The material that sustains nuclear reactions to produce nuclear energy can also be used to make nuclear weapons, and so the development of nuclear energy by a non-nuclear weapons state is considered one of multiple pathways to potential proliferation and represents a subset of issues within nuclear nonproliferation. The risk of proliferation through the nuclear energy development pathway can be described as containing several components: an intrinsic component (e.g., the amount of weapons-usable material produced throughout the particular fuel cycle) and many external components such as host-state factors (e.g., political and regional stability and technical capability) and facility design details (layout of buildings and application of safeguards).
Assessments associated with proliferation risk are performed using subject-matter experts (both technical and nontechnical) who can provide:
• quantifiable data for intrinsic aspects of the fuel cycle and facility data, if the data are available, and
• opinions and knowledge about host-state factors.
This study considers how the current methods of quantification are being used and implemented. It also considers if existing or new methods of quantification can be extended to account for host-state factors. Methods that have been proven to have impact in other complicated systems that include human decisions (such as nuclear safety) are of particular interest.
The Department of Energy Office of Nuclear Energy (DOE-NE) and the National Nuclear Security Administration (NNSA) Office of Nonproliferation and International Security requested that the National Academy of Sciences perform a study to understand the extent and limitations of technical analysis of proliferation risk of nuclear fuel cycles for use by nonproliferation policy makers and decision makers. DOE-NE would like guidance in assessing proliferation risk as one of several critical factors guiding U.S. potential future choices for nuclear fuel cycles (by programmatic decision makers). NNSA would like guidance for better characterizing the near-term and future proliferation risk associated with the use of nuclear technology and nuclear power throughout the world (for nonproliferation policy makers). Both sponsors are considering how technical assessments of foreign host-state proliferation risk contribute to improved programmatic research decisions, nonproliferation policy decisions, and communication between U.S. government agencies, the public, and international partners.
Six findings and three recommendations emerged from this study. These are grouped with the relevant tasks and are presented and summarized below. Further supporting information can be found in the body of the report.
DEFINITIONS AND TERMINOLOGY
The International Atomic Energy Agency (IAEA) defines proliferation resistance as
The characteristics of a nuclear energy system that impede the diversion of undeclared production of nuclear material or misuse of technology by states in order to acquire nuclear weapons or other nuclear explosive devices. . . . The degree of proliferation resistance results from a combination of, inter alia, technical design features, operational modalities, institutional arrangements and safeguards measures. (IAEA 2002).
Proliferation resistance assessments frequently break the fuel cycle down into individual processing steps and follow a specified framework to assess the proliferation resistance at each step. These frameworks typically contain predefined lists of detailed attributes of the fuel cycle and its processing steps and a predetermined approach for scoring and combining the attributes to determine the cycle’s overall proliferation resistance. Attributes can be intrinsic (e.g., inherent to the nuclear fuel cycle’s materials) or extrinsic (e.g., external to the fuel cycle such as safeguards and operations). The committee has termed these methodologies “predefined frameworks.”
Proliferation risk is a more complex concept than proliferation resistance and includes analysis of country-specific issues. Although it does not have an internationally accepted definition, some analysts in the proliferation resistance assessment community have attempted to define proliferation risk through an equation that relates proliferation risk to three terms: the probability that a host state will choose to proliferate along a particular or multiple pathways (L), the probability of success along that path (P), and the consequences of proliferation (C) (NSSPI 2010, NRC 2011a, Charlton 2012). A full risk assessment sums over all possible pathways—a difficult and changing task given that a motivated host state may continuously invent new pathways, which affects the probabilities that attach to some of the above-defined terms. Country-specific factors such as motivation, technical capability, access to technology, and intent are needed to assess L (and would likely include intelligence information). Proliferation resistance could contribute to P and L. Because proliferation resistance is one component of the problem of proliferation risk and because it is not infinite (there is no proliferation-proof fuel cycle), other country-specific factors will determine whether and how a host state proliferates.
FINDINGS AND RECOMMENDATIONS
TASK 1: IDENTIFY KEY PROLIFERATION POLICY QUESTIONS CAPABLE OF BEING ANSWERED BY A TECHNICAL ASSESSMENT OF THE HOST STATE PROLIFERATION RISK POSED BY A GIVEN NUCLEAR FUEL CYCLE, AND DISCUSS THE UTILITY OF THESE QUESTIONS FOR INFORMING INTERNATIONAL NONPROLIFERATION POLICY DECISIONS.
