4
Clandestine Stocks and Production of Nuclear Weapons and Nuclear-Explosive Materials

We concluded in Chapters 2 and 3 that procedures and technology are available to verify with high confidence declarations of stockpiles of nuclear weapons and nuclear-explosive materials (NEM) at declared sites. But undeclared nuclear weapons and NEM could exist as a consequence of retention of undeclared existing nuclear weapons and NEM or could come into existence by the clandestine production of nuclear weapons from existing NEM. In addition, undeclared NEM for weapons might be produced clandestinely or diverted covertly from peaceful nuclear power programs. Current non-nuclear weapon states and possibly terrorist groups might also acquire nuclear weapons or NEM. The potential for clandestine activities in these categories poses the largest challenges to efforts to strengthen transparency and monitoring for nuclear weapons, components, and materials on a comprehensive basis.

This chapter addresses those challenges, asking how and to what degree other states could gain confidence that undeclared nuclear weapons and stocks of NEM do not exist and are not being produced at clandestine facilities. Accordingly, we describe tools and techniques that could be used for two interrelated tasks: (1) to detect undeclared stocks and production activities that might exist and (2) to narrow the uncertainties in declarations that could mask such clandestine activities.

Table 4-1 highlights the key routes by which a state might retain or acquire undeclared nuclear weapons. The most straightforward of these is the retention of existing nuclear weapons at clandestine sites. A state wishing to cheat would simply move some number of weapons to a secret facility and provide a false declaration indicating that these weapons had never been produced or that they had long ago been dismantled. Alternatively, new nuclear weapons could be assembled at a clandestine facility. In this case,



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials 4 Clandestine Stocks and Production of Nuclear Weapons and Nuclear-Explosive Materials We concluded in Chapters 2 and 3 that procedures and technology are available to verify with high confidence declarations of stockpiles of nuclear weapons and nuclear-explosive materials (NEM) at declared sites. But undeclared nuclear weapons and NEM could exist as a consequence of retention of undeclared existing nuclear weapons and NEM or could come into existence by the clandestine production of nuclear weapons from existing NEM. In addition, undeclared NEM for weapons might be produced clandestinely or diverted covertly from peaceful nuclear power programs. Current non-nuclear weapon states and possibly terrorist groups might also acquire nuclear weapons or NEM. The potential for clandestine activities in these categories poses the largest challenges to efforts to strengthen transparency and monitoring for nuclear weapons, components, and materials on a comprehensive basis. This chapter addresses those challenges, asking how and to what degree other states could gain confidence that undeclared nuclear weapons and stocks of NEM do not exist and are not being produced at clandestine facilities. Accordingly, we describe tools and techniques that could be used for two interrelated tasks: (1) to detect undeclared stocks and production activities that might exist and (2) to narrow the uncertainties in declarations that could mask such clandestine activities. Table 4-1 highlights the key routes by which a state might retain or acquire undeclared nuclear weapons. The most straightforward of these is the retention of existing nuclear weapons at clandestine sites. A state wishing to cheat would simply move some number of weapons to a secret facility and provide a false declaration indicating that these weapons had never been produced or that they had long ago been dismantled. Alternatively, new nuclear weapons could be assembled at a clandestine facility. In this case,

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials the required NEM could be supplied from existing undeclared stocks or could be newly produced at undeclared production facilities. We first discuss tools and techniques for detecting clandestine stocks of nuclear weapons or NEM, if they exist, and then turn to the problem of detecting clandestine production of weapons and NEM. In both cases, we attempt to estimate very roughly the maximum size of an undeclared stockpile or production activity that might go undetected. TABLE 4-1 Routes to Undeclared Nuclear Weapons Route to undeclared nuclear weapons Source of NEM   Move existing weapons to a clandestine storage or deployment facility None Required   Transfer of weapons from another state     Assemble new weapons at clandestine facility   Existing, undeclared stocks at clandestine storage facility New, undeclared production at clandestine facility Covert diversion from declared stocks or production facility Transfer from another state These highlighted routes to undeclared weapons are the principal focus of this chapter. Covert diversion of NEM from declared stocks or production facilities was considered in Chapter 3, where we concluded that the application of safeguards like those used by the International Atomic Energy Agency (IAEA) should be able to detect any significant diversion. We also do not consider here the problem of overt breakout—that is, the open diversion of NEM from declared stocks or production facilities—because this would provide timely warning that the state may be producing undeclared nuclear weapons, which is the purpose of the monitoring system. Nor do we discuss the possibility of theft or transfer of nuclear weapons or NEM from other states. We assume that any significant theft would be detected, and the transfer of weapons or NEM would only shift the problem of detecting a false declaration from one state to another (provided that all such states were subject to equivalent monitoring and safeguards arrangements).

