Comprehensive Nuclear-Test-Ban Treaty: Issues and Arguments

Congressional research reportMar 12, 2008

Ask Donna

What actually matters in this document.

Text

ȱȬȬȱ¢Ǳȱ

ȱȱȱ

ȱȱ

ȱȱȱȱ¢ȱ

ȱŗŘǰȱŘŖŖŞȱ

ȱȱȱ

ŝȬśŝŖŖȱ

ǯǯȱ

řŚřşŚȱ

ȱȱȱ

Prepared for Members and Committees of Congress

ȱȬȬȱ¢Ǳȱ ȱȱȱ

¢ȱ

The Comprehensive Nuclear-Test-Ban Treaty would ban all nuclear explosions. It was opened for

signature in 1996. As of March 2008, 178 nations had signed it and 144 had ratified. To enter into

force, 44 specified nations must ratify it; 35 have done so. The Senate rejected the treaty in 1999;

the Bush Administration opposes it. The United States has observed a nuclear test moratorium

since 1992.

There have been many calls worldwide for the United States and others to ratify the treaty. Many

claim that it would promote nuclear nonproliferation; some see it as a step toward nuclear

disarmament. Several measures have been introduced in Congress regarding the treaty; it might

become an issue in the presidential election.

The U.S. debate involves arguments on many issues. To reach a judgment on the treaty, should it

come up for a ratification vote in the future, Senators may wish to balance answers to several

questions in a net assessment of risks and benefits.

Can the United States maintain deterrence without testing? The treaty’s supporters hold that U.S.

programs can maintain existing, tested weapons without further testing, pointing to 12 annual

assessments that these weapons remain safe and reliable, and claim that these weapons meet any

deterrent needs. Opponents maintain that there can be no confidence in existing warheads because

many minor modifications will change them from tested versions, so testing is needed to restore

and maintain confidence. They see deterrence as dynamic, requiring new weapons to counter new

threats, and assert that these weapons must be tested.

Are monitoring and verification capability sufficient? “Monitoring” refers to technical capability;

“verification” to its adequacy to maintain security. Supporters hold that advances in monitoring

make it hard for an evader to conduct undetected tests. They claim that any such tests would be

too small to affect the strategic balance. Opponents see many opportunities for evasion, and

believe that clandestine tests by others could put the United States at a serious disadvantage.

How might the treaty affect nuclear nonproliferation and disarmament? Supporters claim that the

treaty makes technical contributions to nonproliferation, such as limiting weapons programs;

some supporters believe that nonproliferation requires progress toward nuclear disarmament, with

the treaty a key step. Opponents believe that a strong nuclear deterrent is essential for

nonproliferation, that nonproliferation and disarmament are unrelated, and that this nation has

taken many nonproliferation and disarmament actions that the international community ignores.

This report presents a detailed, comprehensive discussion of the treaty’s pros and cons from a

U.S. perspective. It contains an appendix outlining relevant history. It will be updated periodically

with views from protagonists. CRS Report RL33548, Comprehensive Nuclear-Test-Ban Treaty:

Background and Current Developments, by Jonathan Medalia, tracks current developments.

ȱȱȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

Introduction ..................................................................................................................................... 1

Can the United States Maintain Deterrence Under the CTBT?....................................................... 2

Can the United States Maintain the Nuclear Weapons Enterprise Without Testing? ................ 4

Can the United States Maintain Existing Warheads Without Testing?...................................... 9

Does Deterrence Require New Warheads That Must Be Tested? ........................................... 12

Do U.S. Warheads Require New Surety Features? Is Nuclear Testing Needed to Add

Them? .................................................................................................................................. 14

Does the Treaty Provide Adequate Protection Against Cheating? ................................................ 16

What Does the Treaty Ban?..................................................................................................... 17

How Capable Is the CTBT Monitoring Regime?.................................................................... 19

Monitoring Systems and Methods .................................................................................... 19

Additional Evasion Scenarios ........................................................................................... 34

Would Clandestine Testing Confer Military Advantages? ...................................................... 36

What Risks Does a Nation Run if It Is Caught Cheating? ...................................................... 39

The CTBT, Nuclear Nonproliferation, and Nuclear Disarmament................................................ 40

The Treaty’s Technical Contributions to Nonproliferation ..................................................... 42

“Nuclear Umbrella,” New Weapons, and Nonproliferation.................................................... 42

The CTBT and the NPT’s “Grand Bargain”............................................................................ 44

The CTBT and Nuclear Disarmament .................................................................................... 48

Moratorium and Entry into Force............................................................................................ 50

Conclusion: Alternatives, Packages, and a Net Assessment.......................................................... 52

¡ȱ

Appendix A. History of Nuclear Testing, Test Bans, and Nonproliferation .................................. 56

Appendix B. Abbreviations ........................................................................................................... 64

Author Contact Information .......................................................................................................... 64

ȱȱȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

The Comprehensive Nuclear-Test-Ban Treaty, or CTBT, would ban all nuclear explosions.1 It was

opened for signature in September 1996; as of February 2008, 178 nations had signed it and 144

of them had ratified.2 To enter into force, 44 nations with nuclear reactors must ratify it; so far, 35

of them have ratified and another 7 have signed. The United States signed the treaty in September

1996; the Senate rejected it in October 1999.

Nuclear test bans have a long history (see Appendix A). There has been strong international

support for test ban treaties; U.S. opinion has been divided. Most U.S. Presidents have sought

agreements to limit testing. The Eisenhower Administration devoted great, but unsuccessful,

effort to negotiating a treaty. The Kennedy Administration sought a CTBT; when that proved

nonnegotiable, it achieved the Limited Test Ban Treaty (LTBT) in 1963, which bans nuclear tests

in the atmosphere, under water, and in space. The Nixon Administration negotiated the Threshold

Test Ban Treaty (TTBT) with the Soviet Union in 1974, which limits underground tests to a yield

of 150 kilotons.3 The Ford Administration negotiated the Peaceful Nuclear Explosions Treaty

(PNET) in 1976, which extended the 150-kiloton limit to peaceful nuclear explosions. The Carter

Administration did not pursue entry into force of these two treaties, but sought a CTBT; partly

because of strong opposition within the Administration, no treaty was concluded. The Reagan

Administration rejected the TTBT and PNET because of verification concerns, but in 1987 began

to negotiate new verification protocols. The George H.W. Bush Administration concluded

negotiation of these protocols; the Senate approved the two treaties in 1990, and they entered into

force in that year. President Bush also signed into law a provision implementing a nine-month

moratorium on nuclear testing starting in October 1992. President Clinton extended the

moratorium; he had initially thought to pursue a test ban treaty of limited duration and permitting

a low explosive yield, but in 1995 he opted for a CTBT of zero yield and unlimited duration. The

George W. Bush Administration has continued the moratorium but has not pursued the CTBT.

U.S. interest in the CTBT waned after 1999, but has since reemerged. In the wake of 9/11 and the

rise of nuclear programs in Iran and North Korea, the risk of nuclear proliferation has become

more stark; some claim the treaty would curb that risk. An op-ed in January 2007 by Henry

Kissinger, Sam Nunn, William Perry, and George Shultz called for steps toward eliminating

nuclear weapons, including ratification and entry into force of the CTBT.4 The Administration is

pursuing the Reliable Replacement Warhead (RRW), which it argues would make nuclear testing

less likely; some envision a CTBT-RRW bargain. Scientists around the world have made progress

in detecting nuclear explosions, and U.S. scientists have made progress in maintaining nuclear

weapons without testing; both topics were of concern in the 1999 debate. Others hold that

monitoring capability is insufficient and that new weapons requiring testing are needed.

International pressure for the treaty has continued through U.N. General Assembly votes and

For treaty text, see http://www.state.gov/www/global/arms/treaties/ctb.html. For CTBT developments, see CRS

Report RL33548, Comprehensive Nuclear-Test-Ban Treaty: Background and Current Developments, by Jonathan

Medalia.

For status of signatures and ratifications, see http://www.ctbto.org/.

One kiloton is equivalent to the explosive force of 1,000 tons of TNT; for comparison, the yield of the Hiroshima

bomb was 15 kilotons.

George Shultz, William Perry, Henry Kissinger, and Sam Nunn, “A World Free of Nuclear Weapons,” Wall Street

Journal, January 4, 2007, p. 15.

ȱȱȱ

ŗȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

international conferences. The treaty might be an issue in the presidential campaign.5 Several bills

and resolutions in the 110th Congress call for ratification of the CTBT.6

Opinions on the treaty reflect contending views on how to obtain security; the role of nuclear

weapons; nuclear nonproliferation and its relationship, if any, to nuclear disarmament; and

international relations generally. (1) Some opponents would revoke the U.S. signature of the

treaty and resume testing to maintain U.S. nuclear weapons, weapons expertise, and the

credibility of the nuclear deterrent, and to develop new weapons. (2) Some supporters and

opponents prefer to maintain the moratorium because of concern for political and international

ramifications, but would test if necessary to fix a warhead problem. (3) Some supporters favor the

treaty on grounds that it has significant value for nonproliferation and can help the United States

monitor nuclear testing by other nations. (4) Others favor the CTBT as a step toward abolition of

nuclear weapons. While many people of all stripes favor abolition of nuclear weapons as an

ultimate goal, those in the fourth group see abolition as a realistic if long-term possibility and

believe that the CTBT is a critical step toward reaching that goal. These views are on a

continuum, with overlaps and shades of gray between positions. Still others feel the treaty would

make little difference in restraining weapons development because technical advances enable

such development without testing, or that it would make little difference in countering nuclear

proliferation as a stand-alone measure. While the United States has observed a nuclear test

moratorium since 1992, few appear to hold it as their preferred position; instead, the treaty’s

supporters accept the moratorium as better than a return to testing, and opponents accept it as

better than the CTBT.

This report seeks to present information that may help Members understand many CTBT issues

and to assess whether, on balance, the United States is better off with or without the CTBT. It is

organized around three aspects of how the treaty might affect U.S. security that were prominent in

the 1999 debate: the CTBT and deterrence; monitoring and verification; and implications for

nuclear nonproliferation and disarmament. In the public debate since 1999, CTBT supporters

have written extensively on all aspects of the treaty, while opponents have written much less. To

provide balance, CRS has obtained many comments from people representing all perspectives. As

a result, this report contains a substantial amount of new material.

ȱȱȱȱȱȱȱ

ȱǵȱ

During the Cold War, the United States and Soviet Union engaged in an arms competition, often

called an “arms race” or “action-reaction cycle.” This competition was dynamic. The United

States built submarines carrying ballistic missiles; the Soviet Union followed suit. The Soviet

Union built deeply buried bunkers for its leaders; the United States built very high yield weapons

See “2008 Presidential Candidates’ Responses to Seven Key National Security Questions,” Council for a Livable

World, August 16, 2007, at http://www.clw.org/elections/2008/presidential/

2008_presidential_candidates_questionnaire_responses/.

These include Section 3122 of S. 1547, the FY2008 national defense authorization bill, as passed by the Senate but

not included in the final legislation; H.Res. 68, recognizing the dangers posed by nuclear weapons and calling on the

President to engage in nonproliferation strategies designed to eliminate these weapons of mass destruction from United

States and worldwide arsenals; and H.Res. 882, expressing the sense of the House that the Senate should initiate a

bipartisan process to give its advice and consent to CTBT ratification.

ȱȱȱ

Řȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

to destroy them. Scores of such examples could be listed. Despite this effort, U.S. and Soviet

nuclear strategies and programs resulted in a rough parity between the two sides, and the Cold

War passed into history with no nuclear or conventional war between them.

While deterrence has had many permutations over the years, most in the United States supported

it during the Cold War for want of a better alternative. To be sure, some argued that the United

States should seek superiority, while others held that a minimum deterrent sufficed. Others

reluctantly supported deterrence as an interim measure, arguing that while it purports to reduce

the risk of nuclear war, that very outcome could be expected if a low probability per year is

aggregated over many years. Despite these differing views, Congress supported the forces to

implement a deterrent strategy over many decades. The capability to deter the Soviet Union was

by far the most stressing case, so it was seen as more than sufficient to deter other threats.7 In that

environment, nuclear testing served many purposes. Nuclear tests were mainly conducted for

weapons development, but also for safety, weapons physics, stockpile confidence, and

certification of modifications. Tests also served to maintain skills in weapons science,

engineering, and manufacturing, and to demonstrate the credibility of the U.S. deterrent.

With the end of the Cold War and the Soviet Union, the “comfort” of dealing for four decades

with a single more-or-less predictable adversary ended, to be replaced by considerable

uncertainty. R. James Woolsey, in his 1993 nomination hearing to be Director of Central

Intelligence, said “Yes, we have slain a large dragon, but we live now in a jungle filled with a

bewildering variety of poisonous snakes. And in many ways the dragon was easier to keep track

of.”8

Despite this changed situation, there remains wide, but not universal, agreement in the United

States on the need to maintain a nuclear deterrent for the foreseeable future. Lawrence Korb and

Max Bergmann of the Center for American Progress wrote, “To maintain an effective deterrent,

the United States must continue to possess conventional and nuclear forces capable of quickly

and decisively destroying these regimes,” referring to “extreme regimes ... such as Iran and North

Korea.”9 Sidney Drell and James Goodby, in an Arms Control Association report, “estimate that a

U.S. strategic force of some 500 operationally deployed warheads would be more than adequate

for deterrence. ... this force level would be enough to provide a degree of flexibility in a fluid

security environment.” A responsive force of 400 to 500 warheads would supplement this force.10

The Administration’s Nuclear Posture Review of 2001 stated that with the end of the Cold War,

“U.S. nuclear forces still require the capability to hold at risk a wide range of target types. This

capability is key to the role of nuclear forces in supporting an effective deterrence strategy

relative to a broad spectrum of potential opponents under a variety of contingencies.”11

For information on U.S. nuclear policies, see CRS Report RL34226, Nuclear Weapons in U.S. National Security

Policy: Past, Present, and Prospects, by (name redacted).

U.S. Congress. Senate. Select Committee on Intelligence. Nomination of R. James Woolsey. S.Hrg. 103-296, 103rd

Congress, 1st Session, 1993, p. 76.

Lawrence Korb and Max Bergmann, Restoring American Military Power: Toward a New Progressive Defense

Strategy for America, Center for American Progress, December 2007, p. 17.

Sidney Drell and James Goodby, What Are Nuclear Weapons For? Arms Control Association, revised and updated

October 2007, p. 15.

U.S. Department of Defense. Nuclear Posture Review [Excerpts] Submitted to Congress on 31 December 2001, p. 7,

available at http://www.globalsecurity.org/wmd/library/policy/dod/npr.htm.

