The Manhattan Project Resources

U.S. Department of Energy

HSS Office of Classification

Office of History and Heritage Resources

Report of the Fundamental Classification Policy Review Group

Report of the
Fundamental Classification
Policy Review Group

 

January 15, 1997

 

Dr. Albert Narath, Chair

 

Unclassified Version
(Classified Material Has been Removed)

Issued by the Department of Energy
October 1997


CONTENTS

Part I Part II


Figures

  1. Fundamental Classification Policy Review Organization
  2. Fundamental Classification Policy Review Schedule
  3. Classification as a System
  4. Graded Portion Marking
  5. Example Decision Logic Chart for Applying Guiding Principles to Restricted Data


Appendices

Title Classification
A. Tasking U
B. Stakeholder Input U
C. Topics Requiring Greater Protection SRD (not included)
D. Report of the Weapon Design Working Group SRD (not included)
E. Report of the Weapon Science Working Group SRD (not included)
F. Report of the Weaponization and Weapons Production and Military Use Working Group U
Annex A – Overarching Issues
U
Annex B – Specific Declassification/Downgrading/
Transclassification Recommendations
U
Annex C – Classified Rationales for Some Declassification Recommendations
SRD (not included)
Annex D – Proposed FRD Topic Transclassifications from CG-W-5 "Joint DOE/DoD Nuclear Weapon Classification Policy Guide"
CRD (not included)
G. Report of the Nuclear Materials Production
Working Group
U
H. Report of the Military Reactors Working Group U
I. Report of the Safeguards and Security
Working Group
U (OUO not included)
J. Summary of Proposed Topics for Declassification SRD (not included)
K. Organization and Participants U

Part I

The first chapter of this report describes the origins of the Review, its place in the DOE Openness Initiative, and the general plan for carrying out the tasks assigned. Stakeholder concerns are discussed, and limitations on the utility of classification for shielding information from potential adversaries are outlined.

In contrast to the current principle of balancing four aims—assuring defense and nonproliferation; promoting peaceful applications of atomic energy; disseminating environmental, safety, and health-related information; and promoting technology transfer—the Review recommends, in Chapter 2, principles for governing the DOE classification system that are centered on the proposition that:

Classification must be based on explainable judgments of identifiable risk to national security and no other reason.

Six recommendations for dealing with overarching issues are given in Chapter 3. Implementation of these recommendations, together with the principles from Chapter 2, will provide the foundation for a major cultural shift toward better protection for information truly important to national security while making available to the public information that no longer warrants protection.

The results of the detailed work to determine which information must continue to be protected and which should be made available to the public are discussed in Part II of this report.

Chapter 1 – Introduction

"Much of the strength and efficiency of any government in procuring and securing happiness to the people depends on opinion, on the general opinion of the goodness of that government ..."
  Benjamin Franklin to the
Constitutional Convention, 1787

The DOE Openness Initiative

In December 1993, Secretary of Energy Hazel O'Leary launched a comprehensive initiative to move the Department of Energy (DOE) into a new era of government openness. She outlined a broad package of classification and declassification reforms and disclosed for the first time:

The Openness Initiative is DOE's response to President Clinton's commitment to improve public access to government. The President emphasized the importance of accessibility when he stated, "The more the American people know about their government the better they will be governed." After reviewing DOE's approach, the President called it, "a bold new initiative that will allow an informed group of stakeholders to work with the Department of Energy to solve the problems that face our nation." Key elements of the Openness Initiative include:

The announcement of the Openness Initiative was followed by 28 public meetings aimed at garnering public support and trust. Over 1,000 DOE stakeholders responded. During her June 1994 press conference, Secretary O'Leary stated, "We've been listening and what we've heard loud and clear is that openness in government is very important to our citizens. We've also heard that accessibility is important." The Secretary then outlined broad classification reforms and released information about the nation's nuclear materials production program, its nuclear testing programs, and the U.S. nuclear weapons stockpile.

In addition, the Secretary described new actions aimed at declassification and classification reform and accessibility. These commitments included:

The Review Task

"In proportion as the structure of a government gives force to public opinion, it is essential that public opinion should be enlightened ..."
  George Washington
Farewell Address, 1796

The Under Secretary of Energy, in a February 13, 1995, letter to Dr. Albert Narath (Appendix A), requested that Dr. Narath chair a group to carry out a comprehensive, fundamental review of DOE's classification policy. The stated objective was to determine which information must continue to be protected and which no longer requires protection and should be made available to the public. The review was identified as a major component of Secretary O'Leary's Openness Initiative with an overarching objective of making openness a DOE legacy. Twelve months were allowed for the review so that the views of stakeholders both within and outside the government could be solicited and expertise from the Department of Defense (DoD) and other government agencies brought to bear.

All information within DOE's responsibility was included in the review. Within the Department there are four major categories of protected information. Three are derived from the Atomic Energy Act (AEA) of 1954, as amended; the fourth is defined by Executive Order 12958 of April 17, 1995. The charge to the Review Group applied not only to the individual categories, briefly described below, but also to the manner in which they collectively are used to protect information.

Guidance associated with these four categories of information is contained in more than 50 headquarters classification guides and about 800 local classification guides. General policy for classification is given in CG-C-3A, "Classification Policy Guide for Nuclear Programs," dated June 1993. Because the final report to the Secretary was to make clear, concise recommendations for change, along with supporting rationale, review of all policy driving the various guides, and in many cases the details of how that policy is applied, were integral to the study task.

The Approach

"The mutual confidence on which all else depends can be maintained only by an open mind and a brave reliance on free discussion."
  Justice Learned Hand
October 24, 1952

The nature of the task suggested an organization with enough formality to foster completeness and timeliness yet retaining flexibility to deal with prominent issues arising from research and deliberations. The arrangement shown in Figure 1 (page 49) was selected to meet these needs. A steering group, comprising the chair, deputy chair and working group and coordinator group chairs, was formed to direct and coordinate the various activities.

Working groups, suitable for collegial discussions, were formed with experts in the relevant areas of technology and policy. Because the main job of the groups was to determine the boundary between information which can be published without undue risk and that which must continue to be protected, classified research and deliberations were essential. For this reason, working group members were selected from the population holding appropriate clearances. The groups were arranged to span the full spectrum of DOE classified information with some deliberate subject matter overlap among groups.

A schedule emphasizing stakeholder input was developed by the steering group (see Figure 2 on page 50). Distinguished speakers from inside and outside the government followed Secretary O'Leary in addressing the initial steering group meeting in Washington, D.C. Written comments were solicited through the Federal Register and by direct mailing to stakeholders known to the DOE Office of Classification. Another public meeting was held in Oakland, CA, on July 28, 1995, to gain additional stakeholder comment. Again, a distinguished group of public speakers expressed their concerns and suggestions for improvements. Both public sessions were videotaped so that Review participants could carefully consider this important counsel.

Following the initial meeting, the steering group developed a set of interim guiding principles. The 50 or so headquarters classification guides were divided among the working groups. Study areas were in many cases interdependent and in some cases overlapping.

