Coddling the Nuclear Weapons Complex

Christopher Paine

At the National Nuclear Security Administration (NNSA), it is as if the Cold War never ended. Despite the reduced likelihood of a major nuclear conflict, massive cuts in the deployed U.S. arsenal, and a long-standing moratorium on U.S. nuclear testing, the budget of the Department of Energy’s nuclear weapons agency has steadily grown during the last decade. The Energy Department now spends 35 percent more on the U.S. nuclear arsenal each year than it did on average during the standoff with the U.S.S.R. between 1948 and 1991 (when it spent the equivalent of $4.2 billion annually in current dollars). In addition, if the White House prevails with its fiscal year 2005 budget request and stays on the track outlined in its “Five-Year Nuclear Security Program,” the growth in spending for nuclear weapons will continue without respite.

In fact, if President George W. Bush’s request of $6.6 billion is approved, U.S. spending on “weapons activities”—nuclear weapons research, development, testing, and production, as well as administration of the nuclear weapons stockpile—will have virtually doubled over the past decade. If the administration has its way, the trend is not about to end. The NNSA has told Congress that it plans to increase spending on the U.S. arsenal to $7.76 billion by 2009.

Indulging the Weaponeers

U.S. nuclear spending has soared since a 1995 decision to adopt a complex, simulation-based “virtual testing—virtual prototyping” strategy as the paradigm for “stewardship” of the U.S. nuclear stockpile. Since then, the Energy Department has spent tens of billions of dollars on new and upgraded experimental facilities, computers, networks, code development, and computer-controlled production equipment. It remains largely unable to implement this grandiose and technologically aggressive strategy, however, as key experimental capabilities remain unfinished, mired in technical difficulties and huge cost overruns.

The intellectual premise behind this massive expenditure was that absent nuclear test explosions, confidence in the performance of new or modified nuclear weapons could only be gained from building costly new experimental facilities to generate data that would inform three-dimensional computer simulations of the entire nuclear explosion sequence, beginning with the detonation of chemical explosive and ending with the combined thermonuclear burn and fissioning of the weapon’s secondary stage. Moreover, DOE officials insisted that these simulations could be relied upon to predict the performance of stockpiled ‘war reserve’ weapons only if they had first been “validated” by successfully predicting the results of similarly complex integrated experiments conducted in the new “science-based stockpile stewardship” facilities. Examples of such code-validating tests include inertial fusion capsule gain experiments and primary stage implosions, diagnosed with highly-penetrating, three-dimensional, time-resolved radiographic imaging.

Critics at the time charged that the entire “virtual testing” strategy was needlessly complex and costly for the relatively simple task at hand: sustaining confidence in the safety and reliability of a modest but sufficient, enduring stockpile of nuclear weapons, for as long as needed. Detailed studies also noted that the large unclassified research component of the effort to model fundamental physical processes involved in generating nuclear explosions could well exacerbate proliferation of nuclear weapons knowledge.[1] Subsequent events have proven many of these criticisms well founded.

Indeed, because of technical flaws and delays in its more ambitious efforts, the NNSA is slowly reverting to an alternative method put forward a decade ago by its critics: “engineering-based” stewardship. This approach seeks to assure the ability to produce and replace non-nuclear components and subsystems that can be thoroughly tested while attempting to minimize changes to primary- and secondary-stage nuclear components that cannot be tested to nearly the same extent. In this more conservative approach, confidence in the performance of these nuclear explosive components is maintained, not through elaborate computer simulations of uncertain fidelity but through careful engineering efforts that ensure maintenance and remanufacture of weapons at current (or improved) quality levels.

These efforts rely on tried and true techniques for ascertaining that the primary’s high-explosive driven plutonium shell continues to hit the performance benchmarks required for the onset of nuclear fission and subsequent ignition of fusion reactions in heavy-hydrogen “boost gas,” which is confined and heated within the imploding core. The flood of extra neutrons released by fusion “boosts” fission of the plutonium to a level that ensures a superabundance of radiation for imploding and igniting the secondary stage, which typically accounts for the vast bulk of the total energy released in a large nuclear explosion. Very little uncertainty attends the detonation of a proven secondary design in the presence of a minimally adequate primary yield, so under a prolonged test moratorium the major uncertainties tend to focus on the continued performance of the primary stage.