ȱȱȱ

řȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

At issue, though, is what is needed for deterrence. The aim of deterrence has always been to make

an adversary fear it will suffer unacceptable consequences if it takes certain actions. Many believe

that the U.S.-Soviet deterrent relationship worked during the Cold War because threats were

credible and each side understood the consequences of attacking the other. In the post-Cold War,

post-9/11 world, many questions arise. Who is to be deterred, by what threats? What weapons are

needed to make them credible? Is deterrence dynamic, with constant weapons development

needed to respond to changing threats, or is a modest number of nuclear weapons of existing

designs, together with U.S. conventional forces and economic might, more than sufficient? Are

existing nuclear weapons sufficient to deter North Korea, or are new ones needed that could

destroy underground bunkers where leaders might hide, or is the nation so irrational that it is

beyond deterrence, or is a North Korean nuclear attack wildly implausible? Is a satisfactory

outcome possible through diplomacy? What capabilities are needed to deter Iran or to roll back its

nuclear program? Do nuclear forces have any relevance to deterring terrorists or their state

sponsors?

This report now considers CTBT and nuclear testing issues that link to these broader issues of

deterrence.

•

Without testing, can the United States maintain the facilities and skilled

personnel supporting U.S. nuclear weapons? This question is considered first

because these capabilities are the bedrock on which nuclear weapons rest.

•

Can existing weapons be maintained without testing? This is a necessary criterion

for deterrence under the CTBT, as it would take many years to develop and

deploy new weapons.

•

Does deterrence require new weapons that incorporate new military capabilities,

and is testing required to develop them?

•

Do U.S. weapons need more features for safety and security, and is testing

required to add them? Such features might deter terrorist attempts to seize and

detonate these weapons.

ȱȱȱȱȱȱȱȱȱ

ȱǵȱ

The nuclear weapons enterprise is here taken to mean the nuclear weapons complex managed by

the National Nuclear Security Administration (NNSA), a semiautonomous agency of the

Department of Energy (DOE) responsible for the U.S. nuclear weapons program;12 scientists,

engineers, and production staff of the complex; and Department of Defense (DOD) agencies that

deal with nuclear weapons. Collectively, they provide the skills and capabilities that support and

would use nuclear weapons.

Whether the United States can maintain this enterprise without nuclear testing has been at issue

for decades. In 1963, the Joint Chiefs of Staff conditioned their support for the LTBT on four

The nuclear weapons complex consists of eight sites: Los Alamos, Livermore, and Sandia National Laboratories;

Pantex Plant, Y-12 Plant, Kansas City Plant, and Savannah River Site, which together produce, maintain, and dismantle

nuclear weapons; and the Nevada Test Site, which until 1992 was used to conduct nuclear tests but is now used for

other nuclear weapons-related activities and other purposes.

ȱȱȱ

Śȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

“safeguards,” or actions this nation would take within the confines of that treaty. The first three

would help maintain this enterprise: Safeguard A, an aggressive underground nuclear test

program; Safeguard B, technology facilities and programs to attract and retain scientists;

Safeguard C, maintenance of the ability to resume atmospheric testing promptly; and Safeguard

D, improvement of monitoring capability.13 President Kennedy’s assurance to Senators Mansfield

and Dirksen, the majority and minority leaders, that the United States would observe these and

other safeguards14 was instrumental in securing Senate advice and consent to ratification. The

safeguards have been observed over time, though Safeguard C has been modified as the perceived

need for atmospheric tests waned and ended. As Appendix A details, other nuclear test limitation

treaties were negotiated and entered into force between 1974 and 1990.

The Hatfield-Exon-Mitchell amendment, Section 507 of the FY1993 Energy and Water

Development Appropriations Act, P.L. 102-377, mandated a nine-month moratorium on nuclear

testing beginning in October 1992, limited testing thereafter, and directed the President to report

on a plan for achieving a CTBT by September 30, 1996. President Clinton extended the

moratorium several times. In response to the prospect of a permanent halt to testing, Congress, in

Section 3138 of P.L. 103-160, the FY1994 National Defense Authorization Act, and the President,

in Presidential Decision Directive 15, mandated a Stockpile Stewardship Program (SSP) to

maintain U.S. nuclear capabilities in a no-test era.

In 1995, President Clinton announced his decision to seek a zero-yield CTBT. He conditioned the

CTBT on six safeguards: (A) SSP, (B) modern laboratory facilities and nuclear technology

programs to attract and retain scientists, (C) the “basic capability to resume nuclear test

activities,” (D) continued R&D to improve the ability to monitor compliance with the treaty, (E)

continued improvement of intelligence capabilities to provide information on nuclear weapons

programs worldwide, and (F) the understanding that if a key nuclear weapon type could no longer

be certified as safe or reliable, “the President, in consultation with Congress, would be prepared

to withdraw from the CTBT under the standard ‘supreme national interests’ clause in order to

conduct whatever testing might be required.”15 Safeguards A, B, C, and F would help maintain

the nuclear weapons enterprise.

In the 1999 CTBT debate, SSP, as the core of U.S. ability to maintain the nuclear weapons

enterprise without testing, was a major issue. SSP had been in being for a short time, resulting in

uncertainty on its ability to maintain existing weapons. Former National Security Adviser Brent

Scowcroft, former Secretary of State Henry Kissinger, and former Deputy Secretary of Defense

John Deutch questioned whether funding would be maintained and wrote that SSP “is not

sufficiently mature to evaluate the extent to which it can be a suitable alternative to testing.”16

Former Secretary of Defense Caspar Weinberger said, “[i]f we need nuclear weapons, we have to

know that they work. That is the essence of their deterrence.... The only assurance that you have

Testimony of General Maxwell Taylor, Chairman, Joint Chiefs of Staff, in U.S. Congress. Senate. Committee on

Foreign Relations, Nuclear Test Ban Treaty, hearings on Executive M, 88th Congress, 1st Session, 1963, pp. 274-275.

Letter from President John Kennedy to Hon. Mike Mansfield and Hon. Everett McKinley Dirksen, in address by

Senator Dirksen on the Nuclear Test Ban Treaty, U.S. Congress. Congressional Record, September 11, 1963, p. 1679016791.

U.S. White House. Office of the Press Secretary. “Fact Sheet: Comprehensive Test Ban Treaty Safeguards.” August

11, 1995, p. 1.

Letter to Hon. Trent Lott, Majority Leader, U.S. Senate, and Hon. Thomas A. Daschle, Minority Leader, U.S. Senate,

October 5, 1999, in U.S. Congress. Senate. Committee on Armed Services. Comprehensive Test Ban Treaty. S.Hrg.

106-490, 106th Congress, 1st Session, 1999 (hereinafter SASC CTBT hearings, 1999), pp. 100-101.

ȱȱȱ

śȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

that they will work is to test them.”17 John Browne, Director of Los Alamos, argued that

Safeguard F was absolutely essential,18 while Weinberger expressed concern that the President

would not exercise it.19 Six former Secretaries of Defense were concerned that an indefiniteduration CTBT could lead to loss of expertise, the topic of President Clinton’s Safeguard B:

Another implication of a CTBT of unlimited duration is that over time we would gradually

lose our pool of knowledgeable people with experience in nuclear weapons design and

testing. Consider what would occur if the United States halted nuclear testing for 30 years.

We would then be dependent on the judgment of personnel with no personal experience

either in designing or testing nuclear weapons. In place of a learning curve, we would

experience an extended unlearning curve.20

Such uncertainties cast doubt for some Senators on the CTBT. Senator Olympia Snowe said,

“there are [SSP] methods that are yet to be proven and we are years or decades away from

knowing whether or not they are reliable.”21 Senator John Warner said, “there are honest

differences on both sides leaving clearly a reasonable doubt, and I come from the old school that

it should be beyond any reasonable doubt if we are going to take a step that affects our vital

security interests for decades to come, indeed possibly into perpetuity as it relates to this cadre of

weapons.”22

The treaty’s defenders tried to give assurance on these points. Secretary of State Madeleine

Albright said, “We have also now said that [the nuclear weapons laboratories] would have $45

billion over a 10-year period to be able to update and keep going all of the various parts of the

stewardship program,” and the United States would “maintain the capability to test again should

the need ever arise.”23 Secretary of Energy Richardson “stress[ed] that the President, in

consultation with Congress, can withdraw from this treaty if a high level of confidence in the

safety and reliability of a nuclear weapon critical to our nuclear deterrent cannot be certified. As

Secretary of Energy, I would not hesitate to so advise the President in the event it becomes

necessary for our country to conduct tests.”24 Senator Carl Levin also emphasized Safeguard F:

if lab directors and other experts ... cannot certify to us 2 years, 4 years, 6 years, 10 years

from now that this is a safe and reliable stockpile, then we are giving everybody notice who

signs this treaty that under our supreme national interest clause we are prepared to withdraw.

So in a sense this treaty is almost a year to year treaty.25

U.S. Congress. Senate. Committee on Foreign Relations. Final Review of the Comprehensive Nuclear Test Ban

Treaty (Treaty Doc. 105-28), S.Hrg. 106-262, 106th Congress, 1st Session, 1999 (hereinafter SFRC CTBT hearing,

1999), p. 14.

SASC CTBT hearings, 1999, p. 111.

SFRC CTBT hearing, 1999, p. 42.

Letter from James Schlesinger, Richard Cheney, Frank Carlucci, Caspar Weinberger, Donald Rumsfeld, and Melvin

Laird to The Honorable Trent Lott, Majority Leader, United States Senate, and The Honorable Tom Daschle,

Democratic Leader, United States Senate, in SASC CTBT hearings, 1999, p. 57.

SASC CTBT hearings, 1999, p. 43.

SFRC CTBT hearing, 1999, p. 52.

SFRC CTBT hearing, 1999, pp. 90, 92.

SASC CTBT hearings, 1999, p. 107.

SASC CTBT hearings, 1999, p. 87.

ȱȱȱ

Ŝȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

How have President Clinton’s safeguards fared since 1999? Safeguards A and B called for SSP

and facilities and programs to attract and retain scientists. CTBT supporters claim that SSP has

made great progress under NNSA. They cite Thomas D’Agostino, then Acting NNSA

Administrator, who said, “stockpile stewardship is working. This program has proven its ability to

successfully sustain the safety, security and reliability of the stockpile without the need to conduct

an underground test for well over a decade.”26 K. Henry O’Brien, RRW Program Manager at

Lawrence Livermore National Laboratory, called SSP a “dramatic success.”27 SSP has developed

sophisticated computer models of nuclear weapons and explosions, has built some of the world’s

most powerful computers, is building the world’s largest laser, and conducts nonnuclear

experiments. Its surveillance program examines warheads for problems, and its Life Extension

Program (LEP) is designed to correct them by replacing components that are, or are expected to

become, defective. Life-extended W87 warheads have been certified for use in the stockpile.

While the first RRW design, “WR1,” is to replace some W76s, Barry Hannah, Chairman of the

RRW Project Officers Group, called the W76 LEP an “excellent program” that he believes “meets

the Navy’s needs.”28 Richard Garwin, IBM Fellow Emeritus who has been involved with nuclear

weapon issues since 1950, does not “agree with the generally stated assumption that confidence

and the reliability of our existing nuclear weapons will inevitably decline with time as the

weapons age.” Instead, “with the passage of time and the improvement in computing tools, I

believe that confidence in the reliability of the existing legacy weapons will increase rather than

diminish.”29 SSP has permitted 12 annual assessments that the U.S. nuclear stockpile is safe and

reliable. It has permitted design of RRW, as discussed later. NNSA is planning to modernize the

nuclear weapons production complex.30 For FY2001-FY2007, SSP received about $42.2 billion;31

its FY2008 current appropriation is $6.3 billion and its FY2009 request is $6.6 billion.32

CTBT opponents are concerned that without nuclear tests that integrate all phenomena, there is

no experimental basis on which designers can be sure that their understanding of a design

corresponds to what they would learn with a nuclear test. As Kathleen Bailey, former Assistant

Director for Nuclear and Weapons Control, Arms Control and Disarmament Agency, testified in

1998, “Virtual reality cannot replace reality.”33 Without new nuclear test data, in this view,

Testimony of Thomas D’Agostino, Acting Administrator, National Nuclear Security Administration, in U.S.

Congress. House. Committee on Appropriations. Subcommittee on Energy and Water Development. Hearing on the

Department of Energy’s FY2008 budget for programs in the National Nuclear Security Administration, 110th Congress,

1st Session, March 29, 2007, transcript by CQ Transcriptions.

Personal communication, April 2, 2007.

Information provided by Dr. Barry Hannah, SES, Branch Head, Reentry Systems, Strategic Systems Program, U.S.

Navy, telephone conversation, October 23, 2006.

U.S. Congress. House. Committee on Appropriations. Subcommittee on Energy and Water Development. Hearing on

nuclear weapon activities, 110th Congress, 1st Session, March 29, 2007. Transcript by CQ Transcriptions.

See, for example, U.S. Department of Energy. National Nuclear Security Administration. Office of Defense

Programs. “Report on the Plan for Transformation of the National Nuclear Security Administration Nuclear Weapons

Complex.” January 31, 2007, 31 p, Available at http://www.nnsa.doe.gov/docs/Trans_of_NNSA_WC_2007-31-07.pdf.

Data for FY2001-FY2004 are for NNSA annual request documents for FY2003-FY2006; data for FY2005-FY2007

are from U.S. Department of Energy. “FY 2007 Operating Plan by Appropriation,” p. 2, http://www.energy.gov/media/

FY2007OperatingPlanForDOE.pdf. NNSA budget documents list SSP funds as “Weapons Activities.”

U.S. Department of Energy. Office of Chief Financial Officer. FY 2009 Congressional Budget Request. Volume 1,

National Nuclear Security Administration. DOE/CF-024, February 2008, p. 71.

“Testimony of Kathleen Bailey, Senior Fellow, Lawrence Livermore National Laboratory,” in U.S. Congress.

Senate. Committee on Governmental Affairs. Subcommittee on International Security, Proliferation, and Federal

Services. The Comprehensive Test Ban Treaty and Nuclear Nonproliferation, S. Hrg. 105-699, 105th Congress, 2nd

Session, 1998, p. 26.

ȱȱȱ

ŝȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

stewardship tools are unvalidated, so certifications are political statements and it is not possible to

be certain that the stockpile is safe and reliable.34 Supporters say that the computer models are

valid because they fit a vast array of experimental data, notably including the results of the U.S.

nuclear test program; critics respond that while the performance of an individual electronic

component can be validated through repeated testing, a nuclear explosion is an integrated event

that cannot be predicted by analyzing the performance of individual components. It is a different,

and easier, exercise to fit computer models to past tests, they argue, than to see how well a

computer model predicts the outcome of a future test.