Research for the Review was greatly facilitated by two previous studies. The Classification Policy Study, requested by the Under Secretary and carried out by DOE's Office of Classification, examined statutes, Executive Orders, DOE Orders, and other records. A number of findings and recommendations, largely dealing with issues requiring modifications to the Atomic Energy Act, were presented.1 A Committee on Declassification of Information for the DOE Environmental Remediation and Related Programs was convened by the National Academy of Sciences at the request of Secretary O'Leary. The July 1995 final report put forth four basic principles and a number of recommendations.2

After a May 2, 1995, plenary session involving all working groups and associated support staff, detailed investigation proceeded under the direction of the working group chairs. Periodic steering group meetings monitored progress, considered emerging issues, and coordinated efforts.

Informal cooperation building was pursued throughout the process as a means for widening the spectrum of issues and concerns under consideration and for informing stakeholders in both the government and the private sector of the intent and progress of the Review. The report summarizing the work conducted and recommendations developed during this review was submitted in draft form for public comment on February 6, 1996.

Stakeholder Concerns

"Every citizen owes the country a vigilant watch and close scrutiny of its public servants and a fair and reasonable estimate of their fidelity and usefulness. Thus is the people's will impressed on the whole framework of our civil policy; and this is the price of our liberty and the inspiration of our faith in the Republic."
  Grover Cleveland
First Inaugural Address
March 4, 1885

Stakeholder comments came from a diverse group of private citizens, federal and non-federal government personnel, public and private interest groups, and experts from various disciplines. In addition to public meetings, questionnaires and surveys were used to solicit inputs on specific classification policy issues and to invite comment on topics of individual or organizational interest or concern. As the number of responses expanded, a database was established that grouped information into six major categories: accessibility; openness concerns; declassification proposals; human experimentation; nonproliferation and testing; and environment, safety, and health (ES&H).

Analysis indicated major concerns in the following areas:

ES&H concerns were voiced by a broad range of stakeholders including private citizens, environmentalists, historians, and health researchers. The stakeholders felt that access to ES&H data would:

Accessibility was a major concern for historians, public interest groups, journalists, and individuals. The majority stated that the review of pertinent documents must be done more promptly and, upon completion of the declassification process, the documents must be released to the public. In addition, the public should be advised of newly released materials. Specific suggestions included:

All human experimentation data were declassified prior to the Review; however, stakeholder interest remained high. Interested parties included veterans organizations, private citizens, medical researchers, and historians. Stakeholders praised the Department's efforts that began in 1993. The major concerns expressed by this group were accessibility to declassified information and the lack of interagency cooperation in response to information requests. Ongoing litigation and the age of the population concerned make this an extremely complex and emotional issue.

Researchers, historians, environmentalists, and special interest groups requested that data relative to the U.S. nuclear weapons stockpile be declassified. Requests included:

Stakeholders indicated that the release of these data allows independent review of U.S. strategic policy. Weapons storage, stockpile maintenance, and the Science-Based Stockpile Stewardship Program were included in this area of concern.

While stakeholders demanded improved access to both DOE facilities and information, they were in general agreement that critical nuclear weapon design information must remain safeguarded. Several expressed the belief that scientific information should be declassified on relatively short schedules. However, they recommended that all design data should be reviewed and detailed technical information that continues to be critical should be protected in the most stringent manner available. "High fences" around truly sensitive subjects, but declassification and dissemination of a large volume of lesser materials, was a popular theme.

A compendium of stakeholder comments is presented in Appendix B.

Protectability of Information

"He who is firmly in authority soon learns to think security, and not progress, the highest level of statecraft."
  James Russell Lowell
Literary Essay, 1890

Not all information important to national security can be protected by classification. In assessing protectability—the ability to shield information from potential adversaries—it is helpful to consider the nature of the information, which from the standpoint of classification can be placed into three categories: subjective, objective, and technical.3

Subjective information is arbitrary and reflective of human decisions and intentions. Consequently, it is unique in that it cannot be generated independently by an adversary. It can, however, be compromised by espionage or unauthorized disclosure. Operational information such as military plans and shipping dates is of this nature. Some subjective information effectively becomes declassified once the activities described are enacted (e.g., launch dates). Thus, establishing a duration of classification at the time this information is generated is frequently practical. Subjective information, when generated and maintained under strict control and secured by proper safeguards, is unconditionally protectable.

Scientific Information—the systematized knowledge of nature and the physical world—falls in the objective category. It is quite different from subjective information in that it cannot be made unconditionally protectable. It is unique but independently discoverable or determinable by skilled investigators. Consideration of the time and resources needed by potential adversaries to unravel a particular fact of nature is quite appropriate when classifying scientific information. Thus, the credibility of the classification process dictates continual awareness of the progress of outside investigators.

While there is little value in trying to protect scientific information published in learned journals, the details of how that information can be used may be protectable. Scientific information obtained in unusual and exotic ways, for example, nuclear explosive testing, while in theory independently discoverable, may in reality be very protectable for long periods of time. Protectability in this example may best be invoked by the nuclear powers adhering to mutually agreed classification norms or standards. By the same token, scientific breakthroughs may be more protectable than information generated through the steady advance along an established line of inquiry.

Technical Information—information concerning a method, process, technique, or device consistent with the laws of nature, employed to create a product or result—falls between the subjective and objective categories. Like scientific information, technical information is eternal; but unlike scientific information, it may not be unique. There may be several ways to use facts of nature to achieve a certain objective. Consequently, while technical information is not unconditionally protectable, it is frequently more protectable than scientific information. Estimates of time and resources to discover certain technical information can be important considerations in the classification decision process, but they can also be misleading. Since the information is not unique, it may elude methodological investigators for a long time regardless of resources expended. On the other hand, a clever adversary may devise a completely different way of achieving the same objective with little resource expenditure.

To be completely protectable, information must be under strict control. That is, the information holder must be able to restrain access to the information. However, information that has been compromised should not necessarily be declassified. Protectability of some value may still be present, albeit against a smaller set of potential adversaries. Likewise, information revealed by unauthorized disclosure or independent discovery may best remain classified if declassification would remove significant uncertainties on the part of even a subset of potential adversaries. Continued classification, for example, if practiced consistently by all sophisticated nations holding the information, may markedly slow the progress of the less sophisticated.

Chapter 2 – Principles for Governing the Classification System

"Moderation in temper is always a virtue, but moderation in principle is always a vice."
  Thomas Paine, 1792

Classification as a System

Classification policy and practice, like nuclear weapons themselves, exist to support national security policy and objectives. In turn, classification policy must be founded on sound principles. Doctrine for developing, modifying, and applying classification guides—the primary instruments used in classifying and declassifying information—flows from both policy and principles.

Figure 3 (page 51) illustrates schematically how the classification system works. The products generated by the system are classification and declassification actions—determinations by individual authorized classifiers as to the proper classification category and level for documents and materials within their responsibilities. These actions are supported by a hierarchy of classification guides, which are the basic tools of the authorized classifier.4 Local guides—issued by DOE Operations Offices or contractors, after approval by the Office of Classification or Operations Offices, dependent upon subject matter—provide detailed direction in selected technical areas. Program guides are developed for work involving two or more field offices or for cooperative efforts with another government agency. Both local and program guides are prepared and revised in conformance with the DOE Classification Policy Guide—a critical link between policy and individual classification and declassification actions. Changes in national policy must generally flow through this hierarchy before practice is revised at the working level.

The system has several feedback mechanisms for adjusting to changing conditions. A Technical Evaluation Panel—the principal scientific advisory group—advises on specific items submitted for declassification and makes recommendations for modifying and interpreting policy. Formal appraisals are conducted to determine whether practices actually conform to DOE policy and to evaluate the effectiveness of classification personnel in implementing the program.5 DOE also conducts various reviews of documents for declassification and downgrading. These mechanisms were devised to keep the system in tune with slowly evolving Cold War policy.

