Australian Safeguards and Non-Proliferation Office

New verification challenges

John Carlson, Director General, Australian Safeguards and Non-Proliferation Office

Paper published in the Journal of Nuclear Materials Management, Summer 2009, Volume XXXVII, No.4, and presented at the Annual Meeting of the Institute for Nuclear Materials Management, Tucson, 12-16 July 2009.

This paper presents the personal views of the author and not necessarily those of the Australian Government.

Abstract

To date, nuclear verification efforts have largely focused on IAEA safeguards pursuant to the NPT. IAEA safeguards face a number of challenges, especially detection of undeclared nuclear activities - this will also be a major challenge for new verification missions, such as fissile material cut-off. In developing new verification missions, important lessons can be learnt from IAEA safeguards experience. This paper briefly outlines some of the new verification issues and challenges on the horizon.

1. INTRODUCTION

Non-proliferation is not just an end in itself, but an essential contribution towards achieving a world free of nuclear weapons. This paper looks beyond the current IAEA safeguards system, and briefly outlines the new verification missions required in support of this broader goal.

The maxim 'trust but verify' is of fundamental importance to nuclear arms control - to efforts to counter the spread of nuclear weapons and ultimately to eliminate them. Non-proliferation - the commitment not to acquire nuclear weapons, and disarmament - the commitment to phase out and eliminate nuclear weapons - both depend on the highest levels of confidence and trust amongst states. Confidence and trust are underpinned by verification - effective verification is essential to achieving a denuclearized world.

To date, nuclear verification efforts have mostly focused on non-proliferation in the non-nuclear-weapon states (NNWS), in the form of IAEA safeguards. The negotiators of the Nuclear Non-Proliferation Treaty (NPT) recognised that an effective non-proliferation regime is a necessary condition to achieving nuclear disarmament - disarmament will not proceed without confidence that new nuclear threats will not emerge. Moreover, an effective non-proliferation regime will be essential in the post-disarmament world, to counter new nuclear weapons programs, whether by former nuclear-armed states or others. Thus the non-proliferation regime will remain the essential foundation for nuclear disarmament - but its application will need to expand from the NNWS to the states that are currently nuclear-armed, i.e. the recognised nuclear-weapon states (NWS) and the states outside the NPT.1

The world has almost four decades experience with a comprehensive multilateral verification system, the IAEA safeguards system established pursuant to the NPT. By and large the IAEA safeguards system has proven successful, but it faces a number of challenges. The most serious of these is ensuring the capability to detect undeclared nuclear activities. Other safeguards challenges include: the potential spread of proliferation-sensitive technologies (enrichment and reprocessing) to further states; the implications of new fuel cycle technologies; and an ever-increasing workload.

These challenges impact on disarmament in two ways: through their implications for the effectiveness of the non-proliferation regime, essential to underpinning disarmament; and because there are parallel issues for disarmament itself - e.g. in the future, how will it be possible to reach credible assurance that none of the current nuclear-armed states has undeclared nuclear material and activities?

Progress beyond current non-proliferation arrangements will involve a range of new agreements and associated verification arrangements. There will not be a single disarmament treaty, but several steps: cut-off of production of fissile material for nuclear weapons; dealing with fissile material stocks; a ban on nuclear testing; reductions in the numbers of deployed nuclear weapons; dismantlement of nuclear weapons; stockpile stewardship issues; and so on. Each of these will involve particular verification challenges.

2. VERIFICATION MODELS2

The IAEA safeguards system has developed over some five decades, and represents the model in peoples' minds when they think about nuclear verification. Key features may be outlined as follows:

To some extent the institutional arrangements, approaches and methods developed for IAEA safeguards will be adaptable to new verification missions. But novel situations are likely to require innovative solutions. Already there are alternative models that might influence future arrangements:

It is likely technical verification will need to be complemented by suitable transparency and confidence-building measures - see section 5 below.

3. IAEA SAFEGUARDS

IAEA safeguards issues are covered in detail by other authors, and will be discussed only briefly here.