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials DETECTING UNDECLARED STOCKS OF WEAPONS AND NEM In order to gain confidence that declarations are complete and that no undeclared stocks of nuclear weapons or NEM exist, states or an inspection agency must have an ability to collect direct or circumstantial evidence of undeclared stocks if they do exist. Direct evidence would be the actual discovery of undeclared weapons or NEM. Circumstantial evidence would include audits of records and physical evidence from which one could infer that it is likely that undeclared nuclear weapons or stocks of NEM exist at some unidentified location. If a state agreed to declare all of its nuclear weapons and all stocks of NEM, then the discovery of a single weapon or container of NEM not listed in the declaration would ipso facto be a violation of the agreement, unless promptly and satisfactorily explained. As noted in previous chapters, an agreement to tag all declared nuclear weapons and containers of NEM would greatly facilitate monitoring, because the discovery of a single weapon or container without a valid tag would be direct evidence of a violation. If tags are not used, the state might claim that the discovered weapons or containers of NEM were not properly listed in the declaration because they had been moved recently and the declaration had not been updated accordingly, or some other excuse that would be difficult to refute incontrovertibly. National Technical Means The main problem with discovering undeclared stocks is knowing where to look. Existing storage facilities for nuclear weapons and NEM, such as the bunkers shown in Figure 2-3, are distinctive and easily detected and identified using National Technical Means (NTM), in particular high-resolution satellite photography.1 Any such facilities that were omitted from the declaration would be high priorities for challenge on-site inspections. But a state wishing 1   NTM also include other satellite-borne sensors, as well as ground- and sea-based receivers that collect a broad range of signals intelligence. Such information collection is acceptable under international law by sensors located outside the borders or above the sensible atmosphere of the state being inspected. States may also enter into agreements that allow ground- or air-based sensors to be located on or flown above the territory of a state subject to inspection. This is a feature of the Open Skies Treaty and the Comprehensive Nuclear Test Ban Treaty.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials to cheat could be expected to select a building that would give few indications of its contents, such as a warehouse in an industrial complex, with few outward signs of high security or other activities usually associated with the storage of nuclear weapons or NEM. Nuclear weapons and containers of NEM are small—even a house-size building could hold hundreds of weapons (or enough NEM for hundreds of weapons). The number of personnel and traffic to the clandestine facility could be kept to a minimum and given a plausible cover story. NEM components or containers of bulk material need little or no maintenance. Nuclear weapons require periodic inspections and the replacement of limited-life components, such as batteries or tritium, but this could be accomplished without attracting attention. If necessary, weapons and materials could be moved between clandestine facilities using commercial vans or trucks, with no security escort. Nuclear weapons and NEM give few clues about their existence within a facility. NEM emits neutrons and gamma rays, but these radiations are easily shielded and are detectable only at relatively short range (on the order of 100 meters or less) even without shielding. Barring a severe accident, such as the detonation of the high explosive in a weapon, the likelihood that a clandestine storage facility would be detected using NTM probably would be low. One might intercept communications indicating suspicious activities that may be related to nuclear weapons or NEM, but such opportunities would be limited given the relatively low level of activity and small numbers of people required to maintain clandestine stocks. Human Sources The surest way to locate a clandestine storage facility for undeclared nuclear weapons or NEM is through human intelligence—for example, the leak of information by someone who has knowledge of the hidden stocks. Governments have varied in their ability to maintain secrecy. The case of Mordechai Vanunu, a nuclear technician who claimed to have personal knowledge of Israel’s undeclared nuclear weapon activities, was widely publicized. Even oppressive and secretive governments such as Iraq and North Korea have experienced high-level defectors. While in the United States the secrecy of some major programs was apparently well kept, the general outlines of the Manhattan Project, as well as many sensitive technical details, were known to the Soviet Union through individuals associated with the U.S. nuclear weapons pro-

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials gram. Over the years, there have been many serious leaks from highly sensitive U.S. and Soviet programs, motivated by ideology or greed. The extent to which a covert program may have been compromised by individuals within the program or by knowledgeable outsiders can never be known with certainty by a state undertaking the secret program. One may be able to increase the probability that individuals would report activities that contravene international agreements to international authorities. If the individuals most likely to have knowledge of undeclared activities can be identified, inspectors can request confidential interviews with these people. Such interviews have been used by the IAEA to resolve uncertainties and investigate possible violations of safeguards agreements. Individuals also could be encouraged to come forward on their own, by requiring that states pass domestic laws making it illegal for individuals to participate in activities that contravene an international agreement, requiring that individuals report violations to a designated international commission, and guaranteeing immunity for the “whistle blowers.” This concept, which is sometimes called “societal verification,”2 received a favorable review by former Under Secretary General for Disarmament Affairs of the United Nations, Jayantha Dhanapala.3 Societal verification might serve as an additional deterrent to governments contemplating whether to establish a clandestine stockpile of weapons or NEM. A government wishing to cheat would attempt to screen participants to eliminate potential whistle-blowers, but it could never be absolutely and forever certain of their loyalty to the regime. 2   Joseph Rotblat, “Societal Verification,” in Joseph Rotblat, Jack Steinberger, and Bhalchandra Udgaonkar, eds., A Nuclear-Weapon-Free World: Desirable? Feasible? (Boulder, CO: Westview Press, 1993), pp. 103-118. Joseph Rotblat, “Citizen Verification,” Bulletin of the Atomic Scientists 48 (May 1992), pp. 18-20 and Richard L. Garwin, Appendix H “Theater Missile Defense, National ABM Systems, and the Future of Deterrence,” in Naval Studies Board, National Research Council Post-Cold War Conflict Deterrence (Washington, DC: National Academies Press, 1997), pp. 182-200. Available as of January 2005, at: http://bob.nap.edu/html/pcw/Dt-h.htm. 3   Jayantha Dhanapala, “Civil Society and the Verification of Disarmament,” workshop on “Societal Verification” at the Hague Appeal for Peace conference, May 13, 1999. Available as of January 2005, at: http://www.un.org/Depts/dda/speach/13May1999.htm.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials Audits of Records The probability that clandestine stocks would be discovered and directly observed cannot be assumed to be high. Fortunately, there are ways that an inspection agency or other states might gather circumstantial evidence of the existence of undeclared nuclear weapons and stocks of NEM. One can examine the original records for the assembly and disassembly of nuclear weapons and the production and use of NEM, for example, to check their authenticity, internal consistency, and consistency with the declaration and with intelligence information. An examination of operating records is similar to a financial audit, in which a company’s records are examined to verify the legitimacy and accuracy of transactions, and that balances accurately reflect receipts and disbursements. It is reasonable to assume that the assembly and disassembly of every nuclear weapon was documented in the operating records of each assembly facility. The declared inventory of each weapon type should equal the total recorded number assembled minus the number disassembled. In order to hide existing nuclear weapons, the records would have to be falsified to indicate either that the hidden weapons had never been assembled, or that they had been disassembled or explosively tested at an earlier date. The authenticity of original paper records, if they still exist, could be verified using standard forensic techniques, which should be able to detect any recent alterations. The records also could be examined for any unusual patterns—for example, periods of low assembly or high disassembly rates or of highly uniform assembly or disassembly rates—that might indicate attempts to falsify the records to hide the existence of undeclared weapons. Finally, archived national intelligence information could be checked for consistency with the operating records provided. Imagery taken by photoreconnaissance satellites might show activity at variance with the records—for example, movements of large numbers of weapons out of or into the plant during periods when the operating records showed little or no activity. Operating records for NEM production facilities also could provide evidence about the existence of undeclared weapons, or undeclared stocks of NEM that could be used to assemble them. The declared inventories of highly enriched uranium (HEU) and plutonium should be consistent with original records for the production and use of these materials, and the records should be internally consistent. In the case of a uranium enrichment plant, for ex-