The system is currently keyed to the underlying principle of achieving a balance among four aims:6

  1. assuring defense and nonproliferation by controlling declassification;

  2. promoting peaceful applications of atomic energy by dissemination of scientific and technical information;

  3. promoting dissemination of environmental, safety, and health-related information; and

  4. promoting technology transfer for U.S. commercial interests.

Fifteen general policy statements, including criteria for declassification, are provided for authorized users seeking to apply the underlying principle to specific problems.7

The previously discussed DOE Openness Initiative is only one manifestation of policy changes since the end of the Cold War. Defense focus has shifted toward a world where regional powers may attempt to gain local hegemony through aggression or intimidation. These threats cannot, for the most part, be adequately addressed by the United States or any other single nation state.8 In general, the changes in national security policy provide a greater opportunity to emphasize our commitment to open government.9 However, a critical priority of the U.S. is to stem the proliferation of weapons of mass destruction and their missile delivery systems.10 It is largely within this arena of sometimes competing national priorities—international cooperation, openness in government and nonproliferation—that a new DOE classification policy must be formulated.

The Review sought to identify and evaluate fundamental changes that could bring the DOE classification system more closely into line with national policy. A premium was placed on expressing policy and principles so that both practitioners of classification and the public at large can understand which information must be protected and which can be made available to the public. In contrast to the current principle of balancing four aims, the Review recommends 11 general and another 10 area-specific principles aimed primarily at defining what DOE classification must and must not do.

General Principles

  1. Public trust can best be held by providing complete and accurate information in a timely manner and ensuring that only information requiring protection is classified.11

  2. Classification must be based on explainable judgments of identifiable risk to national security and no other reason.12

  3. Information relating to environmental, safety, and health issues should be classified only when national security requirements clearly outweigh the public's need to know.

  4. Classification must never be used to conceal or delay the discovery of violations of law, inefficiency, or errors; to prevent embarrassment; or to restrain competition.

  5. Classification policy must be unambiguously related to national policy and enunciated in a manner understandable by the public.

  6. Classification guidance must be traceable to classification policy and must provide a clear, unambiguous understanding of what information must be classified.

  7. Classification policy and guidance must be reviewed periodically to ensure harmony with national policy and to identify information that can be declassified.

  8. Classification policy and practices must honor U.S. international commitments contained in treaties and other agreements and not inhibit authorized cooperation with other countries.

  9. Unauthorized disclosure does not necessarily provide a basis for declassification.

  10. Information or technology considered obsolete by U.S. standards may still be classified due to its usefulness to terrorists and proliferators.

  11. Open communication of fundamental science does not preclude classification of implementing technology.

Nuclear Weapons – Specific Principles

  1. DOE classification is primarily focused on stemming the flow of information that could materially advance the objectives of nuclear proliferators, terrorists, or saboteurs; assist in significantly improving a nuclear weapon capability; or expose a significant vulnerability or defect in U.S. weapons.

  2. Classification is a major element in assuring that a U.S. nuclear weapon cannot be used in an unauthorized manner.

  3. Classification may delay, but cannot prevent, acquisition of a first-generation nuclear weapon; it can significantly increase the cost and the time to develop more advanced capabilities.

  4. Information confirming the technical merits of various approaches to nuclear weapon development, detailed nuclear weapon design information, performance data gained from nuclear tests, and information on how test data are used to validate models and databases must be closely guarded.

  5. Basic scientific information, not revealing details of weapon design or fissile material production, should not be classified unless there is a compelling reason to believe that disclosure would significantly assist in gaining or enhancing a nuclear capability.

Nuclear Materials Production – Specific Principle

  1. Information that could materially assist a proliferator in producing special nuclear material shall be classified.

Military Reactor – Specific Principles

  1. DOE classification is primarily focused on protecting information that could materially assist potential adversaries in exploiting vulnerabilities in U.S. military reactor systems or in developing or significantly improving their military reactors.

  2. Information on retired programs is generally obsolete and decisions to retain as classified should only be made on the basis of applicability to current programs.

Safeguards and Security – Specific Principles

  1. Sound risk management approaches should be applied in the decision process to classify safeguards and security information.

  2. Safeguards and security information that could result in an adversary obtaining a nuclear weapon or nuclear material, or in nuclear sabotage, or in damage to the health and safety of government employees or the public must be protected.

Chapter 3 – Overarching Issues

"When our Founders boldly declared America's independence to the world . . . they knew that America to endure would have to change; not change for change sake, but change to preserve America's ideals: life, liberty, and the pursuit of happiness."
  Bill Clinton
Inaugural Address, January 20, 1993

Modifying the Atomic Energy Act

The stringent provisions for classification prescribed by the Atomic Energy Act of 1954 (AEA) were appropriate for an era when it was very uncertain where nuclear research might lead. Four decades of intensive technical work, however, provide the understanding needed to simplify those provisions without compromising national security. Changing the AEA is fundamental to changing DOE classification policy.

Restricted Data is defined in the AEA in a very conservative manner. All data meeting the general definition are considered to be "Born Classified." Unlike National Security Information, no decision by an official is needed to render information falling within the AEA definition classified. The very nature of this approach leads to an environment fostering overclassification.

Today, an extensive set of classification guides (briefly described in Chapter 2) provides workers with official direction on which material is to be classified. Consequently, it is now time to make the provisions for Restricted Data more consistent with those governing classified material throughout the government.

Recommendation 1. Amend the AEA definition of Restricted Data by changing the wording of Chapter 2, Section 11 (y) from "all data concerning ..." to "all data designated by the Secretary of Energy concerning ..."

Another shortcoming in the AEA has arisen through examination of the legality of reclassifying previously declassified material. Implementation of Recommendation 1 will make it less likely that a reclassification action is needed, but the Secretary of Energy should have this authority in order to rectify previous decisions that are inappropriate because of errors or changing national security conditions.

Recommendation 2. Amend the AEA to provide the Secretary of Energy authority to classify specific topics in previously declassified Restricted Data categories based on technological developments or changing national security conditions.

The AEA also provides for removing from the Restricted Data category information primarily relating to military utilization of nuclear weapons provided that the information is safeguarded adequately as "defense information," and that it is shared with other nations only pursuant to agreements for cooperation as defined in the AEA.

Information in this category is known as Formerly Restricted Data—a term that has caused much confusion and unnecessary work over the years. Little difference exists between National Security Information and Formerly Restricted Data except for the cumbersome requirement for joint DoD-DOE determinations on declassification and the process for sharing the information with other nations—a process largely redundant with other mechanisms for achieving similar objectives.

The DoD and DOE could benefit by jointly reviewing and separating topics classified Formerly Restricted Data for transclassification to National Security Information or Restricted Data, as appropriate, with the objective of eliminating Formerly Restricted Data from current usage. One result should be a reduced number of topics classified by joint DoD-DOE guidance, thereby removing DOE from classification decisions on military utilization more properly made by DoD.13 Legislation would, however, be required to apply NSI processes unencumbered by stipulated AEA (Sec. 142.d) restrictions on sharing military utilization information with other nations.

Recommendation 3. Amend the AEA to allow designation as National Security Information information now categorized as Formerly Restricted Data.