FINDING 1.1: Technical assessments related to aspects of proliferation risk do make valuable contributions to nonproliferation policy decisions on a broad range of topics such as peaceful international nuclear cooperation, export control, nuclear fuel cycle R&D, and nuclear safeguards. However, technical assessments do not fully answer nonproliferation policy questions. Final decisions also include consideration of a much broader set of political, security, economic, and cultural issues.
The committee approached the study by first understanding a broad set of nonproliferation decisions related to the nuclear fuel cycle, and the factors that U.S. and international decision makers take into account when making these decisions. The types of decisions that require an assessment of proliferation risk include peaceful international nuclear cooperation agreements and treaties, nuclear export control decisions, international safeguards, domestic regulatory decisions and domestic nuclear fuel cycle research. The committee notes that in these cases proliferation risk is one factor among many that policy makers and decision makers must take into account. Indeed, political, economic, environmental, and safety considerations are almost always relevant and, depending on the circumstances surrounding the decision that needs to be made, may be given more weight than issues related to nonproliferation in arriving at the final outcome.
While acknowledging the importance of nontechnical factors, the committee found that for all of the nonproliferation policy issues above, technical assessments related to proliferation risk were routinely used—directly or indirectly—to inform policy makers and decision makers. Most technical assessments requested by nonproliferation policy makers are designed to inform on a specific issue and consequently are focused on a particular nuclear technology or capability in the context of a specific country or region or terrorist group. For these proliferation assessments, multidisciplinary teams of experts, often in close collaboration with the intelligence community, tailor their analysis to the specific question, including systematic assessment of both technical and country-specific issues. While addressing a very specific topic or question, these types of analyses are adapted to the variety of topics that require assessment and are the most commonly used method to inform policy makers and decision makers on technical issues (including issues that extend beyond the proliferation risk associated with nuclear fuel cycles). This approach is not unique to nuclear nonproliferation policy issues. In order to differentiate it from the predefined framework or risk assessment approaches discussed below, the phrase “case-by-case” assessments is used throughout this report.
Among the broad set of nonproliferation-related topics requiring technical assessment, questions about host state proliferation risk of a given nuclear fuel cycle represent a small subset. The technical assessment of proliferation resistance represents an even smaller subset. Regardless, predefined framework methodologies have been developed specifically for assessing the relative proliferation resistance of a given nuclear fuel cycle. They can and have been used to address questions such as:
• Are there significant differences in resistance to proliferation (e.g., time to breakout, cost, physical barriers, safeguard ability, or transparency) associated
with different types of potential future fuel cycles compared with those that exist today?
• Can extrinsic measures, such as physical security and international safeguards and/or intrinsic measures, such as reactor design, or material composition, or new operational concepts significantly increase resistance of a particular nuclear fuel cycle?
• For a given nuclear fuel cycle or facility, where are barriers to proliferation lowest? Where can safeguards be most effective in raising these barriers?
Case-by-case assessments can be, and have been, used to address these same questions.
TASK 2: ASSESS THE UTILITY FOR DECISION-MAKERS OF EXISTING AND HISTORICAL METHODOLOGIES AND METRICS USED BY DOE AND OTHERS (SUCH AS THE INTERNATIONAL ATOMIC ENERGY AGENCY) FOR ASSESSING PROLIFERATION RISK, BOTH FOR CONSIDERING THE DEPLOYMENT OF THESE FACILITIES DOMESTICALLY AND THE IMPLICATIONS OF DEPLOYMENT OUTSIDE OF THE UNITED STATES.
FINDING 2.1: Predefined frameworks have been developed and used to assess the proliferation resistance of partial or full nuclear fuel cycles. These methods provide a useful framework for comparing the intrinsic metrics or “attributes” of existing and potential future nuclear fuel cycles and for identifying where safeguards can be most effective in raising barriers to proliferation. However, these comparisons address a small subset of the wider range of issues faced by policy makers and the committee was able to determine that the frameworks have rarely been used to inform policy decisions. Additionally, there have been shortcomings in their execution.
In addressing this task, the committee notes that the “utility” of a methodology is subjective and dependent on the individual and/or organization. The committee considered the frequency of use as an indirect measure of utility and discussed the apparent impact of the methodologies’ results with policy makers and decision makers.