In the first two decades of the NPT, as the nuclear industry expanded, the focus of IAEA safeguards was on known (declared) nuclear materials and facilities. The IAEA developed a complex verification system, including: inspection, sampling, analysis and monitoring methods; safeguards standards; safeguards equipment; and safeguards performance evaluation. Emphasis was placed on nuclear material accountancy, with containment and surveillance described as 'complementary' measures. A particular interpretation of non-discrimination led to uniformity in safeguards implementation, resulting in inspection resources being concentrated in states with substantial nuclear programs regardless of any proliferation risk analysis.

Since the early 1990s, the illicit spread of sensitive nuclear technologies, especially centrifuge enrichment, has shifted attention to developing the capabilities needed for detection of undeclared nuclear activities. This has prompted fundamental changes in safeguards - including greater use of information, a broadening of verification activity, and the drawing of more qualitative conclusions. At the same time, the growth in workload and in complex facilities has led to containment and surveillance measures assuming greater importance.

Especially important has been the development of 'information-driven' safeguards. Safeguards are moving from uniformity to a state-level approach, under which decisions on verification intensity can reflect state-specific factors. The state-level approach is essential to optimising effectiveness and efficiency, enabling scarce safeguards resources to be deployed to areas of highest priority.

The greatest technical challenge facing IAEA safeguards is establishing a reliable capability of detecting undeclared nuclear activities. The IAEA's technical skills are increasing - but the Agency cannot be expected to find undeclared nuclear activities unaided. Member states have given the Agency vital technical assistance in development of and training in equipment, detection technologies (such as sensors and satellite imagery) and so on. But more is needed in the area of information-sharing. States have substantial information, including intelligence ('national technical means') and export data (encompassing both items supplied and items denied).4 Detecting undeclared nuclear activities - or providing credible assurance of their absence - requires an active partnership between the Agency and states.

Other challenges for the IAEA safeguards system include: the potential spread of proliferation-sensitive technologies to further states; the consequences for safeguards of new fuel cycle technologies (leading to an increasing reliance on containment and surveillance relative to material accountancy); and an ever-increasing safeguards workload.

More than ever, it is essential for safeguards to be complemented by other non-proliferation measures, such as transparency and confidence-building mechanisms. These are discussed in section 5.

4. NEW VERIFICATION MISSIONS

4.1 Comprehensive Nuclear-Test-Ban Treaty

The CTBT prohibits all nuclear explosions, and establishes a verification regime to detect and investigate possible non-compliance. The Treaty serves both non-proliferation and disarmament objectives - complementing the NPT by increasing the difficulty for a NNWS to produce reliable nuclear weapons, and limiting the ability of states with nuclear weapons to develop new warhead designs.

The Treaty was opened for signature in 1996 - to date it has been signed by 180 states and ratified by 148. It will not enter into force until ratified by the 44 states listed in Annex 2 of the Treaty. Nine of these have yet to do so - China, DPRK, Egypt, India, Indonesia, Iran, Israel, Pakistan and the US.

Although the CTBT is not yet in force, the Treaty stipulates that its verification regime must be capable of meeting the requirements of the Treaty when it does enter into force. Most (more than 80%) of the Treaty's International Monitoring System (IMS) has been installed and is in provisional operation. The IMS will comprise 321 seismic, radionuclide, infrasound and hydroacoustic monitoring stations and 16 radionuclide laboratories in 89 states. IMS stations are operated by national agencies, mostly under contract to the Treaty organisation.5 On-site inspection arrangements are also under development.

Conceptually CTBT verification is quite different to the IAEA safeguards model. Judgments on compliance, based on technical verification data, are made by the Treaty Parties themselves rather than the Treaty secretariat. There are no regular inspections - rather, the IMS is looking for events that, it is hoped, will never occur. Hence there is no standing inspectorate - if an on-site inspection is required, an ad hoc inspection team would be drawn by the Organization from a cadre nominated by the Treaty Parties. Data from the IMS stations are available to all Treaty Parties, via the Organization's International Data Centre in Vienna. Any Party therefore is in a position to analyse IMS data and to seek clarification of an event, including through an on-site inspection.