Implementation of the three recommendations discussed above is essential if DOE is to significantly improve classification policy. Reining in "Born Classified" is needed to deal with overclassification. The authority to reclassify is essential to address timidity in the declassification process. Eliminating Formerly Restricted Data simplifies processes and clarifies classification lines of responsibility. All three have been recommended by the Classification Policy Study and the National Academy of Sciences review.

Improving DOE Practices

As a matter of practice, DOE has generally classified entire documents at the highest level of any material contained therein without further identification of the relative sensitivity of the various parts. By contrast, most government agencies use segregation and portion marking of classified material so that the reader can more precisely identify the sensitivity of the various parts. Segregation and portion marking based on a strong system of classification guides offers a significant measure for reducing overclassification today, for more easily declassifying material in the future, and for more clearly distinguishing between classified and unclassified information.

What constitutes an appropriate portion may vary greatly from document to document. Figure 4 (page 52) illustrates a decision logic that can be applied on a case by case basis. Choices on appropriate portions for segregation or marking can best be made by well-informed technical staff with review and assistance by authorized derivative classifiers. Caution must be exercised to ensure that the entire assemblage of all portions in a document that are marked at a given level and category is itself classified at that level and category as a whole, and that the overall marking on the document reflects the highest classification level contained therein and the most restrictive category of information.

Recommendation 4. The DOE should institute a graded system (as illustrated in Figure 4 on page 52) for segregation and portion marking of all new Restricted Data documents by authors and authorized derivative classifiers. (Implementation should begin as classification guides are revised.)

There are four main categories of Unclassified Controlled Nuclear Information (UCNI): security, nuclear materials production, nuclear power production, and nuclear weapons. Implementation of UCNI is confusing; criteria for use are difficult to distinguish from those for classification. For example, the phrase in the AEA, Sec. 148, "... significant adverse effect on health and safety of the public or the common defense and security," is similar to the definition for "Secret" classified information in Executive Order 12958: "... reasonably could be expected to cause serious damage to national security."

The most compelling argument for retaining UCNI is to provide some protection for sensitive material that needs to be shared with those not holding security clearances (e.g., law enforcement officials). Release of UCNI to the general public is prohibited. It is thought that this protection lowers the risk of terrorist attacks on sensitive facilities, operations, and associated personnel.

Our analysis indicates that an 80% reduction in UCNI-marked documents can be achieved if application is limited to safeguards and security needs.

Recommendation 5. The use of UCNI should be limited to nuclear safeguards and physical security information that is clearly unclassified.

High Fences Around the Most Sensitive Information

Some stakeholders, while advocating major classification reductions in less sensitive areas, also urged that strict, and perhaps higher, levels of security be maintained around the more sensitive material. The main effort of the Review was focused on the former objective--identifying information that no longer requires protection. However, our investigations led to the conclusion that the latter admonition is valid, and in fact should be treated as an imperative.

As a result of downsizing the DOE nuclear weapons complex, the nuclear stockpile, and the associated operational forces, far fewer people need access to Restricted Data. Implementation of the recommendations of this Review, together with previous actions, will reduce the amount of information that must remain classified. Fewer people accessing less material makes protection enhancement a viable course of action.

Restricted Data not recommended for declassification, in general, warrant protective measures at least as stringent as those used today. Moreover, there exists a body of information within the Sigma 1 and 2 (theory of operation or design of thermonuclear and implosion-type warheads) and Sigma 14 and 15 (nuclear weapons unauthorized use and use control design and vulnerability assessment) categories that warrants special protection, namely:

Appendix C provides a list of topics judged to require this additional care. (It should be noted that today DoD personnel having clearances based on a National Agency Check can be authorized access to most of this material, and DOE personnel having L clearances can be authorized access to topics denoted SFRD.)

Two main options have been discussed of late for achieving extra protection for this more sensitive material:

  1. Creating a new subcategory (handling caveat) of RD (and perhaps FRD) designated "CRIT," with classification marking and protection requirements published in the National Industrial Security Program Operating Manual; or

  2. Reclassifying this sensitive information to Top Secret so that existing procedures can be used to achieve enhanced protection.

Option 1 requires cooperation of other government agencies, particularly DoD, to ensure the CRIT is recognized and honored with the appropriate clearances and need to know measures. While such processes are in place for Option 2 (Top Secret), implementation of the associated procedures would be burdensome, particularly for computer operations.

Both options raise serious questions as to how to properly deal with previously generated material. A high level of protection must be maintained until review and proper marking can be accomplished. There is also a concern that SRD material not marked for increased protection should continue to be protected at a level not less than that used today. Specifically, Q clearances and strict (or even stricter) need to know control should continue to be required.

Most importantly, regardless of the method of protection chosen for the most sensitive material, discipline must be maintained across all levels and categories of classified material. Procedures must be clear, and all personnel involved must understand the importance of adhering to those procedures.

Recommendation 6. More stringent measures should be implemented for protection of the 137 topics identified in Appendix C as most sensitive.

One major advantage gained by collective implementation of the previous recommendations will be a much clearer understanding of which information must remain Restricted Data and within that category which topics are most sensitive. This will allow greater protection because stringent measures can be applied more effectively around sharply defined areas of importance.

Relying on Protection of Unclassified Information

The provisions for protecting classified material at the Confidential level (disclosure could cause damage to national security) are far stronger than the various measures for controlling special categories of unclassified information. Unclassified Controlled Nuclear Information (UCNI) and Official Use Only information, while curbing information requests under the Freedom of Information Act, do not have protective measures consistent with preventing disclosure to determined adversaries. For example, non-U.S. citizens who are employees of a government subcontractor can access UCNI for purposes of bidding on a government contract.14 Consequently, if information needs to be withheld for reasons of national security, it should remain classified. Declassifiers should not count on protection against disclosure under any of the controls for unclassified information today, and, as worldwide communications systems grow, instant disclosure becomes a real possibility. Classification, as previously noted, can only delay an adversary from obtaining or developing technical information; no delay can be expected when the information is unclassified.


Part II

The results of the detailed examination of current classification guides by the seven working groups are briefly described in Chapters 4-8. In-depth reports of this work are presented in Appendices D through I. Recommended declassification actions on 62 topics are summarized in Appendix J. Members of the FCPR are listed in Appendix K.

Fully mindful of the fact that there is no such thing as an obsolete nuclear weapon design, the Review has placed particular emphasis on protecting information needed to gain or advance a nuclear weapon capability. Sorting information that no longer warrants protection from the larger body of currently classified material required informed debate and judgment. In most cases detailed classified discussion was required, and in many cases the rationale for recommending for or against disclosure could only be properly framed in arguments that are themselves classified. Consequently, much of the material in the reports of the working groups is classified, and the summaries given in the following chapters may appear terse or incomplete to the reader.

Chapter 4 - Nuclear Weapon Science, Technology, Design, Weaponization, and Testing

"Yet in holding scientific research and discovery in respect, as we should, we must also be alert to the equal and opposite danger that public policy could itself become the captive of a scientific-technological elite."
  Dwight Eisenhower
Farewell Address, 1961

One overwhelming conclusion from this Review was that the important information needed to gain or advance a nuclear weapon capability should continue to be carefully protected. The potential threat to the security of the U.S. and its allies of even a primitive single-stage weapon is indeed grave.