As for sustaining the ability to forecast the ultimate force of an explosion with high confidence, the strict nuclear explosive performance criteria used in the Cold War are no longer as relevant as they once were. Even for those who subscribed to the oddly detached moral psychology and doctrines of nuclear war-fighting deterrence, worry about how unexpected degradations in nuclear explosive yield (“unreliability”) might affect “kill probabilities” against hardened “counterforce” targets, such as missile silos and command centers, no longer seems to be of vital importance. Even the Pentagon’s own Defense Science Board has come round to the view that there is no need to rebuild large numbers of high yield “legacy” nuclear weapons to support a credible and effective deterrence policy.[2]

For post-Cold War deterrence, it matters little (except perhaps to the population of the targeted nation) whether or not a nuclear weapon detonates at its full thermonuclear yield (typically tens to hundreds of kilotons of TNT equivalent). Even if the secondary stage fails to ignite, the weapon would still achieve yields in the low kiloton range, sufficient to deter nuclear use by “rogue” states.

All this means that using computer simulations to guarantee that a nuclear blast will reliably yield the greatest possible explosive power is no longer vital for deterrence. Moreover, the vast improvements in guidance technology provided by global positioning system-aided systems can be expected to make up for losses in lethality against most targets from less-than-expected yields. The only exception might be when striking very deeply buried targets, which can only be destroyed by weapons with yields of several hundred kilotons or more. Yet, using such a high-fallout weapon against a non-nuclear-power would be both monstrous and implausible. And its use against a nuclear-weapon state such as Russia or China would be indistinguishable from all-out nuclear warfare, and thus confers no deterrent advantages.

Costly Campaigns

Given the reduced relevance of large-scale nuclear capabilities to countering today’s U.S. security challenges—it is inexcusable that more has not been done to reduce the scale of the size and scale of the nuclear complex. On the contrary, Bush’s budget calls for a series of so-called campaigns bent on resurrecting a modern, highly automated, and networked version of the old Cold War nuclear weapons complex. These plans would cost $1.9 billion in fiscal year 2005 and more than $12 billion over the next five years. That is nearly 50 percent more than the Energy Department will spend on the nuts-and-bolts costs of the U.S. arsenal: the warhead development, refurbishment, and stockpile maintenance tasks that have been agreed upon with the Department of Defense.

A look inside these various NNSA “campaigns” reveals an astonishing world of unaccountable spending, gross mismanagement, and technological excess by a laboratory elite that confuses its own narrow “weapons science” interests with those of the nation and its taxpayers.

Science Campaign

At the base of the stewardship pyramid one finds the so-called Science Campaign, which is really a wide array of applied science and technology projects to improve U.S. capabilities for predicting the performance of nuclear weapons. This account funds such items as improving Nevada Test Site (NTS) readiness to resume underground nuclear test explosions ($30 million in fiscal year 2005) and reviving Lawrence Livermore National Laboratory’s (LLNL) Atomic Vapor Laser Isotope Separation (AVLIS) pilot plant, which was mothballed at the close of the Cold War. The pilot plant will likely separate additional supplies of a highly prized plutonium isotope, Pu-242, which can be used to test the behavior of a bomb’s core of plutonium and high explosive at full scale without producing an explosive nuclear chain reaction. These “hydrodynamic tests,” so named because the compressed material behaves like a moving fluid, are a key tool of stockpile stewardship.

A major preoccupation of the Science Campaign in fiscal years 2005 and 2006 will be fixing the Los Alamos National Laboratory’s (LANL) Dual Axis Radiographic Hydrotest (DARHT) facility, arguably the most technically important facility in the science-based stockpile stewardship program. The facility is designed to provide rapid, high-resolution x-ray photographs simultaneously along two axes. Weapons scientists want to use this detailed three-dimensional imagery of surrogate primary-stage explosions as a benchmark to ensure the accuracy of computer simulations of the implosion process in the absence of nuclear testing.