Predefined frameworks are methodologies designed to consistently and transparently evaluate proliferation resistance of a nuclear fuel cycle through a standardized set of predetermined attributes applied to the nuclear fuel cycle’s processing steps.1 However, these frameworks are not objective models or simulations of nuclear facilities, operations, safeguards, material flows, or proliferation pathways. Rather, they use expert knowledge as the source of data to score the attributes at each step.2 Such data
1 In 2000, the Technical Opportunities to Increase the Proliferation Resistance of Globa Civilian Nuclear Power Systems (TOPS) Task Force formulated a set of qualitative attributes (barriers) relevant to assessing the proliferation resistance of nuclear fuel cycles and termed the approach “attributes methodologies.” The committee has used the term “predefined frameworks” to account for how the attributes are combined.
2 The committee reviewed a broad set of current predefined frameworks starting with the original approach defined by TOPS and including more developed methodologies such as the Generation IV International
are gathered, weighted, and combined into higher-level scores to give an overall measure of proliferation resistance (usually high, medium, or low). Typical attributes of proliferation resistance include technical difficulty of proliferation, cost required to overcome barriers to proliferation, time needed to proliferate, and the probability of detection. The process of scoring, aggregating, and arriving at a final assessment of proliferation resistance may give the impression of objectivity often associated with a model while masking the reliance on subject-matter experts.
When considering that future fuel cycles would be deployed between 20 and 30 years in the future, evaluations necessarily have been for a conceptual future fuel cycle, without details about facility design, operational concept, or the country in which it might eventually be deployed. Therefore, the intrinsic details of the fuel cycles were the main focus of these evaluations with assumptions necessarily made about the type of future host state and the facility’s future physical and operational details. Intrinsic attributes of proliferation resistance assessments do contain measures that can be considered robust, such as the expected decay rates of the radionuclides within the proposed fuels or required materials. The details of the extrinsic components (safeguards, inspections, and facility operational details)—an equally important component of proliferation resistance—can significantly change with time and affect the overall resistance of the fuel cycle. For example, large distances between facilities requiring transportation of material without adequate monitoring or the details of how pipes have been laid between processes with low proliferation resistance will significantly lower barriers to material diversion by a motivated host state. Additionally, the inclusion of country-specific factors will alter the assessment results when the broader concept of proliferation risk is considered.
The committee found only a few examples in which predefined framework assessments were performed with the goal of guiding decisions about research and development (R&D) for future nuclear energy systems (Bari et al. 2007, 2008b; DOE 2008b).3 However, the committee found several examples in other domains within the U.S. government in which decision makers use predefined framework-like tools to inform decisions. Examples include prioritizing countries for engagement on nuclear, chemical, and biological security (the Office of Cooperative Threat Reduction within the Department of State; Dolliff 2012), and using risk assessment methods for optimizing architectures for global nuclear detection (the Domestic Nuclear Detection Office within the Department of Homeland Security; Streetman 2012). These examples show that some policy and decision makers do find utility in frameworks that can deconstruct complex problems into their component parts. In these instances, the policy and decision makers were actively involved in the analysis process and not interested only in the final results. The frameworks provide a structure for organizing complex problems with a large number of variables and assessing which factors are most important to the results.
Forum Proliferation Resistance and Physical Protection scenario-based assessment and Texas A&M’s multi-attribute utility analysis approach. For the full list and discussion of the analysis, see Chapter 3 in the report.
3 These reports were not directly cited by policy or decision makers, so it is not clear to the committee if and how actual decisions were made based on these reports.
The committee considered a set of predefined frameworks and found the following shortcomings in their execution:
• poor and/or undocumented expert elicitation processes, and
• lack of sensitivity and uncertainty analyses
Inherent limitations of applicability included:
• for future fuel cycles, unknown facility and host-state details, and
• limited shelf life of assessments
TASK 3: ASSESS THE POTENTIAL FOR ADAPTING RISK ASSESSMENT METHODOLOGIES DEVELOPED IN OTHER CONTEXTS (SUCH AS SAFETY AND SECURITY) TO HOST STATE PROLIFERATION RISK ASSESSMENTS-INCLUDING BOTH QUALITATIVE AND QUANTITATIVE APPROACHES—THEIR BENEFITS, LIMITATIONS, AND THE CHALLENGES ASSOCIATED WITH ADAPTING THESE METHODOLOGIES TO PROLIFERATION RISK ASSESSMENT.