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials ample, the recorded receipts of uranium entering the plant as feed should be consistent with recorded shipments of HEU and discharges of depleted uranium. Over a given period, the mass of uranium and U-235 in the feed should equal the masses of uranium and U-235 in the HEU and the depleted uranium combined, with some margin of error due to measurement uncertainties. In addition, records of the separative work performed could be compared with the capacity of the plant, which can be verified through design documents and on-site inspections. The material balance will depend on the design and operation of the plant, but for illustration, consider a plant that produces HEU and depleted uranium with U-235 concentrations of 90 percent and 0.35 percent, respectively. For every 1,000 kilograms of natural uranium entering the plant, 4 kilograms of HEU and 996 kilograms of depleted uranium are produced. The feed contains 7.1 kilograms of U-235, of which 3.6 kilograms goes into the HEU and 3.5 kilograms goes into the depleted uranium. Similar methods could be applied to verifying declarations of plutonium production. This would involve examining records of the fabrication of uranium fuel and target rods for production reactors; the design of the fuel and the reactors, typical fuel loadings in the core, and dates of fuel loading and discharge; monthly production of thermal energy; shipments of spent fuel; the design of the reprocessing plants; monthly production of plutonium product; and the volume, isotopic composition, and disposition of the various waste streams. As with weapons, archived intelligence information might be checked for consistency with the records; for example, imagery of a plutonium production reactor or a gaseous diffusion enrichment plant could indicate whether the plant is operating and, at least roughly, the level of production. Production records for NEM could also be checked for consistency with records for weapon assembly. In the case of gaseous diffusion plants, records of enrichment work performed might also be checked against records of electrical consumption. This method of verifying production declarations would depend on the accuracy, completeness, and authenticity of the records that were provided. As noted above, the authenticity of paper records can be verified, but the original records might have been lost or destroyed, and electronic records may be impossible to authenticate. Moreover, even authentic records may be unintentionally inaccurate or incomplete. Record keeping was not exemplary

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials in the early days of most nuclear weapon programs, when the emphasis was on producing material and weapons as quickly as possible. Although record keeping presumably was good for weapons, plutonium, and HEU, it is less likely that accurate records were kept for less valuable materials, such as natural or depleted uranium or reprocessing wastes, which are important for accounting purposes. Physical Evidence A variety of methods exist for gathering physical evidence that could be used to confirm NEM production records and resolve uncertainties or apparent discrepancies in them. These methods are sometimes referred to as “nuclear archaeology.” In the case of graphite-moderated plutonium production reactors, for example, isotope ratios of impurities in the graphite can provide an accurate estimate of the total integrated neutron flux that was available to produce plutonium during the life of a reactor.4 Isotopes of various common impurities capture neutrons and produce heavier isotopes of the same element. The ratios of the resulting stable isotopes can be accurately measured by mass spectrometry. By comparing these isotopic ratios with those occurring naturally, the integrated neutron flux at the point of capture can be determined. This method is particularly attractive in that it does not depend on the original concentration of the impurity in the tested items, but only on the ratios of remaining isotopes. A model of the neutron flux with the reactor, however, is essential to relate such measurements accurately to the plutonium produced during the life of the reactor. Titanium has been identified as a particularly attractive impurity for this purpose. A number of other elements that exist as impurities in graphite would allow independent checks. Measurements of titanium isotope ratios in graphite core samples reportedly can give error margins as low as 2 percent on total plutonium production, assuming that the relevant neutron capture cross sections will be measured more accurately.5 For a relatively small cost, the plutonium production from all of the 13 former graphite-moderated Soviet production reactors, which produced essentially all the Russian Federation’s plutonium, could be esti- 4   Thomas W. Wood, Bruce D. Reid, John L. Smoot, and James L. Fuller, “Establishing Confident Accounting for Russian Weapons Plutonium,” Nonproliferation Review 9 (Summer 2002). 5   Talbert, R. J., et al. “Accuracy of Plutonium Production Estimates from Isotope Ratios in Graphite Reactors,” PNL-TEC 0693 (Richland, WA: Pacific Northwest Laboratory, February 1995).