DARHT began life in the fiscal year 1988 budget request as a modest $30 million upgrade of the existing single-axis x-ray capability to a dual axis machine. Sixteen years later, DARHT is still not finished, but now it costs $327 million, according to the Energy Department’s inspector general, who predicted in May 2003 that the facility “would not be fully operational until June 2004.”[3] As the fiscal year 2005 budget request makes clear, even this recent assessment is obsolete: deep within the document, the NNSA reported that the high-tech second axis is suffering from “high voltage breakdown” and that the “first 2-axis shot in support of stockpile assessment” would not be conducted until fiscal year 2007. That would be 19 years after the project’s inception—a project that the Energy Department claimed was absolutely critical to the success of its stockpile stewardship strategy.

Engineering Campaign

While Science Campaign projects mainly flow to the two nuclear design laboratories, the Engineering Campaign element is largely within the purview of Sandia National Laboratory, managed for the NNSA by Lockheed Martin, the nation’s largest defense contractor. Sandia is tasked with conducting non-nuclear component engineering and weapons system integration of the nuclear components developed by LANL and LLNL. The centerpiece of Sandia’s engineering campaign is the $519 million Microsystems and Engineering Sciences Application (MESA) Complex in Albuquerque.

Its purpose is to develop new microelectronic machine (“microsystem”) components to meet a postulated need for “continual advances in technologies” to improve nuclear weapon “surety” (i.e., built-in technical features that ensure against both unauthorized and accidental detonations). The complex will also endeavor to produce modern, highly integrated miniaturized replacement parts for the larger number of discrete non-nuclear components currently used in nuclear weapons, to meet the needs of the NNSA’s large “refurbishment” programs for enduring stockpile warheads.

All in all, the MESA complex will comprise approximately 391,000 square feet and house some 650 engineers working on new microcomponents for nuclear weapons. How much of this is minimally necessary in order to extend the service life of existing weapons and how much is self-serving empire building by Sandia and corporate parent Lockheed Martin is difficult for outside observers to gauge. We do know, however, that, until two years ago, this project consisted only of a proposed $51 million upgrade for retooling the existing Microelectronics Development Laboratory. One may therefore surmise that some of the recent additional work, such as the new Weapons Integration Facility with state-of-the-art “visualization” facilities for designing new weapon components, is a gold-plated pork barrel add-on that is not strictly required to sustain the nuclear weapons stockpile.[4]

Advanced Simulation and Computing Initiative (ASCI) Campaign

In the NNSA’s fiscal year 2005 request, the Bush administration reveals that it intends to spend $740 million next year on nuclear weapons simulation and computing and an astonishing $4.03 billion through fiscal year 2009—an average of $806 million per year just on nuclear weapons computing alone. Each of the nuclear weapons laboratories now has a new supercomputing center under construction or recently completed. The NNSA is nearing completion of a program to equip them with the second generation of ASCI supercomputing systems.[5]

From the inception of the ASCI program in fiscal year 1996 through fiscal year 2004, the Energy Department spent at least $4.75 billion on nuclear weapons computing. This sum does not include all the costs involved in setting up and gathering data from experiments designed to refine the physics models embedded in the various linked modules of code that attempt to simulate each stage of the implosion-explosion process. From fiscal year 2005 to fiscal year 2009, the NNSA plans to spend another $4 billion on further ASCI hardware and software development, for a total of $8.75 billion, or an average of $2.92 billion to equip each weapons laboratory with “state of the art” simulation capabilities.

In an era when “the network is the computer,” apparently no one in government (save perhaps the General Accounting Office [GAO]) thought to ask why the NNSA weapons labs could not make do with networked access to a single center for “massively parallel” computing, rather than constructing and equipping three such centers, particularly in view of the GAO’s repeated findings in the late 1990s that the Energy Department’s existing supercomputer resources were seriously underutilized: “In 1997, for example, less than 5% of the jobs run on the largest supercomputers used more than one-half of the machines’ capabilities.”[6]

Inertial Confinement Fusion and High-Yield Campaign

After ASCI, the second-largest NNSA “campaign” in Bush’s request in terms of funding ($492 million) is the Inertial Confinement Fusion and High-Yield Campaign. It is slated to account for total spending of $2.43 billion over the course of the NNSA’s projected “Five-Year National Security Plan.” Most of this planned funding is directed toward the National Ignition Facility (NIF), a massive, 196-beam laser facility under construction at LLNL. The funds will be used both to complete the facility and to achieve the technical readiness to begin NIF fusion ignition experiments. NIF is by far the largest single project in the NNSA budget and, quite possibly, the most expensive experimental facility ever built.