Despite concerns voiced by some in the US about the CTBT's technical effectiveness, the IMS even in its partially complete state proved effective in the detection of the Indian and Pakistani nuclear tests of 1998 and the DPRK test of 2006. In addition, the IMS performed well with relatively small scale conventional explosions conducted in Kazakhstan for testing the system in 1999 (100 tonnes) and 2002 (12 tonnes). The main challenge for the CTBT is political, not technical - to gain the outstanding ratifications to enable the Treaty to enter into force. US leadership will be vital here.

4.2 Fissile Material Cut-off Treaty

The objective of the proposed FMCT is to ban production of fissile material for nuclear weapons and other nuclear explosive devices. The negotiating mandate agreed in the UN Conference on Disarmament (CD) is for the FMCT to be 'non-discriminatory, multilateral and internationally and effectively verifiable'. Negotiations in the CD have been blocked for over a decade, but it is hoped statements by the Obama Administration in support of a verifiable treaty will help break the log-jam.6

Since negotiations have not yet commenced, the main provisions of the FMCT cannot be described with certainty. The principal area to be settled is the treaty's scope - which materials and facilities will be covered? As a minimum, the treaty would apply to new (i.e. post entry-into-force) production of fissile material. Verification would apply to:

Just how extensive the coverage of the treaty and the verification arrangements should be - whether there is a need to include other materials and facilities to ensure the effectiveness of the verification system - is yet to be determined. One area of contention is whether the FMCT should also apply to existing fissile material - the issue of stocks. Clearly stocks must be addressed, since otherwise the treaty's effectiveness in capping nuclear arsenals will be limited, but for the purposes of this paper it is assumed stocks will be the subject of further negotiations. Stocks are discussed further in section 4.3 following.

Whatever the scope of the FMCT, it can be seen from the above outline that in concept the treaty should be similar to the current safeguards system:

The verification challenges for the FMCT are expected to be:

As with IAEA safeguards, detection of undeclared fissile material production will be a major challenge. It can be expected that the NWS have extensive information on each other's nuclear programs - sharing of information with the inspectorate will be essential. Unlike the current IAEA system, where information held by the Agency is confidential, there may well be a need for sharing of some kinds of verification information, as is the case with the CWC and the CTBT. Almost certainly formal verification activities will have to be complemented by transparency and confidence-building measures, possibly including mutual inspections and arrangements such as 'Open Skies' - see section 5.

A mechanism for initiating investigative inspections will be required. Disappointment over the lack of use of the IAEA's special inspection mechanism - by which investigative inspections are supposed to be initiated by the Agency - means states are unlikely to have confidence in such a mechanism. It is likely that a challenge inspection mechanism, which can be initiated by a party, will be required for the FMCT, either instead of, or as well as, inspections that can be initiated by the inspectorate.

One area requiring considerable development is that of verification standards and intensity for the FMCT. In the case of horizontal proliferation, the diversion of relatively small quantities of fissile material will be enough for a state to change its status from a NNWS to a nuclear-armed state. The sensitivity of IAEA safeguards - reflected in technical parameters such as goal quantities (e.g. the significant quantity of 8 kg plutonium), detection probability, timeliness goals, and inspection frequency - has been set accordingly.

For the states that already have nuclear weapons, however, the calculus is rather different. These states will be concerned about treaty violations that are of sufficient scale to alter strategic relativities: for a state with hundreds of weapons, it might take a violation of hundreds of kilograms to be strategically significant. On the other hand, for a state with a small arsenal - and the objective of disarmament is that all the NWS will progress to this situation in time - small-scale diversion will be significant. One approach that would meet both these cases is to regard a breakout equivalent to say 1% of the monitored inventory as a strategic change.7 These considerations are likely to be reflected in the development of a state-level approach to verification, building on experience being gained with the state-level approach in IAEA safeguards. With a state-level approach, the technical verification objectives and parameters will be the same for all states, but decisions on verification intensity would take account of state-specific factors.