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials mated in this way. The Russian Federation has made plans to entomb the closed-down production reactors under radiation barriers, after which it will be much more difficult and expensive to carry out nuclear archaeology. As of early 2005, however, these plans had not moved forward, so the opportunity remains. Although there have been discussions of a joint U.S.-Russian experiment on one of the reactors and some proposals have been developed, it had not been funded as of early 2005. The United States has used both graphite- and heavy-water-moderated reactors to produce plutonium. The nine graphite-moderated reactors at Hanford reportedly produced 76.4 tons of plutonium; with an accuracy of 2 percent this implies an uncertainty of about 1.5 tons. The five heavy-water-moderated reactors at Savannah River produced 36.1 tons of plutonium. A comparable technique has not yet been developed for heavy-water reactors, and the fractional uncertainty would most likely be larger. In principle, the cumulative production of HEU could be calculated by determining the total amount of U-235 stripped from the stocks of depleted uranium produced by the enrichment facility. In practice, however, a comprehensive inventory of depleted uranium stocks of U.S. and Russian would be very difficult and costly, and the resulting estimate of HEU production probably would not be sufficiently accurate to be useful. As an indication of the magnitude of the problem, the U.S. Department of Energy stores about 500,000 tons of depleted uranium, mostly in the form of uranium hexafluoride in some 60,000 steel cylinders at three U.S. enrichment plants. This depleted uranium resulted from the production of HEU at various enrichment levels for nuclear weapons and fuel for naval, research, and test reactors, as well as LEU at various enrichment levels for civilian reactor fuel. Moreover, the isotopic composition of the feed varied significantly, including partially depleted and slightly enriched uranium recovered from irradiated reactor fuels in addition to natural uranium. Finally, a small fraction of the depleted uranium discharged from U.S. enrichment plants has been used for various defense and commercial purposes, such as tank armor, armor-piecing munitions, ballast, and radiation shielding. Russian stockpiles of depleted uranium are likely comparable in size and in the complexity of their production history. Information from depleted uranium might be useful in cross-checking production records, however. It would be possible, for example, to date accurately the contents of the individual cylinders of uranium hexafluoride using the decay products of the uranium

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials isotopes, and to determine by the ratios of U-234 to U-235 whether these particular tails resulted from the production of HEU or low enriched uranium (LEU).6 If the depleted uranium cylinders are numbered and if these numbers are documented in the operating records, measurements on a random sample of cylinders could be used to confirm the accuracy of the records. South Africa: Verifying the Completeness of Declarations South Africa presents an important case study in verifying the completeness of declarations. Having built six gun-type nuclear weapons during the 1980s, South Africa decided to dismantle its weapons and join the Nuclear Non-Proliferation Treaty (NPT) as a non-nuclear weapon state. Beginning in July 1990, South Africa dismantled all its weapons, decommissioned its production and assembly facilities, and recast the HEU weapon components into standard shapes for storage and international inspection. In March 1993 President de Klerk announced that South Africa had built and dismantled six nuclear weapons. The IAEA was given a full history of the nuclear weapon program, along with a list of the people involved in it. The agency was granted permission to conduct inspections at any relevant location and to interview former managers and workers about the program. A special team of inspectors easily confirmed that the declared amount of HEU had been placed in storage and that weapon-related activities had ceased at various declared facilities. Verifying that the declaration was complete—that South Africa had dismantled all of its weapons and placed all of its HEU under safeguards—was considerably more difficult, however. The IAEA requested the historical production records from South Africa’s uranium enrichment plant. Inspectors used these records, along with other technical and design documents, to recalculate the plant’s daily production over the entire period and to produce an overall material and isotopic balance. On the basis of these studies, together with examination of the facilities and interviews of facility personnel, the IAEA concluded that the amount of HEU that could have been produced was consistent with the amount in South Africa’s initial declaration within an acceptable margin of uncertainty—less than 25 kilograms, or roughly 5 percent of the declared inventory of HEU. The IAEA therefore con- 6   Steve Fetter, “Nuclear Archaeology: Verifying Declarations of Fissile Material Production,” Science and Global Security 3 (1993), pp. 237-259.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials cluded that there was no evidence that South Africa’s declaration was incomplete; that is, there was no evidence of the existence of undeclared weapons or significant amounts of undeclared HEU.7 A comprehensive assay of the 370 tons of depleted uranium tails, which are stored on-site in some 600 cylinders, would have reduced somewhat the uncertainty in the material balance, but the IAEA decided that the increased confidence provided by such measurements would not justify the considerable expense. Measurements of certain samples were performed, however, which confirmed IAEA calculations (based on the original production records) indicating that substantially more U-235 was contained in the depleted uranium than had been recorded in the official accounting records. These measurements were an important piece of evidence that allowed the IAEA to conclude that there was no evidence of any missing HEU. The IAEA experience in South Africa illustrates that high confidence in the completeness of declarations can be achieved with a high level of cooperation and transparency. But it also demonstrates that uncertainties will exist and may be difficult to resolve. Verifying South Africa’s HEU declaration was complicated by the fact that uranium had also been enriched for nonweapons purposes, and poor records were kept of wastes and materials not valuable for the weapons program. According to then-Director General Hans Blix, “There is inherent difficulty in verifying the completeness of an original inventory in a country in which a substantial nuclear program has been going on for a long time.”8 The irony is that South Africa’s nuclear program, which produced a total of six weapons, may prove to be the smallest and shortest lived of all programs that produced a nuclear weapon. We believe that the difficulties experienced by the IAEA in verifying South Africa’s declaration are likely to be much more difficult for the de jure nuclear weapon states, making it much more difficult to conclude that their declarations are complete. 7   Adolf von Baeckmann, Gary Dillon, and Demetrius Perricos, “Nuclear Verification in South Africa,” IAEA Bulletin, 37 (March 1995). Available as of January 2005, at: http://f40.iaea.org/worldatom/Periodicals/Bulletin/Bull371/baeckmann.html. 8   Hans Blix, Statement to the 36th Session of the General Conference of the IAEA, Sept. 21, 1992.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials The IAEA study estimated releases of a mere 0.01 to 1 gram per year of HEU for a centrifuge enrichment facility producing 25 kilograms of HEU per year, based on data from existing commercial facilities. Even at the upper limit of 1 gram per year, the study concluded that a network of 26 stations could detect with high probability releases from a clandestine centrifuge enrichment facility over only a tiny fraction of a million square kilometer area of the Middle East (even after excluding the areas considered too far from roads, electricity, and people to support such a facility).18 Releases from gaseous diffusion and electromagnetic isotope enrichment plants would be much larger and more detectable. Unless substantial progress is made in reducing the cost or increasing the sensitivity of analyzing uranium particulate samples, wide-area environmental sampling would be a practical method of detecting undeclared uranium enrichment activities only over small areas, such as the Korean peninsula. In summary, we judge that environmental samples taken during on-site inspections can be highly effective in detecting significant undeclared enrichment or reprocessing activities at declared or suspect facilities. Environmental sampling is less able to detect such activities at unidentified facilities, or to provide assurance that no significant, undeclared enrichment or reprocessing activities have occurred in a particular state. Although reliable detection of clandestine reprocessing appears possible at reasonable cost for at least limited regions (assuming that facility operators do not cryogenically trap krypton emissions), reliable detection of clandestine centrifuge enrichment over large areas currently is not possible at reasonable cost. On-site Inspections As discussed in Chapters 2 and 3, routine on-site inspections are vital for confirming declarations of inventories and activities at declared sites. While possibly subject to numerical quotas, such inspections at declared sites should not be subject to refusal by the party being inspected. The inspectors should be guaranteed protection and prompt unimpeded access with any equipment agreed to be necessary for inspection purposes, but agreed limitations on the 18   The network of 26 stations was estimated to be able to detect a change in the U-238:U-235 ratio over 2 percent of the primary focus area, with a 95 percent or greater probability of detection of at least one weekly release at one station and a 5 percent false alarm rate per release.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials extent of access relating to the protection of sensitive weapons information could be acceptable. If evidence of suspect activities at undeclared sites has been discovered by some combination of NTM, human intelligence, wide-area environmental sampling, or information developed in the process of routine inspections and audits of declared activities, challenge inspections would provide a mechanism for clarifying the situation. The right to conduct on-site inspections of suspect sites is provided for in the Additional Protocol to the IAEA safeguards agreements, as well as various other multilateral and bilateral arms control treaties. The modalities of how a challenge inspection would be initiated and conducted would depend on whether an agreement was bilateral or multilateral. An existing example of the mechanism for dealing with such issues is the Joint Compliance and Inspection Commission (JCIC) established by the START I Treaty to resolve questions of compliance. The JCIC and other similar commissions have successfully resolved many issues over the past three decades. In the case of multilateral treaties various procedures have been adopted. In the CTBT, any party can request an on-site inspection for cause but at least 30 of the 51 members of the Executive Committee would have to approve the inspection; in the Chemical Weapons Convention, an inspection would proceed automatically unless it was opposed by a given number of parties. Under the Additional Protocol the IAEA must give notice of inspection at least 24 hours in advance, and give the reasons for the inspection. The rights of inspection are broad, including environmental sampling, radiation surveys, examination of shipping records, visual surveys, and so forth. Protection of “proliferation sensitive information” is provided but this protection must not prevent the agency from collecting credible assurance of the absence of undeclared NEM or of the completeness of the specified declarations. Two questions are frequently asked about challenge inspections. First, would the United States be willing to call for a challenge inspection when the evidence for the “probable cause” could jeopardize sensitive intelligence sources and methods? Even if the source of the original stimulus for suspicion might in fact be very sensitive, we believe corroborative evidence could be made available and locations could be established using commercial satellite photography, which now has good resolution. Second, would a multilateral Executive Committee be willing to authorize an in-