According to the NNSA, when laser system installation is finally completed in September 2008, construction of NIF will have “officially” cost $3.5 billion. The Natural Resources Defense Council (NRDC) estimates that the total actual cost of the project through fiscal year 2008, including ignition target research and development (R&D), production, and diagnostics, will actually be much higher, on the order of $5.2 billion, and further hundreds of millions if not billions of dollars will be required to reach the first “demonstration” of the facility’s namesake mission, fusion ignition, now variously postponed until 2010-2014.

Recall that a driving rationale for the NIF was the ostensibly “critical” need to have an “ignition facility” capable of “propagating fusion burn and modest energy gain” in place by last year in order to help “validate” the three-dimensional computer codes under development in the ASCI program. In initially pointing to the 2003 date, the Energy Department had said that by last year “most of the weapons in the stockpile will be in transition from their designed field life to beyond field life design” and “the number of designers with test experience will be reduced by about 50 percent.”

Readiness Campaign

The NNSA says its Readiness Campaign is intended to “revitalize the nuclear weapons manufacturing infrastructure” by improving both its “responsiveness” and its “technology base.” Claiming marching orders from the Bush administration’s December 2001 Nuclear Posture Review, the NNSA claims that “a truly responsive infrastructure is the cornerstone of the new nuclear defense triad” outlined in this document. “To be considered a credible deterrent, this infrastructure must include a manufacturing capability with state-of-the-art equipment combined with cutting-edge applications of technology, and an ability to quickly provide modified or enhanced capabilities and products to meet emerging threats.”[7]

In fiscal years 2005-2009, the readiness program consists of five subprograms with a projected price tag of $1.65 billion. These include thermonuclear component (“secondary”) manufacturing, high-explosives production and weapons assembly/disassembly, electronic and mechanical components, new computerized production technologies, and readiness to produce tritium for stockpile weapons. Some of this work is necessary for maintaining a nuclear deterrent stockpile, but much of it is not.

Apart from its evident self-serving qualities, there are some logical flaws to the NNSA’s new deterrent construct. To be credible, nuclear weapons need not be produced with “state-of-the-art equipment” or “cutting edge technology.” Indeed, Bush professes to have invaded Iraq to forestall development of what clearly would have been a crudely produced nuclear explosive device, the threat of which he nonetheless found to be credible. Further, one is hard pressed to see how, absent far-reaching changes in the global security environment, the existence of modern production infrastructure per se, rather than actual weapons and forces, would be considered a deterrent to armed attack upon the United States or its allies and friends. If this were true, the administration should have no objection to eliminating the entire U.S. nuclear arsenal and relying on its fearsome industrial capability to reconstitute the arsenal to discourage cheating on a global nuclear disarmament regime.

The NNSA seems to be positing a novel extension or reinvention of the concept of deterrence, one that is more accurately described as “dissuading” or “discouraging” potential rivals for global preeminence from even seeking to acquire nuclear-weapon capabilities commensurate with those of the United States. This approach has nothing to do with classic deterrence of nuclear attack through an assured survivable capability for nuclear retaliation, nor even with “extended” deterrence of conventional conflict through calibrated, “not incredible” threats to use nuclear weapons first. On the contrary, it is a self-serving argument to justify continued work for defense bureaucracies and their contractors.

One particularly wasteful aspect of the Energy Department’s Readiness Campaign has been the expenditure of at least $2.6 billion since 1996 in maintaining and attempting to restore U.S. tritium recycling, production, and extraction capabilities.[8] The Energy Department has spent this huge sum even though U.S. tritium requirements have been declining steadily with continuing reductions in the requirement for “active” stockpile weapons and no fresh tritium was required or produced to support the stockpile throughout this period. The recent spending has been premised on restoring a capability to support the Pentagon’s declared START II stockpile tritium requirement (i.e., 4,900 “active” nuclear weapons with filled tritium reservoirs plus a “five-year reserve” to “surge” deployment of additional weapons or replenish these reservoirs in the event production were disrupted for a prolonged period). The NRDC estimates that this artificially inflated requirement could be met from existing tritium supplies until 2007, at which time the reserve would dip below its “required” five-year support level. So, the lack of fresh tritium supplies would not begin to affect the ability to deploy actual weapons until 2012.