Figure 5 (page 53) schematically illustrates how the principles given in Chapter 2 can be applied to the regime of Restricted Data. In making decisions as to the usefulness of information to potential adversaries, particularly proliferants and terrorists, it is desirable to invoke seasoned professional judgment at each step and to err on the side of protection rather than disclosure.

Current policy is to classify information that would aid a proliferant in developing a nuclear weapon, aid a terrorist in constructing a weapon, aid a nuclear power in improving its nuclear weapons, or reveal vulnerabilities in U.S. weapons. Thus, while general concepts and principles of operation of gun-assembled, implosion-assembled, boosted, and staged (thermonuclear) weapons are unclassified, the details remain classified. The result is that drawings of devices or device parts, details of the high explosive system and its detonation, neutron initiation requirements, pit response to the high explosive, boost requirements (quantity of fuel, pressure, temperature, ratio of deuterium to tritium, fusion yield, etc.) and resulting increase in yield are classified. The yield of the primary and its output, the control (management) of the radiation flow to the secondary, the response of the secondary (motion, temperature, pressure, etc.), the configuration of secondary parts, and the response of materials to radiation from the primary are classified. The mass, shape, and, for some materials, the fact of use in weapons is classified. The current unclassified list of materials in the secondary is hydrogen (any isotope), lithium (any enrichment in lithium-6), and uranium (depleted, natural or highly enriched).

Detailed judgments for updating weapon design classification guidance are given in Appendix D. Little change is recommended except in the area of the association of materials with weapons in general or, in some cases, with specific weapons. Declassifying these associations, even if estimates of the mass used in specific weapons can be inferred from plant feed streams or averages, is considered of small risk as long as the details of how the materials are used are protected. (Declassifying the association of a material with a weapon plant or, in some cases, a specific weapon type will allow declassification of many documents that are classified only because they infer or identify the material and its flow into and out of the plant.)

The spontaneous decay of radioactive materials present in nuclear weapons results in the production of neutrons and gamma rays inside the device. Some of this radiation escapes the device and is called intrinsic radiation (INRAD). Personnel who work or sleep in close proximity to the device, usually sailors, are exposed to this radiation. The present classification guidance allows for the specification of a total dose from INRAD at a single radius outside the device measured from the device center. If that single number is an upper bound at that distance, then the maximum health risk is known and can be evaluated. This has been sufficient so far. To characterize the hazard to personnel more completely, it may become desirable to specify the total dose as a function of polar angle or even on the device surface. This information is currently classified because the orientation of the warhead may be deduced from this data. If the warhead orientation becomes unclassified for a given device, then, and only then, we recommend that the total dose as a function of polar angle could be declassified. The polar axis is taken to be the long axis of the warhead with the origin being on the axis somewhere in the center of the reentry body or bomb.

Classification of scientific information underpinning nuclear weapons design activities must be viewed in a somewhat different fashion. Limited resources have become and will likely remain a significant constraint in managing the acquisition of necessary scientific knowledge. However, the past 40 years have seen a large and sustained growth in areas of general science closely related to nuclear weapons technology—astrophysics, condensed matter, high temperature experiments, and computer design and applications. These resources can be leveraged by encouraging scientific exchange between U.S. researchers and the worldwide community.

With no nuclear testing, the safety and reliability assurance of the stockpile will rest on the ability to attract and retain highly skilled scientists and engineers. Their willingness to center their careers in the nuclear weapons field may be enhanced to the extent that their scientific accomplishments can be recognized and rewarded by their peers in the open and unclassified arena.

At the same time, science and technology information relevant to weapon design and unique to the cleared community—for example, data gained from nuclear tests and information on how such data are used to validate models and databases—must be closely guarded. Classification of scientific information becomes an act of informed balance among these imperatives.

In developing definitions of information that must remain classified, nuclear weapon science activities were considered in three topical areas—materials properties, laboratory experiments, and computer codes. The details of this process and the results are presented in Appendix E; a general summary of the findings and recommendations is given in the following paragraphs.

The relations between the thermodynamic variables of a material—density, temperature, pressure, energy, and entropy—are referred to as equations of state. Understanding nuclear weapon performance is dependent on good equation of state information at very high temperatures and pressures.

Because of the importance of uranium and plutonium to weapon design, the equations of state for the actinides (atomic number greater than 89) should remain classified. All currently classified equation of state information used in weapons design calculations should remain classified because it may embody empirical information gained by comparisons with classified experiments. Otherwise, equation of state information for elements whose atomic number is less than or equal to 89 can be treated as unclassified.

Opacity or optical thickness is a measure of the ability of a given path length of material to attenuate radiation of a given wave length. Capabilities for calculating and measuring opacity are becoming widespread and have application in international scientific areas such as Inertial Confinement Fusion. In nuclear weapons design and evaluation, the opacity of the various materials is needed to calculate the passage of X rays and gamma rays from the primary to the secondary.

Current policy is to classify calculated or measured opacities based on the atomic number and temperature of the material. It is recommended that all currently classified opacity information used in weapons design calculations remain classified because it may embody empirical information gained by comparisons with classified experiments. Otherwise, opacity information for elements with an atomic number greater than 89 can be classified for temperatures greater than those achievable in laboratory experiments. Information for elements whose atomic number is less than or equal to 89 can be treated as unclassified. Theoretical methods will be unclassified except for weapon-specific techniques judged to have been normalized or validated by nuclear test data.

In general, for either equation of state or opacity information, any analysis which uses material properties from an unclassified parameter space—in atomic number, temperature, or pressure—to determine material properties in a classified parameter space is classified.

Laboratory experiments are conducted to gain an understanding of the basic science behind the operation of nuclear weapons. They differ from experiments used to design nuclear weapons in that they usually endeavor to isolate aspects of the underlying science from the integrated operation of a weapon. In the absence of nuclear testing, a new class of experiments will be needed.

Classification of weapon-related experiments, the analysis of the experiments, and their results, should be based on whether they reveal nuclear weapon design information, classified material property data, or information about specific classification models, algorithms, or computer codes.

Determining whether an experiment reveals weapon design information should be based on how closely an experimental configuration "resembles" a nuclear weapon. Weapon-related experimental configurations will differ from weapon components because they are scaled or modified in one of several ways, such as scaling physical dimensions, changing overall geometry, making material substitutions, or changing the driving energy source. Experiments on configurations that differ significantly in physical scale, geometry, material, or energy source from weapons or weapons components, and whose relative dimensions do not give weapon design information, should be unclassified.

Codes are computer programs used to solve the equations describing physical systems. Current classification policy for codes was written when the nuclear weapon design community was unique in access to massive computers and in the ability to develop sophisticated simulations. Today, comparable capabilities exist in many research centers, and collaboration with uncleared researchers can improve and verify methods employed in codes useful for weapon design. There remain, however, certain combinations of methods and special techniques that require protection.

Currently classified codes will remain classified. New codes are classified if they: contain weapon design information, are normalized to nuclear test data or other classified experiments, contain classified material properties, or use (or exhibit references to) weapon-specific special methods or combinations.

Weaponization is the term used to describe those features required to make a safe weapon that will operate efficiently in the intended environment. These features include the nuclear assembly, warhead electrical and boosting systems, initiation and use control components, features for hardening to countermeasures, and support structures. Weaponization also includes the nonnuclear testing and analysis needed to develop these features. The results from consideration of the classification of these topics are presented in Appendix F.