SSP rests on skilled personnel. CTBT opponents point to concerns raised by Carol Burns of Los

Alamos National Laboratory: “In 2006, NNSA indicated that about 40% of nuclear weapons

program technical staff members were eligible for retirement.” She noted a decline in production

of students with doctoral degrees in nuclear science, and pointed to a drop in doctoral degrees

earned at U.S. universities in radiochemistry and nuclear chemistry from 33 in 1968 to 4 in

2003.35 Opponents see problems with LEPs. As Ambassador Linton Brooks, then Administrator

of NNSA, said in 2005, “it is becoming more difficult and costly to certify warhead

remanufacture. The evolution away from tested designs resulting from the inevitable

accumulations of small changes over the extended lifetimes of these systems [i.e., warheads]

means that we can count on increasing uncertainty.”36 John Foster, former Director of Defense

Research and Engineering, raised other concerns:

The Stockpile Stewardship Program has been a lifesaver for the nuclear weapons labs. It has

attracted and maintained scientists and engineers and provided new world-class tools for

understanding nuclear weapon performance and advancing weapon science. But I have three

salient concerns. First, U.S. nuclear weapon pit production was stopped in 1989, leading

quickly to a halt in weapons production. The capability to produce nuclear weapons has

atrophied since then. Second, we have not conducted underground nuclear tests since 1992

and we are running risks regarding the safety, reliability and performance of the stockpile.

Third, periodic surveillance of the aging stockpile has revealed the necessity to initiate Life

Extension Programs to refurbish several warhead types. This process introduces new

materials and components into the warheads, which introduces the possibility of “birth

defects” that raise risks.37

Supporters claim that Safeguard C, the “basic capability to resume nuclear test activities,” has

been met, as NNSA reduced the time needed to conduct a nuclear test from 36-plus months to 24

months.38 Opponents respond that without nuclear testing, the capability to test declines as skills

This view provided by Kathleen Bailey, former Assistant Director for Nuclear and Weapons Control, U.S. Arms

Control and Disarmament Agency, personal communication, April 20, 2007.

“Testimony of Dr. Carol J. Burns, Group Leader, Nuclear and Radiochemistry, Los Alamos National Laboratory,

Before the U.S. House of Representatives, Committee on Homeland Security, Subcommittee on Emerging Threats,

Cybersecurity and Science and Technology, Hearing on H.R. 2631, the Nuclear Forensics and Attribution Act,”

October 10, 2007, pp. 2-4, http://homeland.house.gov/SiteDocuments/20071010175138-84437.pdf.

“Statement of Ambassador Linton F. Brooks, Administrator, National Nuclear Security Administration, U.S.

Department of Energy, Before the Senate Armed Services Committee, Subcommittee on Strategic Forces,” April 4,

2005.

Personal communication, October 22, 2007.

A 2003 NNSA report stated, “Over the past several years the NNSA conducted reviews of the 24- to 36-month test

readiness posture [i.e., the time between a presidential decision to conduct a nuclear test and the actual conduct of that

test] that the NNSA has maintained since Fiscal Year 1996. ... From these reviews, NNSA concluded that because of a

loss of expertise and degradation of some specific capabilities, the U.S. would more likely require about 36 months to

test, with less confidence in being able to achieve the 24-month end of the range. Furthermore, as time passes without

(continued...)

ȱȱȱ

Şȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

atrophy, procedures become outdated, and equipment falls into disuse. Safeguards D and E do not

deal with SSP. One cannot prove whether the United States would withdraw from the CTBT, as

per Safeguard F, especially as it has not ratified the treaty. U.S. withdrawal from the Antiballistic

Missile Treaty in 2002 might make the prospect of withdrawal from the CTBT appear more

credible, though critics see prospects for withdrawal dependent on who is President, and thus

uncertain.

ȱȱȱȱȱ¡ȱȱ

ȱǵȱ

During the Cold War, as noted, deterrence was dynamic, with nuclear moves and counter-moves

by the United States and Soviet Union. Testing was essential for both sides to develop new

weapons. In the 1999 debate, arguments over the treaty and deterrence played a minor, and

predictable, part. Both sides in the debate agreed that maintaining the nuclear deterrent was

crucial. Opponents held that without testing, it would be impossible to do so. As former Secretary

of Defense James Schlesinger testified, “In the absence of testing, confidence in the reliability of

the stockpile will inevitably, ineluctably decline.”39 They questioned whether the United States

could, in 1999 if ever, rely on SSP to maintain weapons. The treaty’s supporters had a different

view. Secretary of State Madeleine Albright said, “Under the treaty, America would retain a safe

and reliable nuclear deterrent.”40 And General Henry Shelton, Chairman of the Joint Chiefs of

Staff, testified:

Senator Levin: What you are telling us is that our top uniformed leadership unanimously

support this Treaty?

General Shelton: I might add, Senator Levin, that we would never say that unless we felt that

we could maintain a credible nuclear deterrent and also a safe and reliable stockpile.41

Since 1999, support has continued for this nation to maintain nuclear weapons as long as it retains

them. There are three main approaches for so doing. Supporters of the Reliable Replacement

Warhead (RRW) program and supporters of the Life Extension Program (LEP) each argue that

their approach will reduce the likelihood of testing while the other will increase it. In contrast,

others believe that neither RRW nor LEP can provide sufficient confidence in the safety and

reliability of current warheads without nuclear testing; they therefore see testing as necessary.

RRW, as a funded program, began in the FY2005 Consolidated Appropriations Act, P.L. 108-447;

it was described as a “program to improve the reliability, longevity, and certifiability of existing

(...continued)

further action, the 36-month posture is viewed as increasingly at risk.” U.S. Department of Energy. National Nuclear

Security Administration. Report to Congress: Nuclear Test Readiness. April 2003, p. 5. In contrast, NNSA said that in

FY2005 it “[a]chieved a 24-month [test] readiness posture.” U.S. Department of Energy. Office of the Chief Financial

Officer. FY 2007 Congressional Budget Request. Volume 1, National Nuclear Security Administration. DOE/CF-002,

February 2006, p. 95. However, the FY2009 NNSA request plans to “maintain a minimum readiness posture of 24 to

36 months.” U.S. Department of Energy. FY 2009 Congressional Budget Request. Volume 1, National Nuclear

Security Administration, p. 147.

SASC, CTBT hearings, 1999, p. 59.

SFRC, CTBT hearing, 1999, p. 72.

SASC, CTBT hearings, 1999, pp. 23-24.

ȱȱȱ

şȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

weapons and their components.”42 In the FY2006 National Defense Authorization Act, P.L. 109163, Congress set as an objective that the program “further reduce the likelihood of the

resumption of underground nuclear weapons testing.” The first proposed RRW, WR1, would be

used in place of some W76 warheads on Trident II submarine-launched ballistic missiles. WR1s

would be designed to meet post-Cold War requirements, such as enhanced safety, increased ease

of manufacture, and high confidence without nuclear testing. However, the FY2008 Consolidated

Appropriations Act, P.L. 110-161, eliminated RRW funds, leaving its prospects unclear. An issue

for any future CTBT debate is which approach—RRW or LEP—is less likely to require nuclear

testing in the long term.43

NNSA claims that RRW will make the need for testing unlikely because of steps to increase

confidence. For example, RRW designers used high margins, basically building in more

performance than is needed, to make material deterioration or design or manufacturing defects

less likely to degrade warhead performance below the minimum required. They argued that they

could do so because the design was unconstrained by technologies and design choices made

decades ago. They view added margin as the single most important goal of the design. Another

basis for confidence is that the design stayed close to past experience. Lawrence Livermore

National Laboratory, which designed the nuclear components of WR1, states that components

very similar to those of the WR1 were nuclear tested in the past. For this and other reasons, “there

is direct nuclear test proof that the [WR1] design will perform properly.”44

NNSA and its labs have expressed concerns that, over the long term, minor changes to current

warheads through repeated LEPs will introduce defects and make it harder to maintain reliability,

possibly requiring nuclear testing. They argue that LEPs replace defective or deteriorated

components with replicas. As Thomas D’Agostino said, “The W76 LEP and the life extension

approach is an exact rebuild of what we’ve had in the Cold War stockpile. We try to mimic the

manufacturing processes exactly the way it was done 30 years ago.”45 The concern is that

components and manufacturing processes cannot be replicated precisely, pushing the warhead

beyond the design envelope validated by nuclear testing.46 This problem could result in defects in

life-extended warheads that could cause them to fail.

LEP supporters question whether RRW will provide high confidence. As Steven Fetter of the

University of Maryland said, “Like most other warheads, RRW will have, or could be expected to

have, birth defects or reliability problems that would be discovered and corrected soon after the

warhead was deployed. No one can say whether the unreliabilities introduced by these birth

defects would be greater or smaller than the unreliabilities that would crop up in the existing

warheads due to their age.”47 They thus doubt that a new-design RRW can be certified without

U.S. Congress. Committee of Conference. Making Appropriations for Foreign Operations, Export Financing, and

Related Programs for the Fiscal Year Ending September 30, 2005, and for Other Purposes, conference report to

accompany H.R. 4818, 108th Congress, 2nd Session, H.Rept. 108-792, 2004, p. 951.

For more detail, see CRS Report RL33748, Nuclear Warheads: The Reliable Replacement Warhead Program and

the Life Extension Program, by Jonathan Medalia.

Information provided by Lawrence Livermore National Laboratory, September 19, 2006.

Testimony of Thomas D’Agostino to House Appropriations Committee, Subcommittee on Energy and Water

Development, March 29, 2007.

On this point, see George Miller, Paul Brown, and Carol Alonso, Report to Congress on Stockpile Reliability,

Weapon Remanufacture, and the Role of Nuclear Testing, Lawrence Livermore National Laboratory Report UCRL53822, October 1987, Chapter 3, “Weapon Remanufacture,” pp. 25-30.

Arms Control Association, “The Future of U.S. Nuclear Weapons: The Weapons Complex and the Reliable

(continued...)

ȱȱȱ

ŗŖȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

testing. Robert Peurifoy, a former vice president at Sandia National Laboratories, stated, “The

present nuclear weapon stockpile contains 8 or so nuclear weapon types. That population has

enjoyed perhaps 100 successful yield tests. These weapons have benefitted from a test base of

perhaps 1,000 yield tests conducted during the 40 or so years when nuclear testing was allowed.

Is the DoD really willing to replace tested devices with untested devices?”48

LEP’s supporters argue that current warheads are reliable, as evidenced by 12 stockpile

assessments, and that LEP can keep them reliable for many years without testing. While problems

emerge, solutions do as well, and LEP supporters argue that SSP has been keeping at least even in

this race. RRW supporters agree that SSP is making progress; an NNSA official stated, “Each

year, we are gaining a more complete understanding of the complex physical processes

underlying the performance of our aging nuclear stockpile.”49 Further, say LEP advocates, current

warheads stay within design parameters validated by nuclear tests. In this view, SSP and LEP can

maintain margins through careful remanufacture to minimize changes. They also state, to general

agreement, that margins for some warheads could be increased in certain ways with no change to

a warhead.50 While RRWs, as new designs, are likely to have “birth defects,” LEP supporters

claim such defects have been wrung out of existing designs.

Some, however, doubt that either LEP or RRW can be assessed as reliable, in the case of RRW

because it will never be tested and in the case of LEPs because small changes will undermine

confidence in reliability.51 In this view, SSP has enabled only political assessments rather than

technical ones. Since SSP emerged after the moratorium on testing began, these critics hold that

its tools were never validated with nuclear tests dedicated to that purpose, so they could lead to

false conclusions. Accordingly, in this view, NNSA will not know for sure if SSP, and thus RRW

or LEP, work until it conducts nuclear tests. With confidence in the U.S. nuclear arsenal—by the

United States, its friends, and its foes alike—central to deterrence, in this view, the United States

must conduct nuclear tests regardless of political concerns because only testing can maintain

confidence.52

This section has discussed three views: RRW is less likely to require testing than LEP; LEP is less

likely to require testing than RRW; and the United States can have confidence in neither RRW nor

LEP without testing. One could argue a fourth view, that both RRW and LEP are unlikely to need

(...continued)

Replacement Warhead,” press briefing, Washington, DC, April 19, 2007.

Personal communication, September 24, 2006.

“Statement of Thomas P. D’Agostino, Deputy Administrator for Defense Programs, National Nuclear Security

Administration, Before the House Armed Services Committee, Subcommittee on Strategic Forces,” April 5, 2006, p. 1.

One such change involves a revised means of dealing with the boost gas, a mixture of tritium and deuterium gases

injected into the pit to increase its explosive energy. A study found, “Primary yield margins can be increased by

appropriate changes specific to each stockpile system. These include changes to initial boost-gas composition, shorter

boost-gas exchange intervals, or improved boost-gas storage and delivery systems. These modifications have been

validated by nuclear test data for the appropriate systems, and they would not place burdens on the maintenance or

deployment of the systems by the military.” National Academy of Sciences, Committee on Technical Issues Related to

Ratification of the Comprehensive Nuclear Test Ban Treaty, Technical Issues Related to the Comprehensive Nuclear

Test Ban Treaty, Washington, National Academy Press, 2002 (hereinafter NAS report), p. 31 See also JASON report

JSR-99-305, Primary Performance Margins, McLean, VA, MITRE Corporation, 1999, p. 2.

Information provided by Robert Barker, former Assistant to the Secretary of Defense for Atomic Energy, November

29, 2006.

Information provided by Kathleen Bailey, November 28, 2006.

ȱȱȱ

ŗŗȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

testing. This view could lead to a mixed LEP-RRW force. As Henry O’Brien of Lawrence

Livermore National Laboratory stated, “Our best approach for a small stockpile and complex

would be to retain a couple of the better current weapon types (i.e., those with relatively higher

margins, more advanced safety and security technologies, and more sustainable materials), and

replace the rest with a small number of RRW types.”53

ȱȱȱ ȱȱȱȱȱǵȱ

CTBT opponents argue that the ability to maintain existing weapons without testing through LEP,

even if it can be done, misses the point. Deterrence, as they see it, requires continuing to hold at

risk assets that enemy leaders prize. However, they argue, current nuclear warheads have many

limitations.

•

Current warheads, which were designed during the Cold War, were given high

yield to destroy hard targets like Soviet missile silos. But that yield, in this view,

could cause the United States to refrain from using these weapons out of concern

for inflicting massive civilian casualties in the target area and beyond. As a 2006

Defense Science Board study stated, “weapons that are not seen as useable and

effective by potential adversaries cannot be an effective, reliable deterrent.”54

•

Current warheads, if exploded near the Earth’s surface, would leave much

residual radiation that would contaminate large areas and kill many people,

barring the United States from using them, the treaty’s opponents believe.

•

The radiation output of current warheads, they argue, differs from that needed for

such missions as destroying chemical or biological agents or generating

electromagnetic pulse.