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials spection that could have serious political or military repercussions? Once the United States requested approval from a multinational committee, we believe it should be possible to obtain the necessary support for an inspection when a convincing case could be presented. When necessary, the request could be backed up by sharing relevant classified information privately with other states or publicly as has been done in the past. In practice, the number of challenge inspections for an agreement related to nuclear weapons or NEM should be relatively few. Unless wide-area environmental sampling is used there are no natural events to trigger the system, as is the case with earthquakes in the CTBT. Finally, there is the question of whether any state would permit an on-site inspection that would prove it was in violation of the agreement. Although the challenged country would probably attempt to block or delay the inspection by various means, the international community would undoubtedly see rejection of a challenge inspection, when a serious case for cause had been presented, as an admission of guilt. If the challenged country were indeed innocent of the charges, the inspection would provide an opportunity to clear itself and possibly shift international criticism to the party calling for the inspection in the first place. Detection of Nuclear Weapon Programs by U.S. Intelligence The historical record of U.S. intelligence in identifying foreign nuclear weapon programs is a valuable reference point for evaluating the likelihood that clandestine weapons programs would be detected in the future. At present the record and even some underlying capabilities of U.S. intelligence relating to recent events are the subject of intense controversy. On the specific question of how early the United States became aware of efforts to acquire nuclear weapons, however, the record is good. In every known case, U.S. intelligence became aware of efforts by other countries to develop nuclear weapons relatively early in these programs and well in advance of the actual achievement of a nuclear device. Although there have often been inaccuracies in detailed estimates, such as the date of the initial fabrication of a nuclear device and the total number of weapons assembled as of a given date, the United States has identified key production facilities before a significant amount of NEM or a nuclear weapon was produced. The following is a very brief summary of the U.S. experience: Soviet Union. Despite extreme secrecy, the existence of the Soviet nuclear weapons program became known