The bulk of weaponization topics are classified Formerly Restricted Data. Under Recommendation 3 (Chapter 3), information relating primarily to military utilization would be transclassified to National Security Information under DoD control. A risk management process is needed, however, to ensure that the impact of DoD classification actions on DOE are fully considered. A process like that given in DoD Directive 3150.1 for DOE acceptance of the DoD-formulated Military Characteristics may be suitable for this purpose.15 Periodic review of DoD classification would be needed to ensure harmony with national security policy.

Under present classification guidance, it is sometimes necessary to classify the presence and quantities of hazardous materials in nuclear weapons. The Review has concluded that revealing the presence of hazardous materials in specific weapons is of little value to potential adversaries. In most cases, declassifying the quantity of hazardous material used in a weapon type involves little or no risk. In those cases where it is deemed important to protect the exact quantity of hazardous material used in a weapon (e.g., plutonium), a "not to exceed value" can be specified to facilitate estimation of environmental impact under various hypothetical scenarios.

Historically, all nuclear weapon quantitative reliability requirements and assessments have been classified to protect operational capabilities. Nuclear weapons are designed to be highly reliable, thus reliability requirements can be considered for declassification jointly by the DoD and DOE on a case-by-case basis. Significant differences between assessed reliability values and design requirements are usually addressed promptly by remedial action; but to ensure that information identifying weaknesses potentially exploitable by an adversary is not divulged, it is recommended that current assessments of stockpiled weapon reliability remain classified.

By this same token, requirements for hardening nuclear weapons to the effects (X rays, neutrons, blast, etc.) from nearby nuclear detonations have been formulated in a very conservative manner. While it is not possible to survive a close nuclear burst, straightforward measures have been implemented to deny any defense the ability to destroy more than one U.S. warhead with one nuclear-tipped interceptor. Hardness requirements should be considered for declassification on a case-by-case basis, but assessed hardness levels should remain classified.

The fact that special features are incorporated in U.S. nuclear weapons to prevent unauthorized use is unclassified. Most detailed information on these use control features is classified to deny potential adversaries even hints of how the various layers of protection are structured. The current policy is considered appropriate. Moreover, new use control concepts should be afforded protection consistent with actual weaponization.

In the area of nuclear testing, information that would reveal nuclear weapon design should remain classified. The association of nuclear test events with weapon programs should, in some cases, be designated National Security Information if DoD chooses to classify certain weapon-specific information relating to military utilization.

The DOE Science-Based Stockpile Stewardship (SBSS) work will rely heavily on new multipurpose facilities. Concerns have been raised about the possibility that unclassified work at these facilities might be counter to U.S. nonproliferation objectives. It is recommended that DOE develop an SBSS facilities classification guide to serve as a basis for local security plans.

Chapter 5 – Nuclear Materials Production

"True opinions can prevail only if the facts to which they refer are known; if they are not known, false ideas are just as effective as true ones, if not a little more effective."
  Walter Lippmann, 1920

Acquisition of fissionable material is a necessary step in the construction of a nuclear weapon. To the extent practical, barriers should be used to increase both the effort and the signature associated with fissionable material acquisition by proliferators and terrorists. Classification of selected DOE information is one of the important barriers that can be applied.

On the other hand, much of the technology associated with the production of fissionable material and light isotopes used in nuclear weapons is now widely available. Classification, as previously noted, can only delay the dissemination of information. In the case of nuclear materials production, only a few items still warrant protection. These matters are addressed in detail in Appendix G and briefly summarized below.

Uranium enrichment core technologies—key features that enable practical or efficient use of a process—should continue to be protected as currently defined. Unclassified definitions of core technologies should be developed and provided to the public so that the rationale for these selective classification actions is understood.

Virtually all information relating to uranium processing has been declassified, but much is currently identified as UCNI. Implementation of Recommendation 2 in Chapter 3 will allow reclassification of truly sensitive processes as Restricted Data. All other information on processing should be unclassified. The fact that intermediate assay uranium is used in weapons is unclassified; the general location where this material is used in an unspecified weapon should be declassified in order to facilitate weapons dismantlement.

During the 1950s, the Atomic Energy Commission declassified virtually all information relating to the extraction of plutonium from irradiated reactor fuels. As concerns about nuclear proliferation have increased in more recent times, DOE has been identifying incremental developments in plutonium extraction as UCNI. For the most part, designating these processes as UCNI does not serve nonproliferation objectives but does, in some cases, prevent dissemination of information on clean-up technologies.

Technologies employed in separating light isotopes are generally unclassified and many are practiced commercially; however, certain specific technologies relating to practicality and efficiency warrant protection. Production of tritium in Savannah River reactors involves processes that can be declassified with little risk. However, technology related to gas retention in high-temperature targets that could be irradiated in commercial reactors should remain classified. Technology for accelerator production of tritium, including the design details of the integrated plant, should not be classified. Details of target technology that would enable significant special nuclear materials production utilizing an accelerator otherwise engaged in legitimate activities should be protected as Restricted Data.

Virtually all tritium processing systems and technology-related information at production sites is classified. However, special classification guides have been developed for the Fusion Program that essentially declassify many areas of tritium processing. This major inconsistency needs to be resolved. Except for extraction systems, it is recommended that all tritium processing information and associated technologies and their applications be unclassified unless the information reveals sensitive physics details of tritium usage in nuclear weapons. The details of tritium extraction systems should be protected if they reveal information that would materially aid in developing a capability to produce tritium in commercial reactors.

The production of nuclear weapons and special nuclear materials involves the use of some highly engineered and often unique nuclear and nonnuclear materials. Classification of the associations of some of these materials with specific sites has greatly complicated waste disposal and reporting activities. Literal compliance with current classification requirements is difficult. A systematic examination concluded that current policy should be revised to specify that the simple association of a material with the nuclear weapons program or with specific sites should not be classified.

Much of the current classification policy on nuclear materials disposition is based on concerns related to revealing details of weapon design. This has led to protection of a large volume of information that might be used to estimate parameters such as average mass of components, material composition, and fabrication technologies. The exact or even approximate configuration and design specifications for specific nuclear weapon special nuclear material parts should remain classified, as should disposition data from which this information could be derived. However, estimates inferred from plant averages are not generally of concern. Declassification of disposition processes should be accomplished on a case-by-case basis.

Chapter 6 – Nuclear Weapon, Special Nuclear Material, and Tritium Inventories

"Two basic requirements necessarily guide U.S. planning for strategic nuclear forces: the need to provide an effective deterrent while remaining within START I/II limits, and the need to allow for additional forces to be reconstituted in the event of a reversal of current positive trends."
  William J. Perry, Secretary of Defense
Annual Report to the President
  and the Congress, 1995

Historically, U.S. inventories of nuclear weapons, special nuclear material (plutonium and enriched uranium), and tritium have been classified at the Secret level or higher. In the early days of the Atomic Energy Commission, stockpile numbers were so sensitive that they were only reluctantly committed to Top Secret documents. The scarcity of weapons and fissile material was considered a very grave national defense matter.16 This sensitivity abated greatly as the stockpile grew to tens of thousands, but the actual numbers continued to be protected.