•

Current warheads cannot destroy key targets that enemy leaders would value

highly, such as hardened and deeply buried bunkers where weapons of mass

destruction, key communications nodes, or the leaders themselves might hide.

WR1 shares these limitations. For example, it would have about the same yield as the W76 it

would replace, and would use a reentry body55 that cannot penetrate the ground.

CTBT opponents see deterrence as dynamic, so that it continues to require new military

capabilities that can only be embodied in new weapons that could only be developed with nuclear

testing. The Threat Reduction Advisory Committee, an expert panel advising DOD, stated that

one reason to test would be “[t]o support certification—prior to quantity production—of new

nuclear weapons, should the decision be made that a new weapon design requiring testing is the

only option to achieve a needed capability.” It provided examples of weapons requiring “tailored

physics package design for nuclear effects for new missions,” including:

Personal communication, November 7, 2007.

U.S. Department of Defense. Office of the Under Secretary of Defense for Acquisition, Technology, and Logistics.

Defense Science Board. Report of the Defense Science Board Task Force on Nuclear Capabilities. December 2006.

Report summary, p. 15, original emphasis.

A reentry body, also called a reentry vehicle or aeroshell, is the cone-shaped device that contains a single warhead

and protects it from heat and other stresses as it reenters the atmosphere on the way to its target.

ȱȱȱ

ŗŘȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

•

Earth-penetrating warheads with reduced collateral effects to defeat hard, deeply

buried targets;

•

Warheads to defeat chemical or biological sites ... while simultaneously

neutralizing released chem-bio agents;

•

Reduced residual radiation warheads.56

The 9/11 attacks brought concerns about nuclear terrorism to the fore, and raised questions about

the link between nuclear weapons and deterrence of rogue states and terrorists. According to the

Nuclear Posture Review of December 2001,

Greater flexibility is needed with respect to nuclear forces and planning than was the case

during the Cold War. The assets most valued by the spectrum of potential adversaries in the

new security environment may be diverse and, in some cases, US understanding of what an

adversary values may evolve. Consequently, although the number of weapons needed to hold

those assets at risk has declined, US nuclear forces still require the capability to hold at risk a

wide range of target types.57

The treaty’s opponents see another value in testing. According to Vice Admiral Robert Monroe

(USN, Ret.), former Director of Defense Nuclear Agency, “an ongoing underground nuclear test

program adds immensely to the credibility of the U.S. deterrent. Conversely, failure to test

virtually destroys the credibility of our nuclear forces. A nation which lacks the strength to test

nuclear weapons will almost surely lack the strength to use them.”58

CTBT supporters hold that current nuclear weapons suffice for deterrence; no adversary leader

would gamble that they would not work, or that the United States would not use them if severely

provoked. At the same time, supporters see nuclear weapons as most unlikely to be used,

regardless of their characteristics or yield, because of the norm that has built up since 1945

against their use. Current nuclear weapons deterred a Russian or Chinese nuclear attack during

the Cold War, it is argued, and will continue to do so, especially as the probability of such attack

must be judged as remote. U.S. conventional forces, the treaty’s supporters claim, deter threats

from other nations. Use of these forces is credible, they can be precisely targeted, and they would

create very much less collateral damage than nuclear weapons.

Further, it is argued, adversaries could readily counter new U.S. nuclear capabilities. Nuclear

weapons to destroy chemical or biological weapons could be defeated by placing the weapons

deep underground; even earth penetrator weapons could not destroy them because the heat and

radiation of the blast would not reach down that far. More simply, the weapons could be moved to

nondescript buildings in cities or to caves in rural areas; U.S. intelligence, in this view, could

locate few if any sites. Earth penetrators could be defeated by deeper burial, greater hardening,

tunneling under a mountain, or dispersing assets to secret aboveground locations.

Threat Reduction Advisory Committee. Nuclear Deterrent Transformation Panel. Underground Nuclear Testing:

Issues Regarding Resumption, approved for limited distribution, October 2003, updated for general distribution, March

2005, p. 6.

Nuclear Posture Review [Excerpts], submitted to Congress on 31 December 2001, at http://www.globalsecurity.org/

wmd/library/policy/dod/npr.htm.

Personal correspondence, November 26, 2007, and January 29, 2008.

ȱȱȱ

ŗřȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

The treaty’s proponents see several congressional actions as implying that Congress would not

support testing to develop new weapons. In the last several years, Congress terminated the

“bunker buster” Robust Nuclear Earth Penetrator (RNEP)59 and the Advanced Concepts Initiative,

widely but erroneously thought to be developing a “mini-nuke.” It specified in the FY2006

National Defense Authorization Act that an objective of the RRW program was to further reduce

the likelihood of a return to testing. It eliminated FY2008 funding for RRW.

ȱǯǯȱȱȱ ȱ¢ȱǵȱ ȱȱ

ȱȱȱȱǵȱ

While there are several definitions, surety is here taken to include safety, security, use control, and

use denial. Safety involves protecting a warhead against accidental detonation; security is handled

through a layered approach that includes everything from warhead features to physical security;

use control permits authorized persons to use a warhead only at the direction of the national

command authority; and use denial prevents any unauthorized use of a nuclear weapon. Surety

has always been the most important characteristic in nuclear weapons design, and its technology

has constantly improved, such as with several generations of permissive action links that require a

user to enter a code in order to arm the weapon, and with various safety enhancements. During

the Cold War, nuclear testing was routine, so the question of whether testing was essential for

incorporating these features was moot.

In 1999, CTBT opponents argued that new surety features could and should be added to U.S.

warheads, and could only be added through nuclear testing. In 1997, Siegfried Hecker, then

Director of Los Alamos, testified that “with a CTBT it will not be possible to make some of the

potential safety improvements for greater intrinsic warhead safety that we considered during the

1990 time frame.”60 Robert Barker, former Assistant to the Secretary of Defense for Atomic

Energy, said in 1999, “Of the nine types of weapons that will remain in the inventory only three

types have all three of the most modern safety features while three types have only one such

feature. These safety deficiencies will remain as long as we cannot conduct the necessary nuclear

tests.”61 Secretary of Energy Richardson, in contrast, stated, “Seven years after our last

underground test our stockpile of nuclear weapons is safe and reliable. Three times since 1996 the

Secretary of Energy and the Secretary of Defense have certified this to the President.... Our

nuclear deterrent will continue to be safe and reliable under the Comprehensive Test Ban

Treaty.”62

Also at issue was the need for new surety features. Sidney Drell, emeritus professor of physics at

Stanford University, said in 1999,

I did not support the CTBT then [in 1990]. I thought of some further safety improvements. I

presented some arguments.

For a discussion of congressional handling of RNEP, see Jonathan Medalia, “Water Power: Why Congress Zeroed

“Bunker Buster” Appropriations,” Comparative Strategy, no. 26, 2007, pp. 231-248.

S.S. Hecker, “Answers to Senator Kyl’s questions,” attachment to letter from S.S. Hecker, Director, Los Alamos

National Laboratory, to Honorable Jon Kyl, September 24, 1997, in U.S. Congress. Senate. Committee on

Governmental Affairs. Subcommittee on International Security, Proliferation, and Federal Services. Safety and

Reliability of the U.S. Nuclear Deterrent. Senate Hearing 105-267, 105th Congress, 1st Session, 1997, p. 84.

SASC CTBT hearings, 1999, p. 175.

“Prepared Statement by Secretary Bill Richardson,” in SASC CTBT hearings, 1999, p. 109.

ȱȱȱ

ŗŚȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

First of all, the Department of Defense had zero interest. It wanted to spend no money on

making them. Second, some of the problems have been retired. Others have been altered by

handling procedures in the Navy, and they have satisfied themselves and the Department of

Defense that the safety requirements are safe and sound now.63

Others took the opposite view. Bailey and Barker argued, “Given the increasing threat of

terrorism, it would seem prudent to ensure that U.S. nuclear weapons are as safe, secure, and

invulnerable to unauthorized use as possible.”64

In the wake of 9/11, surety has become even more important. As Linton Brooks said in 2005, “We

now must consider the distinct possibility of well-armed and competent terrorist suicide teams

seeking to gain access to a warhead in order to detonate it in place.”65 The prompt response,

adding physical security, has been costly. Added use-denial features could reduce the burden on

guard forces.

Surety features, it is argued, would enhance deterrence, though in a different way than during the

Cold War. One form of nuclear attack would be for suicide terrorists to seize a U.S. nuclear

weapon and detonate it in place; another would be for terrorists to seize a U.S. nuclear weapon,

dismantle it, and use its fissile material to build a weapon. It is difficult at best to deter terrorists

by threatening to use nuclear weapons to destroy a city or training camp in response to a terrorist

nuclear attack; they might view U.S. nuclear use as desirable if it turned many nations against the

United States. Instead, it is hoped, enhanced surety features would deter attack by creating an

unacceptable consequence, namely a high probability of failure. In addition, if such attacks were

to occur, enhanced surety might defeat them.

Weapon designers and NNSA argue that the WR1 design shows that surety features can be added

without testing, and see RRW as essential to obtaining them. Livermore states that the relaxation

of weight constraints for WR1, for example, has allowed a design that incorporates revolutionary

advances in safety and security without nuclear testing.66 In contrast, according to NNSA

testimony, “[m]ajor enhancements in security are not readily available through system retrofits

via the LEP approach.”67

CTBT supporters dismiss enhanced surety as an argument for testing. They see current weapons

as safe enough, as shown by 12 assessments and the absence of accidental U.S. nuclear

detonations. They see a goal of as much surety as possible as a recipe for unending generations of

weapons to add new features. They also see scenarios involving terrorist seizure and detonation

of U.S. warheads as far-fetched because of physical security measures, and feel that such

measures could be enhanced to add surety if needed. They doubt that new surety features that can

be added only by testing are so critical as to warrant testing.

SASC CTBT hearings, 1999, p. 180.

Kathleen Bailey and Robert Barker, “Why the United States Should Unsign the Comprehensive Test Ban Treaty and

Resume Nuclear Testing,” Comparative Strategy, no. 22, 2003, p. 132.

“Statement of Ambassador Linton F. Brooks, Administrator, National Nuclear Security Administration, U.S.

Department of Energy, Before the Senate Armed Services Committee, Subcommittee on Strategic Forces,” April 4,

2005.

Information provided by Lawrence Livermore National Laboratory, personal communication, May 10, 2007.

“Statement of Thomas P. D’Agostino, Acting Under Secretary for Nuclear Security and Administrator, National

Nuclear Security Administration, U.S. Department of Energy, Before the Committee on House Armed Services [sic],

Subcommittee on Strategic Forces,” March 20, 2007, p. 4.

ȱȱȱ

ŗśȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

CTBT opponents favor the most surety possible in light of the terrorist threat, and hold that more

surety features can be added with testing than without. While it is possible to add guns, gates, and

guards, so doing would be very costly. They maintain that current warheads are not as safe and

secure as possible, and argue that their surety can only be increased through testing. While RRW

offers more advanced surety features than do current warheads, CTBT opponents hold that the

United States can never know if these features will work without testing. They see testing as

needed also to reveal if new surety features on existing warheads or RRWs would impact

performance.

ȱȱ¢ȱȱȱȱ

ȱǵȱ

Monitoring and verification have been central to the debate and negotiations on nuclear test bans

for a half-century.68 While the terms are often used interchangeably, there is a difference.

Monitoring involves looking for indicators that a nuclear test has taken place. It is a dynamic

contest between hiders and seekers, with CTBT supporters showing that monitoring capability is

improving and treaty opponents raising doubts about that capability and claiming that evasion

capability is improving.

Verification, literally “truth making,” involves deciding whether a nation is in compliance with its

treaty obligations. At issue is not perfect verification but effective verification. In 1988, Paul

Nitze offered a widely-used definition: by effective verification, “[w]e mean that we want to be

sure that, if the other side moves beyond the limits of the treaty in any militarily significant way,

we would be able to detect such violation in time to respond effectively, and thereby deny the

other side the benefit of the violation.”69 Thus monitoring is a technical activity that provides

data, while verification uses the data to form judgments on compliance. It is for this reason that

the CTBT establishes an International Monitoring System and leaves it to individual nations to

determine whether a nation has violated the treaty.

Monitoring capability, the military value of clandestine tests, and effective verification are linked.

If, as a hypothetical example, tests above 0.1 kiloton had significant military value and the

threshold of detection was 10 kilotons, the CTBT could not be effectively verified, but it could be

if the numbers were reversed. Thus CTBT opponents claim the threshold for detection is high and

that for military value is low; supporters make the opposite claim. Accordingly, the following

section examines what the treaty bans; describes several monitoring technologies and arguments

about their capabilities and weaknesses; considers whether clandestine testing would confer

military advantages; and discusses risks a nation might run if it is caught cheating.

The public 1999 debate on ratification did not go into detail on the technical ability to monitor the

CTBT. For example, no scientists with primary expertise in a monitoring technology testified in

For discussions of test ban monitoring and verification issues up to the early 1960s, see Harold Karan Jacobson and

Eric Stein, Diplomats, Scientists, and Politicians: The United States and the Nuclear Test Ban Negotiations, Ann

Arbor, University of Michigan Press, 1966, 538 p.; and Benjamin Greene, Eisenhower, Science Advice, and the

Nuclear Test-Ban Debate, 1945-1963, Stanford, CA, Stanford University Press, 2007, 358 p.

U.S. Congress. Senate. Committee on Foreign Relations. The INF Treaty. S.Hrg. 100-522, pt. 1, 100th Congress, 2nd

Session, 1988, part 1, p. 289.

ȱȱȱ

ŗŜȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

Senate hearings on the treaty. However, members and staff received extensive classified briefings

from scientists from the national laboratories and from the intelligence community.70 Since 1999,

scientists have made many advances in detection capability that have been widely published. The

most important technical report on monitoring was prepared in 2002 by the National Academy of

Sciences (NAS).71 It is generally favorable to the treaty. Two other overviews of technical

progress prepared in 2007 also favor the treaty.72 Many journal articles discuss specific technical

advances. In contrast, few if any unclassified technical reports rebut claims of progress in

monitoring. Nevertheless, CTBT opponents have developed many arguments, so any future

debate on monitoring is likely to be less lopsided than one might infer from the imbalance in

writing.