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials soon after World War II, and the facilities that eventually produced the first NEM were identified by 1948. Knowledge of the Soviet program emerged from analysis of many unrelated sources ranging from very sensitive NTM to the open literature, which taken together pointed unambiguously to a high-priority project to develop nuclear weapons as rapidly as possible.19 The Soviet Union conducted its first nuclear test sooner than most analysts anticipated, but the purpose and general nature of the program were clear well before the first Soviet test in late August 1949.20 China. U.S. intelligence identified all of China’s major NEM production facilities prior to their operation.21 Limited to the use of satellite and aerial photoreconnaissance by the strict secrecy of Maoist China, U.S. intelligence had difficulty in determining the operational status and capacity of these facilities. Although U.S. intelligence correctly identified the uranium enrichment facility at Lanzhou two years before China tested its first weapon (which used HEU produced by this facility), for example, it underestimated the capacity of the plant and judged that the facility would not be operational for three more years.22 19   Nuclear History Program, Berlin Crisis History Session #8, Interview with Spurgeon M. Keeny, Jr., October 30, 1992, Transcript. Center for International Security Studies at the School of Public Affairs, University of Maryland, pp. 320-321 and Henry S. Lowenhaupt, “On the Soviet Nuclear Scent,” Studies in Intelligence, Central Intelligence Agency, Fall 1967, pp.13-29. Secret document, declassified and approved for release. 20   Roscoe H. Hillenkoetter, “Estimate of the Status of the Russian Atomic Energy Project,” Director of Central Intelligence Memorandum to the President, July 6, 1948, p. 1, and David Holloway, Stalin and the Bomb: The Soviet Union and Atomic Energy (New Haven, CT: Yale University Press, 1994), p. 220. 21   National Intelligence Estimate [NIE] 13-2-60, “The Chinese Communist Atomic Energy Program,” December 13, 1960; National Intelligence Estimate [NIE] 13-2-62, “Chinese Communist Advanced Weapons Capabilities,” April 25, 1962; Special National Intelligence Estimate [SNIE] 13-2-63, “Communist China’s Advanced Weapons Program,” July 24, 1963; Special National Intelligence Estimate [SNIE] 13-4-64, “The Chances of an Imminent Chinese Communist Nuclear Explosion,” August 26, 1964; and National Intelligence Estimate [NIE] 13-2-65, “Communist China’s Advanced Weapons Program,” January 27, 1965. All partially declassified, approved for release and available as of January 2005, at: http://nsarchive.chadwyck.com. 22   National Intelligence Estimate [NIE] 13-2-62, “We believe that the Chinese would at some point in their program endeavor to produce U-235, but we have no evidence of U-235 production at present. Latest evidence indicates that a facility at Lanchou [Lanzhou] suspected of being a gaseous diffusion plant has not been completed. If this plant is in fact intended to be a gaseous diffusion facility, it probably could not produce weapon grade U-235 before 1965. The Chinese could probably test an all U-235 or composite device within a year after the activation of a production facility,” p. 4.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials Israel. The United States photographed what was to become the Dimona nuclear facility under construction in the late 1950s,23 and the existence of a plutonium production reactor at the site was clearly established in the early 1960s.24 Any doubts about the purpose of the program25 were removed when Israel refused to allow the United States to carry out meaningful bilaterally agreed inspections of Dimona.26 India. India’s nuclear test in 1974 came as a surprise,27 but the suspicious nature of the secretive Indian nuclear program was recognized a number of years earlier and the facilities supporting the development of the device had been identified.28 23   Avner Cohen, Israel and the Bomb (New York, NY: Columbia University Press, 1998), p. 83. 24   Dean Rusk to John F. Kennedy, “Israel’s Atomic Energy Activities,” Department of State Memorandum, January 30, 1961, p. 1. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://nsarchive.chadwyck.com. 25   William N. Dale to Department of State, “Current Status of the Dimona Reactor,” Department of State Airgram A-742, April 9, 1965, p. 5. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://nsarchive.chadwyck.com. 26   From 1961 until 1969, U. S. scientists conducted regular (roughly annual) inspections of Dimona. From the start, the Israelis denied access to most of the key areas in the facility. After the July 12, 1969, inspection, the U.S. and Israeli governments reached an agreement to discontinue the procedure. See Avner Cohen, Israel and the Bomb (New York, NY: Columbia University Press, 1998), pp. 329-338 and Memorandum of Conversation, “1969 Dimona Visit,” Department of State, August 13, 1969, p. 2. Secret document, declassified and approved for release. Available as of January 2005, at: http://www2.gwu.edu/~nsarchiv/israel/documents/visit/01-01.htm. 27   The U.S. intelligence community did not detect the preparation for the May 18, 1974, “peaceful nuclear explosion” at the Thar Desert test site near the city of Pokhran prior to the test but were well aware that India possessed a capability to test at any time. See National Security Study Memorandum [NSSM] 156, “Indian Nuclear Developments,” September 11, 1972, p. 1. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://nsarchive.chadwyck.com. 28   George McGhee to Dean Rusk, “Anticipatory Action Pending Chinese Communist Demonstration of Nuclear Capability,” Department of State, September 13, 1961. Available as of January 2005, at: http://nsarchive.chadwyck.com. “India’s atomic program is sufficiently advanced so that it could, not many months hence, have accumulated enough fissionable material to produce a nuclear explosion,” p. 2 and George Perkovich, India's Nuclear Bomb: The Impact on Global Proliferation (Berkeley, CA: University of California Press, 1999), p. 52.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials Pakistan. The fact that Pakistan was engaged in a major effort to develop nuclear weapons was apparent in the mid-1970s, early in the program.29 The location and nature of the facilities supporting the effort were also identified early, including the secret uranium enrichment facility at Kahuta. In response to Pakistan’s covert theft of Urenco centrifuge enrichment technology,30 the United States terminated economic and military aid to Pakistan in 1977 and again in 1979, many years before Pakistan was able to produce significant amounts of HEU and assemble its first weapon.31 South Africa. The goals of the South African nuclear program were long suspect, particularly when in the 1970s South Africa developed the Helikon enrichment process capable of producing HEU.32 The United States cut off nuclear cooperation in 1976 due to South Africa’s refusal to sign the NPT. These suspicions were confirmed in 1977 when the United States was informed by the Soviet Union that one of its satellites had discovered an apparent nuclear test site in the Kalahari Desert.33 29   National Security Study Memorandum [NSSM] 202, “U.S. Non-proliferation Policy,” May 23, 1974, p. V-9. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://nsarchive.chadwyck.com and Swarts to Hartman, “Demarche to Pakistan on Nuclear Fuel Reprocessing,” Department of State, January 30, 1976, p. 1. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://nsarchive.chadwyck.com. 30   State Department Briefing Paper, “The Pakistani Nuclear Program,” June 23, 1983, p. 4. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://nsarchive.chadwyck.com. 31   Tracking Nuclear Proliferation, 1998: Pakistan, “Glenn-Symington Amendment” (Washington, DC: Carnegie Endowment for International Peace, 1998). Available as of January 2005, at: http://www.ceip.org/files/Publications/TrackingPakistan.asp. 32   Directorate of Intelligence, “Prospects for Further Proliferation of Nuclear Weapons,” October 2, 1974, p. 5. Classified memorandum partially declassified and approved for release. Available as of January 2005, at http://nsarchive.chadwyck.com and Directorate of Intelligence, “New Information on South Africa's Nuclear Program and South African-Israeli Nuclear and Military Cooperation,” March 30, 1983, p. 1. Secret document, partially declassified and approved for release. Available as of January 2005, at: http://www.foia.ucia.gov. 33   David Albright, “South Africa’s Secret Nuclear Weapons,” ISIS Report (Washington, DC: Institute for Science and International Security, May 1994). Available as of January 2005, at: http://www.isis-online.org/publications/southafrica/ir0594.html and Mitchell Reiss, “South Africa: Castles in the Air,” in Bridled Ambition: Why Countries Constrain Their Nuclear Capabilities (Washington, DC: Woodrow Wilson Center, 1995), p. 10.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials Iraq. Iraq’s nuclear program became the object of suspicion beginning in the late 1970s. Israel was sufficiently concerned about the ultimate purpose of the Osiraq research reactor that in 1981 it destroyed the reactor in a controversial bombing raid. There was considerable evidence of the buildup of the Iraqi nuclear program during the 1980s, and key facilities were identified and destroyed during the 1991 Gulf War.34 Subsequent inspections determined that the nuclear program, while more extensive than previously believed, had not produced a weapon or significant amounts of NEM. North Korea. North Korea became the subject of suspicion when it began construction in the early 1980s of a small graphite-moderated reactor capable of producing significant quantities of plutonium.35 Although North Korea signed the NPT in 1985, this suspicion grew when it delayed signing the required IAEA inspection protocol until 1992. During this period North Korea initiated construction of much larger graphite-moderated nuclear reactors, as well as a building identified by U.S. intelligence as a suspected reprocessing facility,36 and a suspected site where associated reproc- 34   Iraq had three operating reactors including the Osiraq reactor. Two small research reactors, IRT 5000 rated at 5,000 kilowatts and Tammuz 2, rated at 500 kilowatts, were both damaged by Coalition air strikes in March 1991. See Iraq Profile, “Iraqi Nuclear Facilities: Research Reactors,” The Nuclear Threat Initiative, Available as of January 2005, at: http://www.nti.org/e_research/profiles/Iraq/Nuclear/2117_3362.html. 35   Gary Samore, ed. North Korea's Weapons Programmes: A Net Assessment, An IISS Strategic Dossier (Basingstoke, UK and New York, NY: International Institute for Strategic Studies and Palgrave Macmillan, 2004). “Through satellite reconnaissance, the US detected the construction of the 5MW(e) reactor at an early stage, and was able to confirm initial operation of the reactor in 1986 by noting the emission of steam plumes from its cooling tower, which indicated the reactor was venting excess heat,” p. 35. See also Michael J. Mazarr, North Korea and the Bomb: A Case Study in Nonproliferation (New York, NY: St. Martin’s Press, 1995), p. 40. 36   Gary Samore, ed. North Korea's Weapons Programmes: A Net Assessment, An IISS Strategic Dossier (Basingstoke, UK and New York, NY: International Institute for Strategic Studies and Palgrave Macmillan, 2004), p. 36. In July 1989, the U.S. media described a U.S. and South Korea meeting on the suspected North Korean reprocessing facility at Yongbyon, corroborating earlier South Korean media reports. See, for example, John J. Fialka, “North Korea May Be Developing Ability to Build Nuclear Weapons,” The Wall Street Journal, July 19, 1989, p. A16, and Don Oberdorfer, “North Koreans Pursue Nuclear Arms: U.S. Team Briefs South Korea on New Satellite Intelligence,” The Washington Post, July 29, 1989, p. A9.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials essing wastes were hidden. The first IAEA inspections of the facility in 1992 and 1993 confirmed U.S. suspicions that plutonium had been separated from irradiated reactor fuel.37 Iran. Iran received significant assistance from China for its civilian nuclear program beginning in the mid-1980s, and in 1995 Iran contracted with Russia to complete one of the two partially constructed reactors at Bushehr, which were begun with German assistance in the late 1970s but were severely damaged during the 1980-1988 Iran-Iraq War. The United States pressed China and Russia throughout the 1990s to curtail their nuclear cooperation with Iran, citing intelligence information indicating that Iran intended to develop nuclear weapons. Evidence of a secret centrifuge enrichment program began to emerge publicly as early as 1995, and in August 2002 the existence of a secret nuclear facility at Natanz was revealed by an Iranian opposition group.38 Perhaps in response to these leaks, in 2003 Iran opened what it claimed to be all of its facilities to IAEA inspection, including a pilot centrifuge enrichment plant under construction near Natanz.39 The controversy over Iran’s nuclear program continues, but the relevant point here is that the existence of the NEM production program was discovered well before weapons or significant amounts of HEU were produced. 37   Michael J. Mazarr, North Korea and the Bomb: A Case Study in Nonproliferation (New York, NY: St. Martin’s Press, 1995), p. 84. For additional details see: Samore, North Korea’s Weapons Programmes, “The 1992 Plutonium Mystery,” pp. 36-38. 38   News Bulletin, “Mullahs’ Top Secret Nuclear Sites and Weapons of Mass Destruction Projects,” Iran Liberation, The National Council of Resistance of Iran, August 19, 2004, p. 2. Available as of January 2005, at: http://www.globalsecurity.org/wmd/library/news/iran/2002/iran-020819-ncri_news.pdf. 39   Mohamed ElBaradei, “Implementation of the NPT Safeguards Agreement in the Islamic Republic of Iran,” Report by the Director General, November 10, 2003 (Derestricted November 26, 2004). Available as of January 2005, at: http://www.iaea.or.at/Publications/Documents/Board/2003/gov2003-75.pdf.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials Libya. Libya was long believed to have been laying the groundwork for a nuclear weapons program. Under pressure from the United States and the United Kingdom, Libya disclosed details of this effort at the beginning of 2004, took steps to terminate it, and agreed to accede to the Additional Protocol to the IAEA safeguards agreement.40 Although it had received substantial assistance, the Libyan program had not progressed to the point of producing NEM. Other Programs. In the past a number of other states, including Argentina, Brazil, South Korea, Taiwan, Switzerland, and Sweden, initiated programs directed at developing a nuclear weapon capability. All of these programs were subsequently abandoned when the political leadership recognized that the programs would not be in the overall best interests of those states. In each of these cases, the United States had early indications of the evolving plans, which allowed high-level diplomacy to influence political leaders, who in some cases may not have been fully aware of the actions of secret organizations within their own military and scientific communities. Looking ahead, the technical collection capabilities of U.S. intelligence will continue to increase, as they have over the last half century, particularly with the advent of advanced space-based sensor technology. Although substantial uncertainties have and will continue to surround intelligence estimates, it is important to remember that in the context of an international agreement in which all nuclear facilities were declared, these uncertainties could be resolved by requesting an on-site inspection of the suspect facility. As noted in the previous section on environmental monitoring, an on-site inspection would be highly likely to detect any significant undeclared activities, such as the enrichment of uranium or the separation of plutonium. Based on the historical evidence, and with these additional capabilities, we judge that a clandestine nuclear weapons program very likely would not escape early detection. 40   Mohamed ElBaradei, “Implementation of the NPT Safeguards Agreement of the Socialist People’s Libyan Arab Jamahiriya,” Report by the Director General. May 28, 2004. Available as of January 2005, at: http://www.fas.org/nuke/guide/libya/iaea0504.pdf.