Suffice to say, after the force reduction envisioned in the 2002 Strategic Offensive Reductions Treaty (SORT), this tritium requirement scenario is now outdated, but it has yet to be formally replaced with another, more realistic one. Supporting the reduced but still large nuclear force levels called for under SORT, for example, 1,700 “operationally deployed” strategic weapons and 300 operationally “deployable” nonstrategic weapons with a five-year tritium reserve, would not “require” resumption of tritium production to prevent a decline in the reserve until around 2012; actual weapons would not need fresh tritium until 2017. Clearly, the Bush administration and the Congress have a lot of flexibility and additional time to determine both “required” operational nuclear force levels and when or, indeed, whether to resume tritium production.

They should use this time to reconsider the administration’s tritium plans. They could begin with the need for a five-year tritium reserve. The existing requirement makes no economic sense for a costly decaying asset such as tritium. Moreover, the president has the inherent authority under the Atomic Energy Act to direct production of tritium in any one of 100 civilians reactors in the (unlikely) event of a genuine national security “emergency.” So, shifting to a shorter two-year reserve makes more sense and would further extend until 2015 the point at which new tritium production might be needed. Additionally, if the United States and Russia were to make further cuts, say, to a level of 1,000 deployed nuclear warheads each, such a U.S. force could be maintained with a two-year reserve until around 2022 without producing additional tritium.

Additionally, the NNSA’s plans to satisfy future tritium requirements, largely inherited from the previous administration, are fraught with inefficiency, bad management, and a continuing failure to consolidate tritium R&D operations. The NRDC estimates that the tritium capabilities spending outlined in the administration’s budget projections for fiscal years 2005-2009 totals at least $1.2 billion.

Rather than consolidating or eliminating sites where tritium R&D activities are conducted, in which each site requires their own (decaying) minimum inventories of tritium and carries high fixed-overhead costs for security and environment, safety, and health, the Energy Department has continued under President Bill Clinton as well as President George W. Bush to sustain tritium operations at both LANL and LLNL as well as at SRS. Indeed, the Bush administration’s particular contribution is to reinvigorate the tritium R&D facility (B331) within LLNL’s “Superblock Complex.” Successive administrations have allowed the Energy Department to maintain tritium, plutonium, and radiographic hydrotest facilities in triplicate at LLNL, LANL, and NTS, as though the nuclear arms race with the former Soviet Union had never ended.

Pit Manufacturing and Certification Campaign

This weapons complex “campaign,” ongoing since fiscal year 1993, has the immediate goal of “restoring [at LANL] some limited capacity to manufacture pits of all types” that was lost in 1989 when the main Cold War pit manufacturing plant, located at Rocky Flats northwest of Denver, imploded in a multibillion dollar debacle of contamination, criminality, and managerial incompetence.[9] The longer-range NNSA objective is the design and construction of a $2-4 billion Modern Pit Facility (MPF) to support long-term pit manufacturing beginning late in the next decade. This comes even as LANL is well along in a $2.3 billion, decade-long modernization of its pit fabrication and plutonium chemistry complex, which is scheduled to reach a capacity for producing 20 pits a year by 2007. (Another article in this issue looks in detail at the debate surrounding the need for the Modern Pit Facility. See page 10).

Quite apart from the MPF, however, the administration’s five-year national security plan calls for spending $1.3 billion on pit manufacturing and certification from fiscal year 2005 to fiscal year 2009, on top of the $1.2 billion already expended during fiscal years 1998-2004. It is something of a mystery how so little capability could have resulted from the huge amount already expended.