The intense measures to protect quantities of weapons and special nuclear materials resulted in classification of most information concerning production, and later, dismantlement, from which estimates of the stockpile might be made. Production levels and inventories of weapon-grade plutonium, deuterium, tritium, enriched uranium, and lithium were classified as Secret Restricted Data. Financial information (e.g., unit cost of weapons, nuclear components, and most nuclear materials), production plans and capabilities, reactor power levels and discharge histories, and feed stream quantities from which production rates might be derived were also classified. The net result was a vast body of information that could not be disclosed to the public. This restrained public discussion of environment, safety, and health issues and public debate on U.S. nuclear weapon policy in general and on specific matters, such as new production capabilities, in particular.

The principal reason for classifying weapon production rates and stockpile quantities is to conceal military operational strength and capabilities that are determined, in part, by the size and makeup of the stockpile. "Recent international upheavals have not changed the calculation that nuclear weapons remain an essential part of American military power ... Thus, the United States will continue to threaten retaliation, including nuclear retaliation, and to deter aggression against the United States, U.S. forces, and U.S. allies. The notion, however, that nations are motivated by U.S. nuclear forces in making decisions about acquiring nuclear weapons themselves is not valid. Potential proliferators are more likely to be driven by concerns about neighbors' capabilities or the desire for regional hegemony, than by decisions America makes about its nuclear arsenal."17

Consistency between classification policy and U.S. defense policy, as described above, is best achieved by the Department of Defense setting the classification level for stockpiled weapons. Implementation of Recommendation 3 in Chapter 3 would allow designating classified stockpile numbers as National Security Information (NSI), thus avoiding the cumbersome joint DoD-DOE process for Formerly Restricted Data. Production history, current activity, and shipping and receiving information between the DoD and DOE will be classified as necessary to protect the various quantities designated as NSI.

A weapon type, once retirement is authorized and dismantlement completed, is not a factor in military capabilities, and all information concerning production and dismantlement rates and schedules for this weapon type can be declassified.

The U.S. surplus of special nuclear material is so large that classification of these values is no longer warranted. As previously noted in Chapters 4 and 5, estimates made of special nuclear material in specific weapons based on plant averages and feed streams are of little consequence.

Production quantities, cumulative quantities at Savannah River, and the total DOE inventory of tritium have traditionally been (and are still) classified as Secret Restricted Data. Many aspects of tritium production, handling, and use have also been classified. Originally, these very stringent measures were put into place to protect the concept of boosting—the enhancement of a fission reaction by thermonuclear neutrons derived from the fusion of tritium and deuterium.

Classification continued at the same level after the concept of boosting was declassified because of concerns that the quantity of tritium in specific weapons could be deduced from production rates and inventory. The average tritium mass per weapon, it was thought, might help a proliferant nation reduce the number of tests needed to develop a boosted nuclear weapon.

Tritium, like special nuclear material, exists in the U.S. inventory in quantities far in excess of today's needs. However, unlike special nuclear material, tritium decays fairly rapidly—one half of the inventory will be gone in about 12 years. The U.S. has not produced tritium since 1988 and, unless new production is started, eventually there will not be enough to meet stockpile needs. The DOE is currently pursuing two different technologies for a new tritium production facility. If efforts to restart tritium production fail, a vulnerability in the U.S. nuclear deterrent posture could develop, albeit slowly. In view of present concerns and uncertainties, the steering group is not comfortable recommending declassification of the tritium inventory at this time.

Chapter 7 – Military Reactors

"Publicity is a great purifier because it sets into action the forces of public opinion, and in this country, public opinion controls the course of the nation."
  Justice Charles Evans Hughes, 1908

The United States has constructed 51 military reactors in addition to those built for the Naval Nuclear Propulsion Program. Used primarily for testing and research, all but four nonnaval military reactors have been retired. Many military reactor technologies were declassified prior to this Review. However, certain aspects of a limited number of military reactor programs have remained classified to date because of their relationship to military operations and missions.

Classification is considered appropriate only to the extent that disclosure would enable or assist foreign military capabilities; it should not infringe on the use of nuclear reactor technology for civilian applications unless there is a clear national security concern. Information that would aid in identifying or exploiting vulnerabilities in planned or existing U.S. military space systems must remain classified. Nuclear reactor technology that would assist foreign powers in developing or improving advanced military systems must also remain classified. Information that would reveal aspects of classified activities (e.g., nuclear weapons design, military operations, mission requirements) must be classified. When unclassified technology is used for a classified military application, the application itself, not the power source, should carry the classification.

The following paragraphs outline conclusions that result from applying these classification principles to the Army Nuclear Power Program and to space-related reactor systems. Details are presented in Appendix H.

The Army Nuclear Power Program was developed to field mobile and stationary small-to-medium electrical generating plants for use in remote areas. There are no Army nuclear electrical generating plants in use or under development. Most Army Nuclear Power Program information has already been declassified. Some information was classified because declassification might reveal details of related classified programs such as naval nuclear propulsion. The remaining classified information concerned military missions and operations. It was concluded that declassification of Army Nuclear Power Program information is not likely to reveal significant information about naval nuclear propulsion or current military missions or operations. It is recommended that all information associated with the Army program be declassified.

Classification in directed nuclear energy systems (DNES), isotopic heat sources, thermoelectric conversion, space reactor power systems, and related programs was based on protecting technologies that were difficult or expensive to develop and that could be used by other countries to advance their military programs significantly. Classification also protected U.S. military space-based applications and missions. The classification guides in these areas are jointly owned, and DOE must coordinate final classification decisions with DoD and NASA, as appropriate.

The basic physics of controlled (non-explosive) nuclear reactions to generate a laser beam is unclassified. DNES research using power levels and configurations unsuitable for military application or development is unclassified. Information that will not substantially assist others in DNES development, contribute to feasibility assessment of DNES development, or reveal the direction of DNES research is unclassified, as is work not directed toward, nor suitable for, military development.

Detailed design information on the controlled fission portion of a system used to pump a laser, nuclear design information revealing or describing a significant technological advance, and isotopic fuel composition should remain protected as Restricted Data. Design of the laser portion of a system, including quantitative beam characteristics and research directed toward military applications, should remain classified as National Security Information. A stronger differentiation is needed between unclassified research and items having significant military potential. Research in configurations and power levels unsuitable for military applications should be unclassified unless classified information would be disclosed.

Space reactor power technology predating August 1973 is unclassified, as is information publicly released between August 1973 and February 1987. Information concerning the reactor power supply, specifically MultiMegaWatt or particle bed reactor designs, is classified as Restricted Data; some information on military applications is classified as National Security Information. Most theoretical and basic reactor system information has been declassified; some applied technology, while unclassified, is subject to export control.

The potential usefulness for military missions justifies continuing classification protection for some aspects of space reactor power systems. The basic classification policy is sound, but as programs advance, guidance should be revised to provide more specific determinations of key information—those protected aspects, features, or attributes significant to achieving military advantage. Detailed information on fabrication and testing of uranium nitride fuels and particle bed reactors should remain classified. Information concerning integration of thermoelectric materials with nuclear reactor fuels should be examined in conjunction with space reactor power systems and considered for continued classification, dependent upon program advances.

Space reactor impact points may be classified, if needed, to protect nuclear material or sensitive components until recovery is accomplished. Decisions to classify orbital parameters should be made by the office having primary responsibility.

Thermoelectric conversion and isotopic heat source technologies are unclassified. Some applications of these technologies are still useful for military applications and operations, but the technologies are already unclassified and in use for commercial purposes.

Declassification of existing information could be more reasonably considered if options for classifying new developments are preserved. Implementation of Recommendations 1 and 2 (Chapter 3) would greatly facilitate this activity.