ȱȱȱ¢ȱǵȱ

Article I of the CTBT sets out the treaty’s basic obligation: “Each State Party undertakes not to

carry out any nuclear weapon test explosion or any other nuclear explosion....” The treaty does

not define “nuclear explosion.” Yet it is physically possible to conduct tiny nuclear explosions

that cannot be detected without cooperative measures. For example, the United States conducted

several dozen “hydronuclear” tests, many releasing fission energy equivalent to less than a gram

of high explosive, during the 1958-1961 nuclear test moratorium.73 As discussed later, some see

the prospect of undetected tests of very low yield as a concern. As a result, a point of contention

in the 1999 debate was whether the treaty barred very low yield tests. Some CTBT critics argued

that Russian and U.S. definitions of zero differed. Senator Richard Shelby referenced “public

statements from the Russian First Deputy Minister of Atomic Energy that Russia intends to

continue to conduct low-yield hydronuclear tests and does not believe that these constitute

nuclear tests prohibited by the treaty.”74 In this view, then, Russia might conduct militarily useful

low-yield nuclear tests and still consider itself as observing the CTBT.

Administration officials responded that all parties understood the treaty was zero yield. Under

Secretary of State John Holum said that the treaty “does ban any nuclear test explosion or any

other nuclear explosion, and in the negotiating record it is very clear that that means there cannot

be any critical yield from a nuclear event. You can do things that do not go critical; you cannot do

things that do.”75 76 Ambassador Stephen Ledogar, who retired from the Foreign Service in 1997

Personal communication, Bureau of Verification, Compliance, and Implementation, U.S. Department of State,

January 25, 2008.

NAS report.

David Hafemeister, “Progress in CTBT Monitoring Since Its 1999 Senate Defeat,” Science and Global Security 15,

2007, pp. 151-183; and Raymond Jeanloz, “Comprehensive Nuclear-Test-Ban Treaty and U.S. Security,” paper

prepared for delivery at conference, “Reykjavik Revisited: Steps Toward a World Free of Nuclear Weapons,” Hoover

Institution, Stanford University, October 24-25, 2007.

Robert Thorn and Donald Westervelt, “Hydronuclear Experiments,” Los Alamos National Laboratory, LA-10902MS, UC-2, February 1987, p. 4-5.

SFRC CTBT hearing, 1999, p. 56.

SFRC CTBT hearing, 1999, p. 99.

This paragraph explains terms and concepts relevant to the question of what is a nuclear explosion. A fissile material

is one whose atoms split (fission) when struck by a neutron regardless of its speed; uranium-235 and plutonium are the

fissile materials used in atomic bombs. Each nuclear fission releases a tiny amount of energy, as well as two or three

neutrons. A self-sustaining nuclear chain reaction occurs if the number of neutrons produced by fission equals the

number of neutrons that escape the material or are absorbed within it without causing further fissions. “Criticality” is

the point at which this chain reaction occurs; a “critical mass” is the amount of fissile material just enough to support

(continued...)

ȱȱȱ

ŗŝȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

and was the chief negotiator for the CTBT under Presidents Reagan, Bush, and Clinton,

elaborated:

As the name suggests, the treaty imposes a comprehensive ban on all nuclear explosions, of

any size, in any place. I have heard some critics of the treaty seek to cast doubt on whether

Russia, in the negotiating and signing of the treaty, committed itself under treaty law to a

truly comprehensive prohibition of any nuclear explosion, including an explosion or

experiment or event of even the slightest nuclear yield. In other words, did Russia agree that

hydronuclear experiments which do produce a nuclear yield, although usually very, very

slight, would be banned and that hydrodynamic explosions, which have no yield because

they do not reach criticality, would not be banned.

The answer is a categoric “yes.” The Russians as well as the rest of the P-5 [China, France,

Russia, the United Kingdom, and the United States, the permanent five members of the U.N.

Security Council] did commit themselves. That answer is substantiated by the record of the

negotiations at almost any level of technicality and national security classification that is

desired and permitted. More importantly, for the current debate, it is also substantiated by the

public record of statements by high level Russian officials as their position on the question of

thresholds evolved and fell into line with the consensus that emerged.77

The issue remains unresolved. In a 2007 letter, the State Department stated:

the Department of State is not aware of any international agreement on what “zero” yield

means. During the negotiation of the Treaty, the P-5 reached an understanding that

subcritical nuclear experiments would not be prohibited under the Treaty. The United States

also made clear that, in its view, supercritical nuclear explosive-driven device tests would be

prohibited under the Treaty. However, there was no agreement among the P-5 that criticality

would be the basis for determining which activities would be permitted under the CTBT and

which activities would not be permitted. Therefore, it is left to the individual State Party to

decide for itself whether a test that produced more than a zero yield would violate the

Treaty.78

(...continued)

criticality. The amount of material for a critical mass depends on many factors, such as shape, density, impurities that

absorb neutrons, and use of material to reflect neutrons back into the fissile material. A nuclear reactor is an example of

a critical chain reaction; it releases energy in a controlled manner. In contrast, a chain reaction in which the number of

neutrons generated increases over time is said to be supercritical; an atomic bomb exemplifies a supercritical chain

reaction, releasing a vast amount of energy in a tiny fraction of a second. The energy released is expressed as yield. It is

typically measured in kilotons, where one kiloton equals the energy released by the explosion of 1,000 tons of TNT;

modern nuclear weapons typically have yields in the range of tens to hundreds of kilotons. In contrast, several types of

experiments producing little to no nuclear yield have been conducted over the years. Hydronuclear experiments were

conducted during the 1958-1961 nuclear test moratorium. They initially used less than a critical mass of fissile

material; as the amount of this material was stepped up toward criticality from one experiment to the next, some of

these experiments resulted in the release of tiny amounts of energy from fission, even as little as a gram of TNT

equivalent or less. Hydrodynamic experiments implode a pit (the first stage or “trigger” of a nuclear weapon) in order

to examine how the pit behaves; these experiments use non-fissile material as a surrogate for fissile material, so they

cannot become critical. Subcritical experiments examine how plutonium behaves when subjected to a spike in pressure,

such as when struck by an explosive-driven metal plate. The plutonium is configured in a way, such as by its shape and

quantity, that it cannot go critical.

SFRC CTBT hearing, 1999, pp. 16-17.

Enclosure, in letter from Jeffrey T. Bergner, Assistant Secretary, Legislative Affairs, U.S. Department of State, to

The Honorable Jon Kyl, United States Senate, August 9, 2007.

ȱȱȱ

ŗŞȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

ȱȱ ȱȱȱȱǵȱ

ȱ¢ȱȱȱ

Because of concerns that states parties to the CTBT could cheat and thereby change the strategic

balance, the ability to monitor the treaty has always been an integral part of the debate over the

treaty. Monitoring has always been more difficult for underground nuclear tests than for tests in

other environments. Radioactive particles in the atmosphere (fallout) are readily detectable in

trace amounts. Sound waves in the oceans travel great distances. Tests in space can be detected by

national technical means. It is for this reason that the LTBT banned tests only in the atmosphere,

in space, and under water. Accordingly, much of this section focuses on detection, and evasion of

detection, of underground tests. This section presents a technical background and contending

views for several monitoring technologies.

The treaty contains complex provisions in an effort to monitor compliance with its basic

obligation of conducting no nuclear explosions. It establishes a Comprehensive Nuclear-Test-Ban

Treaty Organization (CTBTO) that would begin operation upon the treaty’s entry into force. Its

elements are a Conference of States Parties; an Executive Council to promote implementation of,

and compliance with, the treaty; and a Technical Secretariat for monitoring. The secretariat is

deploying an International Monitoring System (IMS) to detect nuclear tests;79 an International

Data Center (IDC) to analyze data and disseminate the results to member states; and a Global

Communications Infrastructure to transmit data to, and reports from, the IDC. The treaty provides

for on-site inspections (OSIs) if 30 of the 51 Executive Council members approve. In 1996, the

signatory states established a Preparatory Commission for the CTBTO to implement the

organization, the IMS, and the IDC, and to prepare for OSIs, so that the CTBTO would be fully

operational upon the treaty’s entry into force.

The treaty calls for the IMS to have 321 stations worldwide to monitor signals that might indicate

a nuclear explosion: 170 seismic stations to monitor seismic waves in the Earth; 11 hydroacoustic

stations to monitor underwater sound waves; 60 arrays of infrasound detectors to monitor very

low frequency sound waves in the atmosphere; and 80 radionuclide stations to detect radioactive

particles that a nuclear explosion might produce; as well as 16 radionuclide laboratories to

analyze radioactive samples. Of the seismic stations, 50 are to be primary stations to provide data

to IDC continuously and in real time, while 120 are to be auxiliary stations to provide data when

requested by the IDC. As of November 26, 2007, 37 primary seismic stations, 76 auxiliary

seismic stations, 10 hydroacoustic stations, 37 infrasound arrays, 47 radionuclide stations, and 9

radionuclide laboratories had been certified. That is, they are completed and meet the technical

requirements of the Preparatory Commission. They transmit data automatically and continuously

to the IDC, excepting for the auxiliary stations and the radionuclide laboratories, which transmit

data as requested by the IDC.80

The United States has operated its own system to detect nuclear tests since the 1940s. The present

system, the U.S. Atomic Energy Detection System (USAEDS), is operated by the Air Force

Technical Applications Center (AFTAC). AFTAC states that USAEDS is a “global network of

For a map of IMS stations, at http://www.ctbto.org see Verification Regime > Monitoring Facilities > Map of

Facilities.

Information provided by Annika Thunborg, Chief, Public Information, Comprehensive Nuclear-Test-Ban Treaty

Preparatory Commission, personal communication, November 26, 2007.

ȱȱȱ

ŗşȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

nuclear event detection sensors” including underground, underwater, atmospheric, and space

sensors.81 NNSA provides technical support for satellite- and ground-based nuclear explosion

monitoring. Other organizations are conducting research on nuclear explosion monitoring as

well.82 While 21 USAEDS seismic stations were part of IMS as of August 200783 (i.e., they

provide data to IDC), USAEDS also has other capabilities, such as detectors on satellites, that are

not part of IMS. USAEDS and IMS are to some extent complementary. USAEDS, as a national

system, focuses on areas of concern to the United States; IMS, as an entity of an international

treaty, maintains a worldwide detection network so no nation feels singled out for special

monitoring attention. IMS makes available to all states signatories, including the United States,

data from its network; some data are from sites that the United States could not access. Further,

IMS data may be more credible to some of those nations than data from USAEDS. The former

come from a transparent, internationally-controlled system, while USAEDS data might be less

convincing to Executive Council members if they suspected that the United States was releasing

information selectively or if the sensors and resulting data were unfamiliar and thus difficult for

some council members to interpret. As the State Department said, “In the case of the DPRK

[North Korean] test, several countries have noted that the combination of IMS and IDC data and

analysis with U.S. national data and analysis provided them with greater confidence in assessing

the event than would have been the case with the U.S. data and analysis alone.”84 In addition to

IMS and USAEDS, academic institutions and national governments operate thousands of other

seismic stations worldwide.85 Some of these stations may feed information to IDC on an ad hoc

basis.

There is general agreement that IMS will be able to detect most nonevasive tests at 1 kiloton or

less. C. Paul Robinson, then director of Sandia National Laboratories, said in 1999, “The

detection threshold that was used informally by treaty negotiators as an unofficial target for the

IMS was about 1 kiloton, non-evasively tested, in environments other than outer space. Although

IMS coverage will not be uniform over the entire globe, it is expected to generally achieve that

informal target.”86 A National Academy of Sciences report places the threshold for nonevasive

underground tests at “significantly better than 1 kt [kiloton]” and says, “For most of Europe, Asia,

and Northern Africa, the detection threshold is down in the range from 30 to 60 tons [i.e., 0.03 to

0.06 kilotons] in hard rock.”87 The detection of the 2006 North Korean nuclear test, with a yield

the United States placed at less than a kiloton,88 by IMS and non-IMS stations supports the claim

of a low detection threshold for nonevasive underground tests.

U.S. Air Force. Intelligence, Surveillance, and Response Agency. “Fact Sheet: Air Force Technical Applications

Center.” June 2007. http://www.afisr.af.mil/library/factsheets/factsheet.asp?id=10309.

These organizations include the Air Force Research Laboratory, the Army’s Space and Missile Command, the Office

of Naval Research, the Special Geology Program of the U.S. Geologic Survey, and the U.K. Atomic Weapons

Establishment’s Forensic Seismology Group. In addition, other organizations are conducting research relevant to

nuclear explosion monitoring. Information provided by Bureau of Verification, Compliance, and Implementation, U.S.

Department of State, personal communications, February 1 and 4, 2008.

Letter from Jeffrey T. Bergner, Assistant Secretary, Legislative Affairs, U.S. Department of State, to The Honorable

Jon Kyl, United States Senate, August 9, 2007, enclosure, answer to question 2.

Ibid., answer to question 9.

See, for example, Incorporated Research Institutions for Seismology, “Stations & Instrumentation,” available at

http://www.iris.edu/stations/.

SASC CTBT hearings, 1999, p. 131.

NAS report, p. 42.

U.S. Office of the Director of National Intelligence. Public Affairs Office. “Statement by the Office of the Director

of National Intelligence on the North Korea Nuclear Test,” ODNI news release no. 19-06, October 16, 2006.

ȱȱȱ

ŘŖȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

ȱ¢ȱ

Seismology has been used for decades to detect and differentiate between earthquakes and

explosions, though it is very difficult for seismology to differentiate between conventional and

low-yield nuclear explosions. Earthquakes and explosions generate many types of seismic waves

that propagate through the Earth. Various techniques are used to obtain more information from

these waves. For example, seismic arrays are typically groups of 5 to 30 seismometers spread out

over several square kilometers linked to a central point.89 Because of the distance between

seismometers, seismic waves from an event arrive at each seismometer at slightly different times.

These differences can be used to calculate the direction from which the waves arrived. This

technique has been in use for decades.

Other techniques also help extract information. Some seismic waves are teleseismic, detected

even at distances over 9,000 km.90 For example, an IMS station in South America detected

seismic waves from the 2006 North Korean nuclear test.91 Some teleseismic waves travel along

the Earth’s surface, while others travel through the interior. Of the latter, some are shear waves;

an earthquake generates them strongly as the two sides of a fault slide past each other. Others are

pressure waves; an explosion generates them strongly as the pressure of an explosion radiates

outward. The appearance of shear and pressure waves on a seismogram differs, giving a clue

whether an event is an earthquake or explosion. Another difference is that the first waves from an

explosion arrive suddenly, while those from an earthquake build up over a short time. More

recently, regional seismic waves have come into use to differentiate between earthquakes and

explosions. These waves are generally observed at distances of up to 2,000 km; they can often be

detected even when teleseismic waves from an event cannot be.

The direction from which seismic waves from an event arrive at multiple seismic stations around

the world can be used to determine the approximate location of the event. The magnitude of

seismic waves can also be used to calculate the yield of an explosion, though with considerable

uncertainty. The CTBT limits the area of an OSI to 1000 sq. km,92 and the CTBTO Preparatory

Commission stated that in the case of the North Korean nuclear test, “analysis of all available

data allowed for the identification of a potential inspection area of considerably less than 1000

square kilometers” despite the low yield of the explosion.93

Contending views. CTBT critics point to “decoupling” as a method of evading seismic

detection. It dates from the late 1950s.94 This technique involves setting off a blast in an

underground cavity large enough to absorb the force of the blast elastically, thus muffling the

resulting seismic signal. Critics point to a 1966 decoupling experiment conducted in a salt dome

in Mississippi in which a 0.38 kiloton explosion generated a seismic signal that appeared to be

NAS report, p. 40.