FINDING 3.1: Some of the identified deficiencies in the implementation of the existing predefined frameworks for assessing proliferation resistance could be improved by adopting expert elicitation and data-gathering practices developed by other fields of risk assessment. For example, probabilistic risk assessment (PRA) methods have well-defined and established processes for gathering, quantifying, analyzing, and presenting data that could benefit the nonproliferation community. However, the challenges of an adaptive adversary and of lack of data on proliferation events in particular limit applicability of all of the risk-based methodologies to the assessment of proliferation risk considered by the committee.
Probabilistic risk assessment (PRA) methods provide a compelling example of how a risk assessment methodology can be applied to engineered systems. PRAs have developed proven processes for gathering, quantifying, analyzing, and presenting data. Additionally, PRA experts have significant and relevant experience identifying data that might otherwise be overlooked. Predefined framework methodologies could be improved by adopting the practices established by PRA practitioners, especially regarding the elicitation of expert knowledge and accounting for uncertainties.4
The committee is aware of efforts to understand factors that may influence a country to pursue nuclear weapons, including some efforts to develop models of how factors such as security concerns, type of government, and technical capability influence decisions (Way 2012; Gartzke 2012; Coles et al. 2009). The committee also found that
4 For example, expert elicitation practices that have proven to be effective in gathering knowledge and assigning uncertainties have been defined in (Kotra et al. 1996; EPA 2009).
there is a significant amount of research being performed on modeling an adaptive adversary (e.g., risk-based, social science-based, game theory-based).
Assuming that the limitations in execution are addressed, the assessment of proliferation resistance via predefined framework methodologies can provide a relative comparison between nuclear fuel cycles of the difficulty to evade barriers. However, two significant challenges exist in extending predefined framework methodologies’ assessment of proliferation resistance to host-state proliferation risk assessment via the application of risk-based methodologies: 1) assessing host state factors such as motivations and intent and 2) the lack of data on the process of proliferation. Since the process of proliferation is generally carried out in secret, detailed data on the factors relevant to the probability that an adversary would choose to proliferate along a particular path or the probability of success along that pathway are not readily available. The challenge of assessing the changing conditions of a host state exacerbates these problems. Choices about proliferation pathways could change over time, based on the ability to overcome safeguards and avoid detection, as well as the ability to acquire technology covertly. The resources a host state might allocate to achieving success would depend on the strength of its motivation, which in turn would depend on its perception of the benefits of nuclear weapons. Because of the limited number of cases of proliferation in general, data on all these factors are very limited. Historical case studies provide some relevant data about countries that have given up nuclear weapons programs in the past, but limited data are available on successful efforts.
Although there is a significant amount of work being performed on the modeling of adaptive adversaries (e.g., probabilistic risk assessment, game theory), the committee did not see results that were supported by evidence of success in real-world applications and therefore does not endorse the inclusion of this assessment into the predefined frameworks’ analysis at this time. Furthermore, the committee notes that although increased proliferation resistance can raise barriers to proliferation, a highly motivated, technically capable state will be able to overcome them.
RECOMMENDATION 3.1: DOE-NE and NNSA should consider whether elements of a formal PRA approach could improve multidisciplinary assessments of proliferation risk, especially the quantification of uncertainties. Although the committee concluded that work on understanding motivations to develop nuclear weapons and modeling an adaptive adversary do not have evidence-based records of success in real-world situations, it supports the inclusion of such approaches into proliferation risk analysis when and if they have an established quantitative basis.
TASK 4: IDENTIFY R&D AND OTHER OPPORTUNITIES FOR IMPROVING THE UTILITY FOR DECISION MAKERS OF CURRENT AND POTENTIAL NEW APPROACHES TO THE ASSESSMENT OF PROLIFERATION RISK.
FINDING 4.1: The committee has identified several specific applications as opportunities in which the current predefined frameworks could provide value and utility to decision makers as long as the shortcomings in their execution are addressed. While aware of their existence, decision and policy makers have rarely
used predefined framework assessments to inform their decisions. They have noted that the predefined frameworks have not highlighted previously unknown proliferation issues related to nuclear fuel cycles. Because these issues are not addressed by expansion or further development of existing predefined frameworks, the committee does not support a new or expanding R&D program.
The committee found that nonproliferation policy and decision makers routinely use multidisciplinary teams on a case-by-case basis and their own knowledge to provide analysis across a broad range of topics including proliferation risk. Policy makers have found that proliferation resistance assessments using predefined frameworks address a subset of their issues, and the predefined frameworks have not produced significant results that were not previously understood. Additionally, some policy makers have expressed concern with predefined frameworks that integrate many factors or produce a single-valued result for a complex problem. They are also concerned that they would be bound by the results. Therefore, in general policy makers find little utility in using predefined frameworks or formal risk-based approaches for such applications.