In the next budget year, the NNSA intends to manufacture “at least six certifiable W88 pits” to augment the six being produced this year, with the goal of having these certified for use by 2007 as replacement pits for the 450-kiloton thermonuclear warheads for the Trident II submarine-launched missile. Manufacture of these six “certifiable” pits, not including the cost of the plutonium itself, will cost $132 million in the next fiscal year, or $22 million per pit, which amounts, at current prices, to roughly 480 times the value of the pit’s weight in gold.[10]

By comparison, the Manhattan Project produced the first significant quantities of separated plutonium in human history and manufactured it into pits within three years. Now, we are supposed to believe that a half century of experience later, and tens of thousands of pits later, that LANL legitimately requires 11 years and more than $2.5 billion to confidently manufacture and certify one “war reserve pit” for the nuclear weapons stockpile?

Recommendations

Given the many problems with the NNSA’s approach to the stockpile stewardship program, Congress needs to take action. It should defer consideration of any new facility or weapons refurbishment request until the administration has submitted and Congress has examined, debated, and approved the essential contours of a plan to reduce the nuclear weapons stockpile to levels that make sense for the post-Cold War world. Congress needs to cease ducking the nuclear weapons issue and have the debate, examining what can and should be changed. There are potentially billions of dollars in annual savings to be had from shifting to a far more compact, less alert, and less deployed nuclear force, supported by a smaller and more sustainable nuclear complex. This money could be used to reduce the federal budget deficit or improve U.S. national security by substantially increasing the $500 million annually NNSA spends on nonproliferation initiatives in Russia and other foreign countries.

Once Congress has approved a revised nuclear stockpile plan, it must determine how to rationalize and consolidate the NNSA complex to eliminate Cold War redundancies, reduce its environmental footprint, and reduce the security and safeguards overhead burden, which is growing rapidly. A key element will be to streamline and simplify the current, needlessly complex “virtual testing” paradigm for stockpile stewardship. After a decade, the current program has not built the kind of disciplined and conservative protocols that would permit indefinite, confident retention of a nuclear weapons stockpile at minimum cost.

The NNSA and its laboratories have mystified the basic tasks of sensible stockpile stewardship in ways that are beyond the bounds of fiscal sanity and reason. Far from boosting confidence, the current program is actually structured to function as a constant wellspring of uncertainty, thereby fostering continuing costly “requirements” and “milestones” for ever more elaborate experimental and computational facilities. Of course, the need to resolve the accumulated conflicts and inconsistencies in the data generated by numerous scaled experiments and weapons computations could just as easily lay the groundwork for what will appear to be an earnest case for a “limited” return to nuclear explosive testing to “resolve the uncertainties” and “calibrate the codes.”

Indeed, the notion at the heart of the current bloated stewardship effort—that the United States must continually expand its nuclear weapons knowledge and capability while threatening or attacking others whom we suspect of doing the same thing —is an Orwellian mismatch with US nonproliferation objectives. The Administration’s errant initiatives to increase readiness to resume nuclear test explosions, build a large Modern Pit Facility, and modernize thousands of Cold War legacy weapons—including conversion of high-yield strategic bombs to “robust” nuclear earth penetrators—all share the distinction of being both wasteful and politically destabilizing. Congress should decline funding for these efforts while taking meaningful steps to restructure and reduce the nuclear weapons complex.

NOTES

1. See Christopher E. Paine, “A Case Against Virtual Nuclear Testing,” Scientific American 281, no. 3 (September 1999): 64; Christopher E. Paine and Matthew G. McKinzie, “Does the U.S. Science-Based Stockpile Stewardship Program Pose a Proliferation Threat?” Science and Global Security, vol. 7 (1998): 151-193; Matthew G. McKinzie, Thomas B. Cochran, and Christopher E. Paine, “Explosive Alliances: Nuclear Weapons Simulation Research at American Universities,” Natural Resource Defense Council (NRDC) Nuclear Program, Washington, DC, January 1998.

2. “The nuclear weapons program as currently conceived—a program focused primarily on refurbishing the legacy stockpile—will not meet the country’s future needs.” Defense Science Board, “Report of the Defense Science Board Task Force on Future Strategic Strike Forces,” Washington, DC, February 2004, pp. 1-10.