Relationship to comprehensive safeguards

It is envisaged that the FMCT will be a universal treaty, i.e. the parties will not be limited to the nuclear-armed states, but will also include the NNWS. In most respects the comprehensive safeguards agreements (CSAs) which NNWS conclude under the NPT should be sufficient to meet the requirements of the FMCT, and additional verification should not be necessary.

However, negotiation of the FMCT provides an opportunity to address some important issues facing the IAEA safeguards system. For example, a concern with NPT safeguards agreements is that their duration is tied to the state's membership in the NPT. If the state withdraws, the CSA lapses. Consideration could be given in the FMCT to an irreversibility provision, that once nuclear material and facilities become subject to peaceful use commitments they would retain this status in perpetuity. Another improvement over current CSAs would be introduction of a challenge inspection mechanism.

4.3 Fissile material stocks

Irreversible submission of excess fissile material stocks to a verified commitment against explosive use will be essential to complement the proposed FMCT and further disarmament steps. As long as there remain significant fissile stocks outside any such commitment that can be drawn on to produce new nuclear weapons, there will be concerns about the durability of limits on the number of weapons. If excess stocks are not covered by the FMCT - they are not included in the current negotiating mandate - a further agreement covering stocks will be needed in due course.

In general there are no particular verification challenges with stocks, other than workload. Parties will declare fissile stocks they regard as excess to military requirements, and these would be verified - initially and on a continuing basis. While these stocks might be stored for a period, it is expected that in due course they would enter the civil fuel cycle, or would be conditioned for disposal. Since the declarations would be voluntary, there would be no requirement for verification activities to establish the completeness of the declarations (i.e. that there are no undeclared materials). However, as a transparency measure, the nuclear-armed states might declare the quantities of fissile materials they hold outside verification.

Subsequent steps

The verification task becomes more complicated if it is decided, as disarmament progresses, that the nuclear-armed states should declare and submit to verification all fissile material except material in the form of declared nuclear weapons. This would raise two issues:

4.4 Naval propulsion programs

This is raised as a problem for FMCT, but it is also a potential problem for current IAEA safeguards. CSAs allow NNWS to remove from safeguards nuclear material intended for non-proscribed military use, under arrangements to be agreed with the IAEA.8 For CSAs, this provision has not been put to the test, as to date no CSA state has introduced nuclear naval propulsion, but Brazil has indicated an interest in acquiring nuclear-powered submarines.9

Some NWS operate naval reactors with HEU fuel, so there is the possibility that HEU production could continue under the FMCT for this non-proscribed military purpose - although it would be an advantage in terms of verification tasking if the states concerned concluded that their existing HEU stocks were sufficient and they had no need for ongoing HEU production.

The problem for verification arises because states with naval reactors regard the design of naval fuel, and factors such as core loadings and range between refuelling, as highly classified. While concern about security is understandable, it is essential to develop appropriate verification arrangements so that naval programs don't present an opportunity for diversion. Diversion is not an issue just for HEU fuel - because LEU could be used as feedstock for high enrichment in an undeclared facility, verification arrangements would also be needed for LEU-based naval programs. And, as noted in section 4.3, the need for verification might apply not only for new production for naval programs, but for existing HEU stocks.

Because of the sensitivities, verification for naval programs will require novel approaches. However, the problems are not insurmountable - the Trilateral Initiative10 between the US, Russia and the IAEA demonstrates the practicability of innovative approaches to verifying fissile material of sensitive composition, shape and mass. Formal verification may be complemented by transparency arrangements, e.g. it is easy to check that a vessel is at sea (and therefore has been fuelled).

4.5 Stockpile stewardship activities

A complication for FMCT is that some nuclear-armed states may require to recycle plutonium from nuclear weapons - either to remove the build-up of americium-24111 (a practice known as 'clean-up'), or to manufacture new weapons (it is not expected that FMCT would prohibit production of new weapons from fissile material in existing weapons). In the past, NWS have simply withdrawn warheads as they reach the end of their service life and produced further plutonium to replace them (one reason why the NWS have accumulated so many warheads), but under FMCT this will no longer be an option.