NAS report, p. 39.

“The CTBT Verification Regime Put to the Test—The Event in the DPRK on 9 October 2006,” Comprehensive

Nuclear-Test-Ban Treaty Preparatory Commission, 2007, available at http://www.ctbto.org/press_centre/

featured_articles/2007/2007_0409_dprk.htm.

Protocol to the Treaty, Part II, On-Site Inspections, Section A, General Provisions, Paragraph 3.

CTBTO Preparatory Commission, “The CTBT Verification Regime Put to the Test.”

See Jacobson and Stein, Diplomats, Scientists, and Politicians, pp. 151-154.

ȱȱȱ

Řŗȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

from an explosion one-seventieth as large.95 Larry Turnbull of the Central Intelligence Agency

said,

In judging whether this evasion scenario is credible, both the feasibility of constructing a

large cavity and of containing the debris from the nuclear explosions must be examined ...

construction of large cavities in both hard rock and salt is feasible, with costs that would be

relatively small compared to effort to produce the material for a nuclear device ... containing

both particulate and gaseous debris is feasible in salt, and more difficult—though not

impossible—in hard rock. Therefore, we judge that the cavity decoupling evasion scenario to

be credible and should be factored into any underground CTB monitoring.96

CTBT supporters respond that while decoupling works for very low yield explosions, it is much

harder for larger ones. The National Academy of Sciences (NAS) report raised ten difficulties in

conducting a decoupled test, such as constructing a cavity clandestinely, predicting the signals

from the test, ensuring that the yield of the device is not greater than planned, and containing

radionuclides. It finds, “Accepting the possibility of a cavity decoupled test, we conclude that

such an underground nuclear explosion cannot be reliably hidden if its yield is larger than 1 or 2

kilotons.”97

CTBT critics believe that decoupling could be concealed. Kathleen Bailey, former Assistant

Director for Nuclear and Weapons Control, Arms Control and Disarmament Agency, and Robert

Barker, former Assistant to the Secretary of Defense for Atomic Energy, reject the claim that the

earth and rock removed to create a cavity would be an indicator of decoupling: “In India, where

the very test site used had been closely observed, no such activity was detected prior to a nuclear

test.”98 CTBT advocates respond that this example is not a valid indicator of U.S. capability to

detect the excavation for decoupling because the test was not decoupled, and a decoupled test

would require excavation of far more material. For example, a cavity 37 meters in radius would

be needed to decouple a 3-kiloton device, with a volume of 212,175 cubic meters. In contrast, a

shaft 10 feet in diameter and 600 feet deep, possible dimensions for a non-decoupled 3-kiloton

test, has a volume of 1,327 cubic meters.99 CTBT opponents reply that excavated material may

not be observed by satellites if someone wants to hide the fact that digging is occurring. Material

could be removed when satellites are not overhead, or it could be moved underground in existing

tunnels. Aqueous excavation could be used to create large cavities in salt domes. In particular,

according to the State Department, “Iran presents particular challenges from a seismic detection

perspective. Iran’s vast numbers of salt domes offer an effective decoupling environment, making

detection particularly difficult in the absence of close-in sensors.”100

NAS report, p. 46.

Larry Turnbull, Central Intelligence Agency, “U.S. Monitoring Goals for the Comprehensive Test Ban Treaty,”

address to the Council on Foreign Relations, March 16, 1998, in SASC CTBT hearings, 1999, p. 204; see also p. 200.

NAS report, pp. 47-48.

Bailey and Barker, “Why the United States Should Unsign the Comprehensive Test Ban Treaty and Resume Nuclear

Testing,” p. 135.

Source for radius of a spherical cavity for full decoupling: “The Soviet Union carried out a partially decoupled test of

about 8 to 10 kt in 1976, in a cavity (in salt) of mean radius 37 m (sufficient to fully decouple about 3 kt).” NAS report,

p. 46. Source for dimensions of a shaft: According to one report, “[underground] tests are conducted in vertical drill

holes up to 10 feet in diameter and from 600 ft to more than 1 mile deep.” U.S. Congress. Office of Technology

Assessment. The Containment of Underground Nuclear Explosions. OTA-ISC-414, October 1989, p. 16. Note that the

diameter of the shaft depends on the drilling equipment used, not on the yield of the device.

100

Letter from Jeffrey T. Bergner, Assistant Secretary, Legislative Affairs, U.S. Department of State, to The Honorable

Jon Kyl, United States Senate, August 9, 2007, enclosure, answer to question 9.

ȱȱȱ

ŘŘȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

Supporters of the treaty point to numerous advances in seismological capability that would help

monitor the CTBT. Foremost is the ongoing rollout of the IMS; many of its seismic (and other)

stations around the globe provide data to IDC in real time. As the IMS is an international system,

many of its stations are in areas that the United States could not access, such as in Iran. Further, it

is important that the seismic stations will contribute regional as well as teleseismic data because

regional data is of particular value in detecting low-yield tests and decoupling. One source states,

“Regional waves enhance the ability to detect cavity decoupling because higher frequency waves

are more observable at regional distances and decoupling is smaller at higher frequencies ...

compared to teleseismic waves ...”101 Regional stations have proven more valuable than was

expected; according to U.K. seismologists,

When the IMS was negotiated, the rationale for auxiliary seismic stations [those that provide

data only when interrogated, not on a continuous basis, to the International Data Center] was

that these stations would improve the ability of the IMS to locate seismic events, and to more

finely characterize the seismic source. With the ongoing deployment of the IMS,

seismologists have discovered that the auxiliary stations are of particular value for

identifying the source of a seismic signal as an earthquake or explosion because they pick up

certain seismic waves that can be used in identification. In addition, it has turned out that

having many seismic stations, such as those in individual national or university networks,

complements the IMS stations and increases the availability of data.102

CTBT supporters note that other signatures in addition to characteristics of seismic waves help

differentiate between earthquakes and explosions. Finding that the epicenter of an event is more

than 10 km deep rules out an explosion, as does finding the epicenter at sea in the absence of

hydroacoustic waves indicative of an explosion. Other characteristics specific to local geology aid

determining whether an event is an earthquake or explosion. The CTBTO Preparatory

Commission states that IMS stations around the world detected the North Korean nuclear test of

2006, and IMS was able to locate the test to well under 1000 square km. As another indicator, the

seismic record shows a clear difference between that explosion and an earlier earthquake.

According to seismologists Paul Richards and Won-Young Kim,

The seismogram of 9 October [2006, the North Korean test] has three important features.

First, it shows an impulsive onset of compressional waves ... characteristic of an explosion.

Second, peaks indicative of shear waves in the [Earth’s] crust, which would be typical of an

earthquake, are very weak ... And third, short-period ‘Rayleigh waves’ are apparent. They ...

are known to be excited only by sources at a depth not much more than about 3 or 4 km,

which is much shallower than typical earthquakes.103

Critics point to evasive tactics and weaknesses in seismic monitoring that open prospects for

clandestine testing. Yield can be calculated from the magnitude of seismic waves. Yet many

factors affect the intensity of seismic signals in addition to the yield of a nuclear device. The NAS

report states, “[regional] waves are dependent on local properties of the Earth’s crust and

uppermost mantle—which can vary strongly from one region to another.”104 For example, a

device detonated in soft rock can have ten or more times the yield as one detonated tamped (fully

101

Hafemeister, “Progress in CTBT Monitoring Since Its 1999 Senate Defeat,” p. 160.

Information provided by seismologists at the U.K. Atomic Weapons Establishment, personal communication,

October 11, 2007.

103

Paul Richards and Won-Young Kim, “Seismic Signature,” Nature Physics, January 2007, p. 5. For a seismogram of

the North Korean test, see CTBTO Preparatory Commission. “The CTBT Verification Regime Put to the Test.”

104

NAS report, p. 39.

102

ȱȱȱ

Řřȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

coupled) in hard rock, yet the seismic signals from each can indicate the same apparent yield

because soft rock transmits seismic energy much less efficiently than does hard rock.105 An

evader, knowing this from the unclassified literature, would consider this difference in selecting a

test site. While CTBT supporters note that regional seismic signals can aid in detecting loweryield nuclear detonations, opponents reply that Russia and China did not permit IMS stations to

be located within hundreds of kilometers of their nuclear test sites, at Novaya Zemlya and Lop

Nor, respectively. The closest IMS station is 1,112 km from Novaya Zemlya, and 783 km from

Lop Nor. In contrast, the three IMS stations closest to the Nevada Test Site (NTS) are at distances

of 249, 380, and 417 km.106 The State Department observes,

There is no doubt that we would be better off if we had close-in seismographs around Lop

Nor and Novaya Zemlya. If IMS were allowed to install three seismographs surrounding Lop

Nor at the distances similar to those surrounding the NTS, it would be much easier not only

to detect smaller events, but also to identify the nature of smaller events and to determine a

better location as well as the origin time.107

Iran has numerous salt domes many hundred of miles from the IMS station near Teheran. Critics

argue that Iran could easily create cavities for decoupling by using water to dissolve salt. It has

extensive experience in drilling for oil, which is often found near salt deposits. As such, it is

argued, it is well equipped to excavate cavities for decoupling. Further, much of Iran is

seismically active, making it easier for Iran to conduct a test during an earthquake to mask the

explosion’s signals. Others respond that hiding a test in an earthquake requires holding the test in

readiness, possibly for years, for the “right” earthquake to come along, and it may still be possible

to distinguish signals from an earthquake from those of an explosion.

Other techniques can also reduce seismic signals from underground nuclear tests. Don Linger,

Senior Scientific Advisor, Advanced Systems Concepts Office, Defense Threat Reduction

Agency, and former director of the Defense Nuclear Agency’s nuclear effects testing program,

provided the following information.108

One technique for reducing seismic signals is “geologic preconditioning.” A nuclear test in

hard rock will fracture or microfracture the surrounding rock to distances of several hundred

meters, fragmenting it and changing the shock propagation and attenuation characteristics.

As a result, a test conducted underground in a hard rock geology region in which a previous

nuclear test was conducted will in effect be conducted in fragmented rock, which absorbs

much more energy than undisturbed rock, weakening the seismic signal. This attenuation

was observed in experiments using 100 tons of chemical explosive, conducted by the U.S.

Departments of Defense/Defense Nuclear Agency (now the Defense Threat Reduction

Agency) in a series of tests in Kazakhstan during the closing of the former Soviet Nuclear

Test site in 1993 to 2002. Moreover, the Russian test site at Novaya Zemlya, which is

comprised mainly of similar hard rock, has similar regions of preconditioned hard rock

created by previous tests that could be used to muffle seismic signals of clandestine tests.

This is a proven technology, clearly understood by the testing community.

105

NAS report, pp. 41-42.

U.S. Department of State. Bureau of Verification, Compliance, and Implementation. “Response to Medalia

Questions of 27 Nov [2007].” January 14, 2008. Hereinafter “Response to Medalia Questions.”

107

“Response to Medalia Questions.”

108

Information provided by personal interview and emails, December 3-13, 2007.

106

ȱȱȱ

ŘŚȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

A second technique to reduce seismic signals, “radiation spectrum tuning,” is to reduce the

radiation coupling of the nuclear device to the ground. The amount of energy that a nuclear

device deposits into the surrounding geology is very sensitive to specifics of its radiation

output spectrum, and strongly affects the manner in which the blast is coupled to the ground,

causing large changes in the ground shock and seismic signature. Radiation spectrum is

entirely different than yield. For a given test cavity, a 10-kiloton weapon with energy

concentrated in the thousand-electron-volt range will produce a significantly lower seismic

signal than a 10-kiloton weapon with electromagnetic energy concentrated in the tens-ofmillion-electron-volt range. Nuclear explosives have been designed with different energy

spectra. For example, the U.S. Plowshare program of nuclear explosives for peaceful

purposes, and the parallel Soviet program, developed nuclear explosive devices with energies

concentrated in a part of the electromagnetic spectrum different than that of typical nuclear

weapons.

CTBT proponents respond that geologic preconditioning may be of use to Russia or China, which

have a “stockpile” of cavities left by nuclear test explosions, and possibly to India and Pakistan,

which may have a few small cavities, but not to other nations. Opponents dismiss this argument

because they view the prospect of Russian or Chinese covert testing as the greatest threat.

Proponents, in turn, reply that the decoupling capability of geologic preconditioning would vary

greatly depending on specifics of the surrounding rock and the extent of its fracturing, which

would be extremely difficult to determine. Regarding radiation spectrum tuning, proponents ask if

modifications to the test device that would be needed to reduce the seismic signature would

interfere with the purpose of, and results from, the test so much as to diminish its value

significantly.

Seismic monitoring entails other arguments. Critics state that the ability to detect lower-yield tests

increases many-fold the number of seismic events that must be analyzed as possible nuclear tests.

Supporters reply that improved seismic detection and data analysis capability rule out most such

events as possible explosions, and that low-yield tests are of little military significance. Critics

respond that low-yield explosions have military significance, as discussed below, and that it

would be easier for IDC to miss a low-yield explosion among thousands of low-magnitude

earthquakes than to miss a higher-yield explosion. Supporters retort that the North Korean test of

October 2006 was clearly detected even though it had a yield of less than a kiloton; critics counter

that it was not conducted evasively.

ȱȱȱȱ

Nuclear explosions generate a great variety of radioactive atoms, or radionuclides, some of which

are gases. Of special interest are radioactive isotopes of noble gases, such as argon-37, krypton85, xenon-131, and xenon-133. The background level of these gases is extremely low. Because

noble gases are chemically inert, they do not bond with the rocks and soil surrounding an

underground nuclear explosion. As a result, they work their way to the surface and disperse into

the atmosphere, where they may be detected thousands of miles away. For example, the

Automated Radioxenon Sampler/Analyzer, in use by IMS, concentrates and measures minute

quantities of the isotopes of radioactive xenon.109 Once a detection system has accumulated a data

archive of background levels of radioactive noble gases, a spike above that level can indicate a

109

“Breakthrough Systems to Detect Nuclear Explosions Worldwide,” Pacific Northwest National Laboratory, press

release, July 24, 1998, at http://www.pnl.gov/news/1998/Bnw98_24.htm. See also Hafemeister, “Progress in CTBT

Monitoring Since Its 1999 Senate Defeat,” p. 168.