3. Office of Audit Services, Office of Inspector General, U.S. Department of Energy, “Audit Report: Dual Axis Radiographic Hydrodynamic Test Facility,” DOE/IG-0599, May 2003, p. 4. The report cites an official “total project cost” of $270 million, plus $57.5 million in additional costs “associated with work elements transferred outside of the project.” But NRDC estimates that the final cost of the fully commissioned DARHT facility will be even larger, probably on the order of $500 million.

4. The funding history of the MESA project supports this view. The fiscal year 2003 omnibus appropriations bill provided $113 million, an increase of $38 million above the request. The fiscal year 2004 Energy and Water Development Appropriations Act, P.L. 108-137, enacted December 1, 2003, provided $87 million for MESA, an increase of $25.2 million above the request. These increases were almost certainly done at the behest of the senior senator from New Mexico, Pete Domenici (R), chairman of the appropriations energy and water subcommittee.

5. The first generation of ASCI machines deployed from 1996-2000 had processing speeds ranging from roughly 1-10 teraOPS (trillions of floating point operations per second); the current second generation has speeds in the 10-100 teraOPS range; and the third generation, starting in 2005, will have speeds of hundreds of teraOPS to “petaOPS.”

6. U.S. General Accounting Office, “Problems in the Management and Use of Supercomputers,” GAO/T-RCED-99-257, July 14, 1999.

7. NNSA 2005 CBR, p. 158 (Weapons Activities/Readiness Campaign). See ibid., app. 20 (details of the Energy Department’s efforts in “virtual prototyping” and increased automation of the nuclear weapons “enterprise”).

8. Tritium, a radioactive isotope of hydrogen, is used to boost the fission reaction in the primary stage of nuclear weapons, enabling the use of lesser amounts of plutonium and high explosive and making the weapons more resistant to accidental nuclear detonation. Tritium decays at the rate of 5.47 percent per year, so the stockpile must ultimately be replenished when stockpile reductions can no longer keep pace with the decline in the tritium inventory.

9. In 1989, nuclear weapons production at Rocky Flats was abruptly halted because of serious environmental, safety, and health problems at the plant. Operations at the plant and the site contractor, Rockwell International, were the subject of an intensive two-year federal grand jury investigation that began in 1989 after FBI agents raided the plant. Rockwell, which had operated Rocky Flats for more than a decade ending in 1989, later pled guilty to 10 environmental crimes and paid an $18.5 million fine.

10. With gold priced at $405 an ounce.

Expensive Life Support for the U.S. Nuclear Arsenal

Extending the life of the current massive U.S. arsenal does not come cheap. If the NNSA sticks to its present “life extension” plans, over a seven-year period that will end in 2009, its direct costs for maintaining and updating nuclear warheads will amount to $11 billion, including $3.2 billion for 2,500-3,500 upgraded submarine-launched ballistic missile warheads; almost $2 billion for 1,700-2,500 air-delivered gravity bombs; $1.6 billion for 1,500-2,500 cruise missile warheads; and $1.3 billion for up to 1,500 ICBM warheads. On the other hand, it will spend only $235 million retiring and dismantling weapons.

Yet, that only scratches the surface. Looking at the next five years, only a little more than one-quarter ($8.2 billion) of the $36 billion that the Bush administration has proposed spending on the nuclear weapons complex will go toward the actual work that is needed for military commanders to be sure that their weapons are in working order. These tasks, “directed stockpile work,” are spelled out in the president’s annual Nuclear Weapons Stockpile Memorandum.

The rest of the five-year nuclear weapons budget, amounting to some $27.5 billion, represents a 330 percent “overhead charge” on top of the direct costs for the NNSA’s “deliverable“ product: stockpiled nuclear warheads.

Tellingly, the largest single category—almost $8.6 billion—in the Bush administration’s five-year, $36 billion spending plan for nuclear weapons is not for actual programmatic work but for an amorphous category called “Readiness in Technology Base and Facilities.” According to the budget request, this spending is used to operate and maintain a wide range of “program infrastructure and facilities…in a state of readiness, ensuring each capability (workforce and facility) is operationally ready to execute programmatic tasks identified in ‘Campaigns’ and ‘Directed Stockpile Work.’”

Christopher Paine is a Senior Nuclear Program Analyst at the Natural Resources Defense Council.