Verification measures for FMCT will need to be able to distinguish between recycle of existing plutonium and new separation of plutonium from irradiated fuel or targets - new separation for weapons would be prohibited. For national security reasons, access to recycle facilities will not be allowed. But measures such as sampling of gaseous emissions would readily indicate if new separation was being undertaken.

4.6 Nuclear weapon reductions and dismantlement

The US and Russia have substantial experience with mutual verification of numbers and class of deployed strategic warheads, though to date these agreements have dealt with delivery systems rather than weapons as such. Nonetheless, this experience can be applied to new arrangements to verify declarations, where numbers of weapons deployed or stored are declared, and numbers of weapons to be dismantled are declared. Far more challenging would be verification measures for completeness of declarations, i.e. that there are no undeclared nuclear weapons. This is touched on in section 4.7 following.

While there is no direct experience of verifying dismantlement - that weapons committed for dismantlement are in fact dismantled and the fissile material placed under non-explosive/verification commitments - verification requirements have been studied in the US12 , and are currently under study by the UK and Norway.

Conceptually, what is required is continuity of knowledge (or 'chain of custody') over the item being dismantled and its fissile core. Verification elements would include:

The Trilateral Initiative, mentioned in section 4.4, has demonstrated that fissile material can be satisfactorily verified against unclassified attributes without revealing sensitive information to inspectors.

Following dismantlement, the fissile core will need to be processed to alter mass, shape and probably isotopic composition, e.g. by blending in a facility under 'black box' arrangements. This will require knowledge of all materials entering and leaving the process area. There will be complexities - e.g. how to ensure that information on inputs of unclassified blending material doesn't allow calculation of sensitive details of the weapons material - but the problems are not insurmountable.

4.7 Verification measures against undeclared fissile material

Today nuclear-armed states are not required to declare military fissile material holdings, but this will be required as disarmament progresses - raising the issue of how to ensure that declarations are complete. Detection of possible undeclared fissile material holdings is by far the greatest verification challenge, particularly in the NWS where nuclear weapon programs have operated for decades and have involved many facilities and large quantities of nuclear material.

There is some experience with verifying that all nuclear material in a weapon program has been accounted for, when South Africa renounced nuclear weapons and joined the NPT in 1991. The South African program had been very small - only six nuclear weapons, using HEU, had been produced. The IAEA's verification that all nuclear material was accounted for is regarded as a success, but the practical difficulties of achieving accuracy were significant.

Conducting such an exercise with much larger programs will be very difficult. This is illustrated by the historical study of plutonium production in the US military program, published in 1996.13 This study showed 'inventory differences' of some 2,800 kg, or 2.5%, in military plutonium accounts over the period 1944-94. These differences were mainly attributed to measurement uncertainties, and 68% of the differences occurred in the period prior to the late 1960s.

A similar study was conducted by the UK and published in 2000.14 This study concluded that 'Overall, confidence in the completeness and accuracy of the information available is very high for the 1980s and 1990s, but less so before the mid 1960s.' The study found a discrepancy of 290 kg of plutonium, or 1.7%, between recorded movements into and out of the Atomic Weapons Establishment at Aldermaston. In fact this discrepancy was a surplus - more plutonium was found than expected - but the error margin shows the difficulties of establishing accurate nuclear accounts for historical periods.

It could be some time before verification of the absence of undeclared warheads or fissile material in the NWS becomes critical. By then it is to be hoped that transparency and confidence-building amongst these states will have progressed to the extent needed to provide confidence in the formal verification results.

5. TRANSPARENCY AND CONFIDENCE-BUILDING MEASURES 15

There are limits to what verification can deliver - e.g. verification cannot provide definitive assurance about a state's future intent; and it is difficult to prove a negative, i.e. to verify the absence of undeclared nuclear material and activities. Accordingly, technical verification measures are not likely to be sufficient in themselves to build the required confidence as disarmament proceeds. It is likely technical verification will need to be complemented by suitable transparency and confidence-building measures.