The Naval Nuclear Propulsion Program is a joint DOE-Department of Navy organization tasked with the responsibility to design, build, operate, maintain, and manage the nuclear-powered warships and facilities supporting the U.S. nuclear-powered fleet. Classification within the program is governed by a single guide, CG-RN-1, Rev. 2, DOE-DoD Classification Guide for the Naval Nuclear Propulsion Program, which contains about 600 topics.

At the time the Fundamental Classification Review began, the Naval Nuclear Propulsion Program was already engaged in a detailed review of their classification policy. Therefore, the Military Reactors Working Group did not review CG-RN-1. However, the working group did establish a dialogue with Naval Reactors personnel and exchanged draft reports with them.

The Knolls and Bettis Atomic Power Laboratories conducted the first portion of the Naval Nuclear Propulsion Review. This work consisted of an examination of each topic contained in the classification guide. Working groups at the two laboratories compared their separately formulated assessments against current guidance and results were exchanged and reevaluated in a series of conferences until consensus was reached. A similar review was independently conducted at Naval Nuclear Propulsion Headquarters. Proposed changes were then compared with those from the laboratories and the merits debated by the division directors until a consensus was reached on each item. The final determination was made by the director.

Twenty-one items have been recommended for declassification or downgrading, including information on ship design, materials and metallurgy, chemistry, power plant design, and reactor servicing. Guidance has been reemphasized to separate information concerning public health, safety, and environment from protected information to facilitate release to the public.

The Military Reactors Working Group has reviewed the draft Naval Nuclear Propulsion Program classification review report and judges that the results are consistent with the DOE Fundamental Classification Policy Review.

Chapter 8 – Safeguards and Security

"But a democracy can with only great difficulty regulate the details of an important undertaking It cannot combine its measures with secrecy or await their consequences with patience and they are precisely the quantities by which a nation, like an individual, attains a dominant position."
  Alexis de Tocqueville
Democracy in America, 1835

"Safeguards" refers to physical protection and material control measures to deter, prevent, detect, and respond to sabotage and unauthorized possession or use of nuclear materials. Security refers to policies, activities, programs, and systems for the protection of classified or sensitive information and material and DOE facilities, property, and equipment. A large portion of the classified information associated with safeguards and security falls under Executive Order 12958, National Security Information (NSI). Where Restricted Data (RD) is mixed with NSI, documents are generally marked RD without further elaboration. Unclassified Controlled Nuclear Information (UCNI) is also used widely.

Classification guidance was found to be compliance-based with little or no flexibility to accommodate the wide variation that exists throughout the DOE complex due to different missions and activities. A performance-based approach allowing site-specific guidance based on risk assessment would provide more balanced protection. Current DOE policy allows for a range of classification (e.g., Unclassified-Secret) in certain areas and requires stringent compliance in others. Site-specific classification levels could be developed locally using risk assessment to achieve balanced protection in accordance with revised DOE implementing instructions.18

In the area of safeguards and security, a myriad of measures—classification being only one—can be applied to reduce the risk to national security. Tradeoffs exist between physical security elements and security information protection. A fragile element may require a higher level of classification, whereas a more robust feature may require no classification at all. Risk assessment can be used to make performance-based cost effectiveness determinations on site-specific classification decisions—an approach superior to uniform, complex-wide, compliance-based guidance.

A logic chart similar in nature to the one given in Figure 5 on page 53 can be used in the risk assessment process. (A Safeguards and Security version of the logic chart is presented in Appendix I, along with the detailed findings and recommendations of the working group.) Commercial software (e.g., Expert Choice™) is widely available to help with the decisions that risk assessment demands. It is recommended that a classification decision logic chart be placed in Safeguards and Security classification guides for use as the first step in the site-specific classification determination process.

A graded portion marking methodology can also be applied to achieve the flexibility DOE needs with National Security Information. It is recommended that an agreement be negotiated with the Information Security Oversight Office19 to allow a graded approach to be applied to NSI, using the same methodology recommended for Restricted Data (Recommendation 4, Chapter 3).

From a Safeguards and Security perspective, UCNI provides valuable protection in certain areas but implementation is complex, confusing, and needs restructuring. UCNI is beneficial to the Nuclear Emergency Search Team in protecting participants, procedures, and equipment. It also helps the DOE Transportation Safeguards Division20 in a similar manner. Major UCNI simplification can be achieved by eliminating categories not directly related to Safeguards and Security, eliminating the topic reference system, restructuring topical guidelines into performance-based topics, and having local site Internal Guidelines.


1 "Classification Policy Study," U.S. Department of Energy, July 4, 1992.

2 "A Review of Department of Energy Classification Policy and Practice," National Academy Press, July 16, 1995.

3 This construct is based on the work of Arvin Quist: "Security Classification of Information, Volume 2, Principles for Classification of Information," Oak Ridge Gaseous Diffusion Plant, K/CG-1007/V2, A. S. Quist, April 1993.

4 DOE Order 5650.2B, "Identification of Classified Information," December 31, 1991.

5 DOE Classification Appraisal Procedure Guide, February 1988.

6 CG-C-3A, "Classification Policy Guide for Nuclear Programs," DOE Office of [De]Classification, June 1993.

7 Ibid.

8 "Annual Report to the President and the Congress," W. J. Perry, February 1995.

9 Executive Order 12958, "Classified National Security Information," April 17, 1995.

10 The White House, A National Security Strategy of Engagement and Enlargement, July 1994.

11 This first principle acknowledges our understanding of the importance of disclosure. The right of the public to government information is defined in statutes (e.g., Freedom of Information Act). A number of types of information, including classified information, are excluded from disclosure under these statutes. So while the public does not have a right to classified information, we recognize that classification is proper only for information requiring protection. The second principle tightens the concept.

12 It follows that an identified benefit or need is not required to justify declassification and release to the public. Rather, the federal government must establish the reason for classification. If information does not meet classification criteria, it should be released to the public unless otherwise restricted by law.

13 A process for ensuring that the impact of DoD classification decisions on DOE is fully considered is discussed in Chapter 4. If Recommendation 3 is not accepted, a principle on military utilization and system capabilities must be added.

14 "Unclassified Controlled Nuclear Information at Sandia National Laboratories," July 1995.

15 DoD Directive 3150.1, "Joint Nuclear Weapons Development Studies and Engineering Projects," December 27, 1983.

16 "On April 3, 1947, the first chairman of the Atomic Energy Commission, David Lilienthal, informed President Truman that there were no nuclear weapons available for immediate use ... Secrecy on early stockpile numbers was intense, even mystical ..." Danger and Survival, McGeorge Bundy, Vintage Books, Inc., 1990.

17 "Annual Report to the President and the Congress," William J. Perry, February 15, 1995.

18 The Joint Security Commission recently observed that security decisions are frequently linked to assumptions based on an all-knowing, highly competent enemy. They, too, concluded that a more rational and enduring framework for security decision-making can be achieved using risk management. "Redefining Security," a report to the Secretary of Defense and Director of Central Intelligence by the Joint Security Commission, February 28, 1994.

19 Within the National Archives and Records Administration, the Director of the Information Security Oversight Office is responsible for directives implementing Executive Order 12958 and for review and approval of agency implementing regulations.

20 The DOE Transportation Safeguards Division is responsible for shipment of nuclear weapons and certain weapon components and materials.