ȱȱȱ

Řśȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

release from a nuclear reactor or nuclear explosion. Computer models of global atmospheric

conditions in the days before a spike can then be worked backwards to provide a general location

of the source.

At entry into force of the CTBT, the IMS is to have 80 radionuclide stations around the world; all

are to monitor radioactive particles and upon the treaty’s entry into force 40 of them would have

capability to monitor radioactive noble gases. Sixteen laboratories would analyze samples from

these stations. The CTBTO PrepCom states: “The relative abundance of different radionuclides in

these [air] samples can distinguish between materials produced by a nuclear reactor and a nuclear

explosion.... The presence of noble gases can indicate if an underground explosion has taken

place.”110

Contending views. The treaty’s supporters claim that the 2006 North Korean nuclear test shows

the value of noble gas monitoring and the capability of the IMS. An IMS radionuclide system at

Yellowknife, Northwest Territories, Canada, collected samples two weeks after the test that, upon

analysis, indicated a trace amount of xenon-133. By comparing this amount to data in its archive,

analysts were able to determine that the level was elevated. By examining wind currents for the

preceding two weeks, and data on releases from the Chalk River Laboratories, a Canadian nuclear

research site several thousand kilometers southeast of Yellowknife, analysts were able to conclude

that the xenon-133 was “consistent with a release from the location and time of the DPRK

event.”111

Opponents see numerous ways to evade detection of radioactive noble gases. They recognize that

noble gases will reach the surface if there is no effort at containment, but believe containment can

work. They point to a statement by Donald Barr, a retired Los Alamos radiochemist with over 50

years of nuclear testing and related experience: “Deep burial of a nuclear device, combined with

gas blocking techniques, virtually eliminates the seepage of noble gases to the surface, though

some such gases might occasionally be detected, but only at the surface above the detonation

point.”112 Burying a nuclear test device at greater depth than would be typically used for

containment would also delay the time when these gases would reach the surface, providing more

time for radioactive decay to reduce the amount reaching the surface. Certain geologies, such as

salt domes, would more readily seal the cavity, blocking the escape of these gases.

CTBT supporters point to experimental data to buttress their claim that it is very difficult to

contain noble gases following an underground nuclear explosion because they rise to the surface

through faults or fractures, especially during periods of low barometric pressure.113 Opponents

would note that the experiment in question used surrogate gases (sulfur hexafluoride and helium3), not argon and xenon. Further, the report stated that the decay of argon-37 to chlorine-37 “will

limit the sampling ‘window’ during which surface detection is possible,” and that “selecting the

timing of a challenge inspection to include the arrival of weather fronts may be necessary to

110

Preparatory Commission for the Comprehensive Nuclear-Test-Ban Treaty, “Verification Technologies:

Radionuclide,” at http://www.ctbto.org, link to “Verification Regime” > “Monitoring Technologies” > “Radionuclide.”

111

P.R.J. Saey et al., “A Long Distance Measurement of Radioxenon in Yellowknife, Canada, in Late October 2006,”

Geophysical Research Letters, vol. 34, L20802, doi: 10.1029/2007FL030611, October 16, 2007, p. 5 of 5.

112

Personal communication, November 21, 2007.

113

C.R. Carrigan et al., “Trace Gas Emissions on Geological Faults as Indicators of Underground Nuclear Testing,”

Nature, vol. 382, August 8, 1996, pp. 528-531.

ȱȱȱ

ŘŜȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

optimize the possibility of detection.”114 An evader, knowing this, might try to delay inspections

beyond the time such a front is due to arrive.

ȱȱȱȱ

Underground nuclear explosions may vent radioactive particles (fallout) into the atmosphere,

where they may travel for thousands of miles, depending on wind, rain, particle size, and other

factors. Fallout analysis has provided a clear indication of a nuclear test for many decades. For

example, the United States learned of the first Soviet nuclear test (an atmospheric test) in 1949,

and learned much about the design of the first Soviet thermonuclear device in 1953, through

collection and analysis of these particles.115 The ease of detecting fallout particles was a main

reason why the United States, Soviet Union, and United Kingdom were able to negotiate the

LTBT in 1963, and worldwide protests against fallout were a main impetus for the treaty.

Contending views. CTBT supporters assert that containment of radioactive debris from a

nuclear test is difficult, and many techniques are learned through trial and error. Geologic

features, such as faults, can provide a path through which debris can vent. Certain types of soil or

rock are better for containment than others. Underground water, turned to steam by an explosion,

generates a great deal of pressure. Depth of burial must be adequate. Elaborate methods must be

used to prevent debris and gases from escaping through the shaft dug for the test.116 Despite

extensive experience with contained underground tests beginning in the 1950s, many U.S.

underground tests through 1970 released radioactive material.117 CTBT supporters therefore argue

that it would be difficult for Russia or China, and much more so for first-time testers, to have high

confidence that they could contain a clandestine test.

CTBT opponents respond that Russia and China would have high confidence in their ability to

contain a nuclear test because of their test experience. Opponents point to a U.S. example.

Following the “Baneberry” test of 1970, which vented a large radioactive cloud, the United States

took further steps to contain underground tests, and of the 386 post-Baneberry tests conducted at

the Nevada Test Site through 1992, only 2 resulted in accidental release of radioactivity detected

outside the test site.118 Even nations without nuclear test experience could learn much about

containment from the open literature, and could make containment more likely by burying the test

device more deeply, examining geologic characteristics in selecting a test site, and building a

large margin of error into containment techniques.

The treaty’s supporters point to data on Soviet nuclear tests at Russia’s only nuclear test site,

Novaya Zemlya in the Arctic Ocean, to show the difficulty of containment. Using the period

beginning in 1971 so as to be comparable to U.S. post-Baneberry tests, 30 underground tests were

conducted from 1971 to 1990, with data unclear for two. Of the other 28, 10 vented radioactive

gases offsite, another 7 vented such gases onsite only, 1 vented radioactive gases and debris

114

Ibid., p. 531.

Richard Rhodes, Dark Sun: The Making of the Hydrogen Bomb, New York, Simon and Schuster, 1995, pp. 370-372,

524.

116

For a detailed discussion of containment, see Office of Technology Assessment. The Containment of Underground

Nuclear Explosions, pp. 31-55.

117

U.S. Department of Energy. Nevada Operations Office. United States Nuclear Tests, July 1945 through September

1992, DOE/NV-209, rev. 15, December 2000, pp. 2-63.

118

Department of Energy, United States Nuclear Tests, pp. 64-88.

115

ȱȱȱ

Řŝȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

offsite, and 10 were contained. The treaty’s opponents counter that there was a sharp

improvement in containment. Of the 28 tests, for the period 1971 through August 1978, 10 of 16

tests vented offsite, 1 vented onsite, and 5 were contained; for September 1978 through 1990, 6

vented onsite only, 1 vented offsite (both gases and particles), and 5 were contained.119

ȱ¢ȱȱȱǻ Ǽȱ

This technique was developed in the early 1990s to study ground deformation around

earthquakes. In it, a satellite-borne radar sends out microwave radar beams to a swath of ground

some 100 km wide, and records, pixel by pixel, what is in effect the distance between the satellite

and each point on the ground.120 If another radar picture of the same terrain is taken later from

nearly the same point in space,121 one image can be digitally subtracted from the other, with any

difference shown as bands of color that reveal ground motion. According to the technical

literature, InSAR can detect ground deformation of less than 1 cm122 and can take pictures

through many types of clouds. Because it does not use visible light, it can take pictures night or

day. This technique has also been used to detect ground deformation due to oil and gas reservoirs

and to measure the stability of retaining walls around a reservoir in London.123

While IMS does not use satellite monitoring techniques, the CTBT (Article IV, section A,

paragraph 5) permits the use of national technical means. According to David Hafemeister,

professor emeritus of physics at California Polytechnic State University, “InSAR is now a widely

adopted technology, available to all CTBT States Parties at reasonable prices from commercial

vendors.”124 The depression formed by an underground nuclear test—assuming the rock or

ground above the test does not collapse into the cavity left by the test, leaving a clearly visible

crater—may be 1 to 2 km across and one to several cm deep.125

Contending views. CTBT advocates hold that InSAR complements other monitoring

techniques. It can monitor large areas for subsidence. It can localize a suspicious site, even with a

test of yield less than 1 kiloton (depending also on other factors such as depth of burial and

geology) to within 100 meters, thus helping to guide an OSI.126 It can discriminate between an

earthquake and an explosion based on changes in ground deformation revealed by InSAR; an

earthquake produces a more or less linear pattern caused by the two sides of a fault sliding past

each other, while an explosion produces a roughly circular depression. It can help find

119

Vitaly Kjalturin et al., “A Review of Nuclear Testing by the Soviet Union at Novaya Zemlya, 1955-1990,” Science

and Global Security, no. 13, 2005, pp. 40-42.

120

For brief descriptions of InSAR, see Gabriele Rennie, “Monitoring Earth’s Subsurface from Space,” S&TR, April

2005, pp. 4-11; and U.S. Department of the Interior. U.S. Geologic Survey. “Using Satellites to Monitor Deformation:

Radar Interferometry,” updated October 11, 2007, at http://volcanoes.usgs.gov/insar/more_insar.html.

121

The U.S. Geologic Survey states, “It isn’t possible to steer a satellite accurately enough to return it to exactly the

same point in space on different orbits, but it’s relatively easy to get within a few hundred feet and then do the

necessary geometric corrections.” Ibid.

122

Rennie, “Monitoring Earth’s Subsurface from Space,” p. 5.

123

On the latter point, see European Space Agency, “Groundmotion: Service Examples,” at http://www.eomd.esa.int/

booklets/booklet183.asp.

124

Hafemeister, “Progress in CTBT Monitoring Since Its 1999 Senate Defeat,” p. 169.

125

Paul Vincent et al., “New Signatures of Underground Nuclear Tests Revealed by Satellite Radar Interferometry,”

Geophysical Research Letters, vol. 30, no. 22, November 2003, p. SDE 1-1.

126

Hafemeister, “Progress in CTBT Monitoring Since Its 1999 Senate Defeat,” p. 171.

ȱȱȱ

ŘŞȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

construction of a decoupling cavity, as ground above the cavity may subside slightly. The wide

availability of InSAR data would arguably make a request for an OSI based on this data more

convincing to the CTBTO Executive Council.

Critics respond that InSAR requires before-and-after pictures of the same piece of ground in order

to detect slight subsidence. If only an “after” picture is available, the technique is thought to work

only for nuclear tests of 20 kilotons of yield or so, a level that seismic techniques can easily

locate, rendering InSAR superfluous. The State Department points to other limitations.

NASA, [Lawrence Livermore National Laboratory], Canadian Space Agency, and European

Space Agency all have InSAR systems and should have libraries of data covering much of

the world, at least up to middle latitudes. However, in some areas where there is rugged

terrain, terrain shadowing will likely cause large areas to be uncovered. Additionally, one

would need to have “before” images that are fairly recent to do an accurate comparison. If

significant changes have occurred in the terrain (other than those caused by the test) by wind,

rain or other natural factors, the “before” image will not be useful in constructing an InSAR

image. Furthermore, this is complicated by the fact that the subsidence may not occur until

some time after the test, perhaps years. So, whereas libraries do exist, without specific

tasking, they’re unlikely to be good enough.

Further, “It is particularly noteworthy that no evidence of subsidence was observed by the InSAR

technique after the North Korean test.” For these and other reasons, State concludes, “the

potential of InSAR in assisting detection of a nuclear explosion is limited and cannot be

considered a useful technique in many test scenarios.”127

Critics assert that subsidence could occur too late to aid an OSI. They also argue that some very

low yield tests, the kind an evader is most likely to attempt, conducted at Nevada Test Site did not

form depressions,128 and that deep burial and certain geologies (e.g., deep inside a granite

mountain) may preclude subsidence. Supporters reply that InSAR is of value if it helps deter

evasion, and that it may reduce the value and increase the difficulty of clandestine tests by forcing

a would-be evader to dig deeper and use smaller nuclear devices in order to avoid detection by

InSAR.

ȱȱȱ

A nuclear test requires much preparation. The testing nation must survey the site to determine if

the geology is suitable, bring drilling and diagnostic equipment to the site, drill the shaft, set up

the diagnostic equipment with its many cables, emplace the device, seal the shaft, and so on.

While IMS does not detect pre-test activities, national technical means of verification could.

Satellite photography and communications intercepts, CTBT supporters argue, can detect such

activities, and Article IV(D) of the treaty permits use of national technical data as well as IMS

data as grounds for requesting an inspection. CTBT opponents recognize that satellites might

detect preparations for a clandestine test, but argue that some activities may appear normal, such

as mining in a mining area, other activities may be hidden, land lines can prevent access to

communications, etc.

127

128

“Response to Medalia Questions.”

Hafemeister, “Progress in CTBT Monitoring Since Its 1999 Senate Defeat,” p. 171.

ȱȱȱ

Řşȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

Ȭȱȱǻ ǼǱȱȱȱ

The treaty and a protocol provide for OSIs, in which international inspectors would travel to the

site of a suspected nuclear explosion to search for conclusive evidence of such explosion. For

example, if the inspection team is able to drill into the cavity formed by a nuclear explosion, it

would have conclusive proof that a test occurred, and radiochemical analysis (such as the ratio of

different isotopes) could provide its approximate date. The treaty and protocol go into extensive

detail on OSIs, specifying procedures by which the Executive Council would authorize the start

and continuation of an inspection, the timeline for an inspection, the number of team members,

and equipment they may and may not use. These procedures represent a compromise between

those who wanted highly intrusive inspections that could be conducted quickly and those who

feared that such inspections would reveal military secrets.

Contending views. Much of the Senate debate on OSIs in 1999 involved the ease of securing

Executive Council permission for an OSI. According to Article II of the treaty, once a state party

has requested an OSI, 30 of 51 members of the Executive Council would have to approve to order

the inspection. Ambassador Jeane Kirkpatrick questioned the competence of the council to make

technical decisions related to the treaty. Each member of the council would have one vote. Since

the council would be based on geographic representation, many nations on it would have little or

no nuclear experience. Further, “there will be a technical support group ... chosen by the same

executive council ... which is chosen by people the overwhelming majority of whom do not

themselves have any experience or competence with nuclear questions, much less nuclear

weapons.”129 She also noted, “U.N. bodies are very highly political bodies.”130 Senator Richard

Shelby said that it would be hard to obtain the 30-vote supermajority needed for an OSI to go

forward,131 while Senator Joseph Biden provided an analysis of likely council voting and

concluded that “it seems to me pretty darned easy to get to 30 votes, not because 30 nations love

us, but because it is in their naked self-interest.”132

Another contentious topic is how the provisions of the treaty and its protocol specifying

procedures for OSIs might affect the success of inspections. Opponents assert that many of these

provisions impair the technical effectiveness of an inspection. Some such provisions are listed

here, along with a few comments made in 2007 by the State Department:

•

The protocol limits the inspection team to 40 members except when it is drilling,

and limits an inspection to 130 days. The State Department observes, “the

availability of acceptable, technically qualified and trained inspectors and

inspection assistants, operating as a cohesive team, is a factor affecting the

adequacy of the OSI timeline.”133

•

The treaty requires the team to submit a progress report within 25 days of the

council’s approval of the OSI; the inspection will continue unless a majority of

the council votes not to do so. But according to the State Department, “there is no

129

SFRC CTBT hearing, 1999, pp. 10-11.