The purpose of such measures can be described as 'to introduce transparency and thereby predictability in relations between states by clarifying national intentions, reducing uncertainties about national activities, and/or constraining national opportunities for surprise.'16 These measures can be grouped as follows:

There is insufficient space here to discuss the range of possible confidence-building measures. Proposals for further strengthening the non-proliferation regime - which will also be relevant to the nuclear-armed states as the process of disarmament proceeds - include:

Transparency and confidence-building arrangements will need to be negotiated amongst the parties most concerned, and may be specific to particular bilateral or regional situations. Possibilities could include:

6. CONCLUSIONS

A number of the prospective new verification missions discussed here are analogous, technically, to IAEA safeguards - although the sensitivity of some of the materials and facilities will require innovative approaches, as seen already with the Trilateral Initiative. In establishing new verification arrangements it will be important to draw on IAEA experience to the extent appropriate.

As new kinds of agreements are negotiated, the parties will need to address not just technical aspects but also institutional arrangements. Issues to be considered include:

As discussed in section 5, transparency and confidence-building mechanisms are expected to have an essential role in complementing formal verification arrangements.

Author's biography

John Carlson is Director General, Australian Safeguards and Non-Proliferation Office, a statutory office which he has held since 1989. He is a graduate in law from the University of Sydney with post-graduate qualifications in jurisprudence and modern logic. He has been a member of the Australian Government's Senior Executive Service since 1976. He chaired SAGSI (the IAEA's Standing Advisory Group on Safeguards Implementation) from 2001 to 2006. He is Alternate Governor for Australia on the IAEA Board of Governors, and a member of the Advisory Board of the International Commission on Nuclear Non-Proliferation and Disarmament. He is a Fellow of the INMM and has received the INMM's Distinguished Service Award.

1. In this paper, unless otherwise indicated, the term 'nuclear-armed' is used to cover both the NWS and the non-NPT states.

2. For a more detailed analysis of IAEA safeguards relative to other treaty verification systems see the author's 'Experience and challenges in WMD treaty verification: a comparative view', in Verifying Treaty Compliance, pp. 213-234, ed. R.Avenhaus et al, Springer, Berlin, 2006.

3. Brazilian-Argentine Agency for Accounting and Control of Nuclear Materials.

4. Knowledge of procurement attempts can be invaluable to the IAEA's understanding of a state's nuclear program, declared or otherwise.

5. Currently the Preparatory Commission for the CTBT Organization.

6. For a discussion of FMCT verification issues see the author's 'Can a Fissile Material Cut-Off Treaty be Effectively Verified?', Arms Control Today, January/February 2004, 25-9.

7. The 1% figure was informally adopted in the Trilateral Initiative, see section 4.4.

8. See INFCIRC/153 paragraph 14.

9. Canada considered nuclear-powered submarines in the 1980s but did not proceed.

10. See e.g. T.E.Shea, 'The Trilateral Initiative: A Model for the Future?', Arms Control Today, May 2008.

11. Pu-241, with a half-life of 14.4 years, decays to Am-241.

12. See e.g. E.R.Gerdes, R.G.Johnston and J.E.Doyle, 'A Proposed Approach for Monitoring Nuclear Warhead Dismantlement', Science and Global Security, Volume 9 pp 113-141.

13. 'Plutonium: The First 50 Years', US DOE 1996.

14. 'Plutonium and Aldermaston - An Historical Account', UK Ministry of Defence, 2000.

15. For a discussion of transparency in IAEA safeguards, see J.Larrimore, M.Kratzer, J.Carlson and B.Moran, 'Transparency and Openness: Roles and Limitations in the Nuclear Nonproliferation Verification System', Journal of Nuclear Materials Management, Fall 2006, Volume XXXV, No. 1, pages 36-51.

16. Definition suggested by United Nations Institute for Disarmament Research (UNIDIR).

17. Under the CWC (Art. IX.17) an inspection can be disallowed by a 3/4 vote of the Executive Council.