARHT, if built, is expected to be available for experiments in approximately the year 2000.

30.  Dr. Daniel Kerlinsky, personal communication.

31.  Los Alamos National Laboratory, op. cit., p. 51.

32.  The overall outlines of the labs' proposal can be found in a presentation by Larry Woodruff of LLNL to the National Security subgroup of the Galvin Panel on August 8-9, 1994.  The responsibilities of Rocky Flats, Savannah River, and the Y-12 Plant would all be moved to LANL and LLNL under this plan, with Sandia taking over for Pinellas and the Kansas City Plant.  The production capacity would be 150 weapons per year, with LANL specializing in fabricating parts made from fissile materials. 

33.  Detailed discussion of this point will soon be available from the present writer and Tri-Valley CAREs.

34.  Congressional Budget Office, "The Bomb's Custodians," July 1994, p. x; Christopher Paine, personal communication.

35.  See Defense Week, 8/15/94.  Dr. John Deutch, Undersecretary of Defense, and Dr. Vic Reis, Assistant Secretary of Energy, were and may still be pressing for development of a new warhead.  This notion differs from the military's preferred road map for maintaining reliable nuclear weapons, which does not involve reliance on an untested weapon.


^ back to top 2901 Summit Place NE Albuquerque, NM 87106, Phone: 505-265-1200, Fax: 505-265-1207