Proliferation resistance assessments for potential future fuel cycles have limited information on technical design features, operational modalities, institutional arrangements and safeguards measures. The cost and time of executing a predefined framework or any other detailed technical assessment to inform R&D decisions is difficult to justify. For R&D decisions, it would be more useful to simplify the analysis to addressing a few key questions, tailored to the level of known detail. As the technology is developed and comes closer to deployment, many more details may be known and the use of a predefined framework may be justified.
The committee found several examples of nonproliferation policy makers and decision makers using a checklist of high-level questions to consistently address proliferation risk-related issues. For example, all Nuclear Cooperation Agreements (NCAs) have nine requirements to be addressed in the Nuclear Proliferation Assessment Statement (NPAS), and the NNSA’s Office of Nuclear Controls uses a set of seven questions that must be answered for each export review that is performed.
RECOMMENDATION 4.1: The committee recommends that fuel cycle R&D decisions include proliferation resistance (rather than proliferation risk) as one factor among others (such as cost and safety) to guide those decisions. Technical assessments are limited by the availability of technical details associated with future nuclear fuel cycles. Therefore, the committee recommends that DOE-NE and NNSA jointly decide upon a set of high-level questions comparing the proliferation resistance of proposed future fuel cycles to the current once-through fuel cycles to determine as early as possible in their development whether the former have significantly different intrinsic proliferation resistance (either for the better or for the worse). Assessments should be revisited at key milestones throughout the technologies’ development and eventual deployment; they should become more detailed as appropriate as new and better information and data emerge.
In determining the set of questions, DOE-NE and NNSA may consider existing metrics or questions. It may not be necessary to develop a new set of high-level questions.
TASK 5: IDENTIFY AND ASSESS OPTIONS FOR EFFECTIVELY COMMUNICATING PROLIFERATION RISK INFORMATION TO GOVERNMENT AND INDUSTRY DECISION MAKERS, AS WELL AS THE PUBLIC AND THE NON-GOVERNMENTAL ORGANIZATION (NGO) COMMUNITY BOTH WITHIN THE UNITED STATES AND INTERNATIONALLY.
The nonproliferation community consists of many segments: U.S. and international policy makers and decision makers, technical analysts, non-governmental organizations, academics, as well as the interested public. Communication among these communities is often impeded by issues such as lack of common vocabulary, differences in technical background, knowledge of social or political issues, lack of access to classified information, and overall distrust.
FINDING 5.1: The terms “proliferation risk” and “proliferation resistance” frequently are used interchangeably and incorrectly when discussing nuclear energy systems. In addition, technical methods for assessing proliferation resistance are often referred to as methods for assessing proliferation risk. This creates confusion, is misleading, and impedes communication.
FINDING 5.2: Predefined framework assessments provide a structured approach that can enhance communication and education as long as their purpose, scope, assumptions, and limitations are clearly stated and understood.
Predefined framework methodologies can facilitate communication about proliferation resistance by providing a structure for organizing and discussing complex data. For example:
• Domestic and international communities: Predefined frameworks provide a common lexicon and vocabulary during international expert discussion, thereby facilitating communications (e.g., Gen IV International Forum Proliferation Resistance and Physical Protection Working Group).
• Policy makers, public, and international partners: Predefined frameworks can help establish a common lexicon and provide a useful structure for communicating how a large number of factors contribute to proliferation resistance, and facilitate communication of policy decisions to NGOs, the interested public, and international partners.
Predefined framework assessments can also be useful for training academics and next-generation policy makers on proliferation-relevant features of the nuclear fuel cycle, the role of international safeguards, and approaches to increasing proliferation resistance.
However, the purpose, scope, and assumptions of framework methodologies must be clear, if results are to be interpreted appropriately. It should be clear that assessments of proliferation resistance or risk are not absolute, evolve with time, and are one factor among many that contribute to nonproliferation decisions.
RECOMMENDATION 5.1: To build trust and increase transparency with domestic and international stakeholders, policy makers and decision makers should refrain from technical jargon in communicating proliferation risk, refer to information available to all parties whenever possible, and always include discussion of the assumptions and limitations inherent in any assessment.