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials CONCLUSIONS As discussed in Chapters 2 and 3, procedures and technology are available to verify with high confidence the declarations of stockpiles of nuclear weapons and NEM at declared sites. Independent knowledge of the actual total size of the Russian stockpiles of nuclear weapons and NEM, however, is sufficiently uncertain that verified declarations alone cannot preclude the possibility that undeclared stockpiles might exist at undeclared locations. Although there is a good chance that the possible existence of undeclared stockpiles would be revealed by a combination of National Technical Means, human intelligence, audits of records, and physical inventories, the theoretical possibility that significant numbers of undeclared weapons had been retained at hidden sites would remain for many years, and perhaps several decades. Undeclared nuclear weapons also could come into existence by the clandestine production of NEM. Past experience indicates a high probability that this would be detected over time. Though not perfect, NTM, human sources, and environmental monitoring form a fairly tight net with which to catch evidence of the large programs and facilities required to produce NEM. Given the extensive knowledge of existing nuclear programs, the massive amounts of additional information that would result from the process of verifying declarations, the new inspection capabilities provided by the IAEA Additional Protocol, and the demonstrated capabilities of NTM, it is very unlikely that any state, including Russia, participating in a cooperative fashion involving detailed declarations could develop a complete, stand-alone covert nuclear weapon production program that would not be discovered over time. If, however, undeclared stocks of NEM exist or can be diverted without detection from civilian stocks or production facilities, it is much more likely that the assembly of new weapons could escape detection. The synergistic effect of the approaches discussed in Chapters 2, 3, and 4 in a cooperative environment, coupled with robust NTM capabilities, would substantially reduce current uncertainties in U.S. assessments of foreign nuclear weapon and NEM stockpiles over time. Nevertheless, in view of the sheer size and age of the Russian stockpile (where current uncertainties amount to the equivalent of several thousand weapons), Russia probably could conceal undeclared stocks equivalent to several hundred weapons. In the case of other countries with much smaller programs, abso-

OCR for page 183
Monitoring Nuclear Weapons and Nuclear-Explosive Materials lute uncertainties would be much less, leading to the possibility that these countries could conceal undeclared stocks equivalent to one or two dozen weapons in the case of China, and at most one or two weapons in the case Israel, India, and Pakistan. Confidence that declarations were accurate and complete, and that covert stockpiles or production facilities did not exist, would be increased by the successful operation of a monitoring program over a period of years in an environment of increased transparency and cooperation.