SFRC CTBT hearing, 1999, p. 48.

131

SFRC CTBT hearing, 1999, p. 55.

132

SFRC CTBT hearing, 1999, p. 96.

133

“Response to Medalia Questions.”

130

ȱȱȱ

řŖȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

guarantee that the Executive Council will consider ‘progress’ (not defined) to be

sufficient to justify the OSI entering the continuation phase of the inspection.”134

•

The protocol permits specified inspection techniques but does not provide for the

adoption of new ones. This omission may become more significant as new

technologies emerge.

•

The protocol permits one overflight that may last at most 12 hours and may only

use field glasses, passive location-finding equipment, video cameras, and handheld still cameras, unless the state being inspected agrees to more overflights and

the use of other equipment. The State Department observes, “a State Party that

conducts a test will most likely employ all available means to evade initial

detection and, following approval of an OSI, restrict to the maximum extent the

use of technologies and techniques that might otherwise result in detection.”135

•

The inspected state has “[t]he right to make the final decision regarding any

access of the inspection team ...,” apparently referring to areas within the area to

be inspected that the inspected state deems sensitive. To protect them, the

inspected state may shroud sensitive equipment and restrict radionuclide

measurements and the taking of samples to those relevant to the inspection.

•

The inspection team may gain access to sensitive facilities if “the inspection team

demonstrates credibly to the inspected State Party that access to buildings and

other structures is necessary to fulfil the inspection mandate.” Opponents doubt

that the inspected state would agree that any such demonstration was credible.

The treaty’s supporters recognize that the inspection provisions represent a compromise between

the ability to find evidence of a clandestine test and the ability of inspected states to protect

sensitive facilities and guard against espionage. Supporters observe that these provisions protect

the United States as well as other nations. They note that many provisions of the Protocol

facilitate inspections.

134

135

•

Inspectors may inspect an area of 1,000 square kilometers; supporters argue that

this is large enough given the ability of monitoring technologies to limit the area

to be inspected.

•

Inspectors shall be chosen “on the basis of their expertise and experience”;

supporters note that other possible criteria, such as representing regional

groupings of states, were not used.

•

The protocol permits many technologies to be used, including visual observation,

video and still photography, multi-spectral imaging, measurement of

radioactivity, environmental sampling, passive seismological monitoring for

aftershocks.

•

Unless the Executive Council disapproves by a majority vote a request to

continue the inspection, it may also use active seismic surveys and magnetic and

gravitational field mapping.

•

If the council approves, inspectors may drill for samples.

“Response to Medalia Questions.”

ȱȱȱ

řŗȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

•

Subject to certain limitations, the inspection team has the right to collect, remove,

and analyze samples. Supporters note that a nuclear explosion would create many

forms of evidence, and that techniques for analysis of samples are highly

sensitive.

•

While the Executive Council may terminate an inspection after 25 days,

supporters of the treaty see that outcome as unlikely given that 30 of 51 members

of the council had to approve the inspection, and argue that the evidence needed

to gain approval by a supermajority would necessarily have been compelling.

Ǳȱȱȱ

While the 1999 debate considered procedural aspects of OSIs, it made little reference to their

technical aspects. Yet that issue has been raised for a half-century. For example, in 1960

testimony, a witness pointed to clues of value for an OSI. A nuclear explosion may produce very

different surface phenomena than an earthquake. If there are no signs of human activity in the

area, an explosion can be ruled out. There are dozens of signatures of a nuclear test, such as

disrupted vegetation, radioactivity, melted snow, pebbles in bushes, and road and fence

displacement. The witness pointed out difficulties as well. The most conclusive evidence of a

nuclear test is radioactive debris obtained by drilling into the radioactive zone left by a nuclear

explosion. Yet, he calculated, the radius of the radioactive zone of a 1.7-kiloton explosion is about

60 feet, and it would be necessary to drill 63 holes to have a 100 percent chance of finding this

zone in an area 500 feet in radius.136

Contending views. Technical capability to support OSIs has improved over the years. Satellite

imagery could reveal human activity. Seismologists have developed techniques to extract more

information from seismic data, helping to distinguish earthquakes from explosions and more

precisely locating the epicenter of an explosion. Radioactive isotopes of noble gases might be

discovered at the test site even if they were in such low concentration that they could not be

detected at a distance. InSAR could greatly narrow the search area.

It may, however, be difficult for an OSI to find the most conclusive proof of a clandestine test,

drilling into the cavity created by an underground explosion and retrieving radioactive debris. A

10-kiloton test would produce a cavity some 60 meters in diameter;137 depending on geology and

depth of burial; a lower-yield device would produce a smaller cavity. The test might or might not

result in a crater on the Earth’s surface. Such craters are caused when a cavity collapses and the

overburden above it collapses into the resulting void all the way up to the surface. Deeper burial

and careful attention to the geology of the test area would reduce but not eliminate the risk of

crater formation or of some signs of a test appearing at the surface. OSIs could encounter

practical problems. According to a prediction based on an experiment, xenon-133 and argon-37

“would be detectable, respectively, about 50 and 80 days after the detonation” for a 1-kiloton

136

“Statement of Richard M. Foose, Chief, Department of Earth Sciences, Stanford Research Institute,” in U.S.

Congress. Joint Committee on Atomic Energy. Special Subcommittee on Radiation and Subcommittee on Research and

Development. Technical Aspects of Detection and Inspection Controls of a Nuclear Weapons Test Ban. Hearings, 86th

Congress, 2nd Session, 1960, pp. 282-305.

137

Information provided by Lawrence Livermore National Laboratory, personal communication, August 31, 2007.

ȱȱȱ

řŘȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

explosion.138 By that time, an OSI might be completed. Livermore presents another problem with

detecting argon-37:

There is another “smoking gun” in lieu of drilling. That is argon-37. This is a noble gas

isotope produced by bombardment of calcium with neutrons. It gets formed during an

underground explosion, has a fairly long half life and is unique to an underground test (i.e.

the background is low to nonexistent). The only problem is that it is difficult to detect and

measure because you have to shield the sample from ambient background to a high degree

(i.e. put the sample in a lead-lined chamber of some kind to do the measurements). The

procedure discussed in OSI circles has been to take extensive air samples from surface

cracks at the suspected site, separate the noble gases from the air, remove the radon, and then

measure for argon-37. This would be difficult to do in the field.139

CTBT advocates claim that OSIs, by offering proof of a clandestine nuclear test, would act as a

deterrent. If a nation fears that it would get caught, the reasoning goes, it would be less likely to

conduct a nuclear test. Further, supporters argue, the deterrent effect would be magnified because

evaders would not know the thresholds at which various U.S. and international monitoring

capabilities could detect various test signatures, so they would have to compensate by deeper

burial, great efforts at containment, lower yield, and the like. Moreover, it is argued, evaders with

little or no test experience would have little confidence in their ability to predict yield or to

contain nuclear explosions, forcing them to take still more conservative measures to evade

detection. Such measures, it is argued, could make testing so difficult, costly, and risky as to be

not worthwhile.

CTBT critics respond that careful attention to evasion would defeat OSIs and would deter other

nations from requesting them. If a nation were not sure that it could locate a test with an OSI, or

even that a test had taken place, it would be reluctant to risk its credibility by requesting an OSI.

Further, in this view, while the U.S. monitoring system, USAEDS, may be able to detect faint

signatures that IMS cannot, the United States may be unwilling to use this evidence to make the

case for an OSI to avoid revealing capabilities. Thus an evader would not need to worry about the

maximum capability of USAEDS. At the same time, a prospective evader could learn the

capabilities of IMS because states parties to the CTBT receive IMS data. It could, for example,

conduct a large mining explosion and see how it registers with IMS. As a result, the treaty’s

opponents maintain, the prospect of OSIs would merely force an evader to pay close attention to

evasion techniques, something it would do anyway. Perversely, then, the CTBT’s provision for

OSIs would allow evaders to use the absence of a request for an OSI, or the conduct of an

unsuccessful OSI, as evidence that it was not evading.

CTBT critics challenge the validity of debating technical issues of monitoring, verification, and

evasion on an unclassified basis. Robert Monroe, former Director of Defense Nuclear Agency,

said:

Verification cannot be usefully addressed in unclassified documents. Verification is a twosided game. On the one hand, many of those around the world who are working to improve

verification are operating in an unclassified environment, and the arms control community

trumpets every advance in sensor locations, sensitivity, networks, etc. On the other hand, our

138

C.R. Carrigan et al., “Trace Gas Emissions on Geological Faults as Indicators of Underground Nuclear Testing,” p.

528.

139

Information provided by Lawrence Livermore National Laboratory, personal communication, August 31, 2007.

“Noble gases” are chemically inert; they include helium, neon, argon, krypton, xenon, and radon.

ȱȱȱ

řřȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

adversaries or potential adversaries who wish to develop or improve their nuclear weapons

while maintaining test deniability, are working with highest priority to improve their evasion

techniques. They are working in absolute secrecy, taking every precaution against being

detected. The only organization the U.S. has to counter them is the intelligence community,

and every scrap of its information collected on evasion improvements is highly classified.

Therefore an unclassified study will acquire a great deal of information on verification

improvements and almost nothing on evasion improvements. This could lead the unwary to

conclude that we are now able to verify a CTBT. My own impressions, based upon many

decades of close involvement with nuclear weapons, are exactly the opposite. I believe the

evaders have an easier problem to solve, that they are now in a comfort zone for undetected

testing, and that they expect their advantage to improve in the future.140

On the other hand, as Senator J. William Fulbright once said, “the mere fact that information has

been classified does not make it necessarily true.”141 Similarly, the treaty’s supporters would note,

the fact that information is unclassified does not make it invalid. Supporters argue that advances

in monitoring capability, many of which are unclassified, are likely to reveal clandestine testing

or preparations for it. Having unclassified information, such as from thousands of seismometers

around the world, publicly and promptly available increases the number of people who may find

evidence of testing. While technical monitoring cannot provide information that human

intelligence can on motivations, plans, and budgets, human intelligence can be misleading

because of disinformation, misinterpretation, reliance on unreliable sources, and incomplete

information. Basing conclusions on the absence of evidence it is argued, may be hazardous.

Former Secretary of Defense Donald Rumsfeld reportedly said, “the absence of evidence is not

evidence of absence.” 142 However, the absence of evidence cannot be construed as evidence.

Clearly, the Senate would consider classified information in any future debate on the treaty, but

classified details of evasion techniques would have to be balanced against classified monitoring

capabilities, and both would be only two of many elements of a net assessment.

ȱȱȱ

For decades, supporters of nuclear testing treaties have argued that monitoring capability is good

enough to permit effective verification, while critics have responded in part by setting forth

scenarios that, they maintained, would defeat verification. This report discussed one scenario,

decoupling, earlier and now turns to two others.

ȱ ȱȱ

One scenario envisions conducting one or more tests that would be detected but could not be

attributed. Robert Barker, former Assistant to the Secretary of Defense for Atomic Energy,

postulates a scenario that

involves conducting a test in a remote ocean area long after identifiable national vessels had

left the scene. The testing nation would expect the international monitoring system to detect

the test and announce the yield, and by virtue of its participation in the monitoring

140

Personal communication, November 5, 2007.

U.S. Congress. Senate. Committee on Foreign Relations. Subcommittee on International Organization and

Disarmament Affairs. Strategic and Foreign Policy Implications of ABM Systems. Hearings, 91st Congress, 1st Session,

1969, p. 183.

142

Walter Pincus, “Report Details Errors Before War,” Washington Post, September 9, 2006, p. 12.

141

ȱȱȱ

řŚȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

community the testing nation would have access to any debris collected, for its own analysis

of performance. It would be impossible to positively attribute the test to a nation if the

testing nation took care to ensure that materials were not used in the test such that debris

could be uniquely traced back to the testing nation. Indeed, a clever cheater would place

materials that are unique to different nations in close proximity to the bomb so that the debris

might look Israeli or Indian or even U.S.143

The National Academy of Sciences study stated,

Attribution is likely to be more problematic for an underwater or atmospheric test, since a

nation with a nuclear explosive could detonate it on a ship or a plane and the effects on the

surrounding media would be more ephemeral. Though such a test would likely be detected

and located, it might be attributed only with difficulty to the nation responsible. ... To

confidently evade attribution, a tester would need to believe that the United States, working

with other nations, did not have the capability to track ships and planes in the vicinity of the

test location, and would not intercept communications relating to the test.144

Arguments on this scenario can be played out at length. Donald Barr, a retired Los Alamos

radiochemist, states,

It is virtually impossible to disguise (spoof) the signatures of a nuclear explosive detonation.

This is because of the broad range of fission product and actinide radionuclides which are

produced instantaneously and then evolve with time according to well-known radioactive

decay laws. Any attempt to tamper with either or both of these distributions would produce a

discordance of the radiochemical data suite. The likely nature of such an attempted spoof

would become apparent through comparison of the radiochemical data with the extensive

data base of U.S. tests coupled with ever-improving model calculations of nuclear

explosives.145

Critics state that attribution depends on matching a sample of radioactive material with a sample

from an archive of such materials. If the sample does not match any in the archive, this method

provides no basis for attribution. Supporters counter that detection of debris from a nuclear test

would trigger an immediate, all-out effort by the United States, other nations, and the CTBTO to

attribute the test. The list of potential testers would be quite small, easing the task, and debris

could reveal information about weapon design, providing further clues as to the testing nation.

Supporters argue that a nation would probably need a test series to have confidence in a warhead

design, increasing the odds of attribution; opponents reply that one successful test might suffice

to confirm a simple implosion design, and an unsuccessful test might not be detected.

ȱȱȱ

While many signatures could reveal a test, it might be possible to conceal them all by conducting

a nuclear test in a large cavity excavated in a mining complex deep underground. A large cavity

would permit decoupling. Excavating the cavity deep underground, especially in rock, would

guard against a depression in the Earth’s surface that could be detected by standard or InSAR

satellite photography. Deep burial would arguably trap noble gases and particles; the open

143

Personal communication, April 28, 2007.

NAS report, pp. 38-39.

145

Personal communication, November 29, 2007.

144

ȱȱȱ

řśȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

literature has much information on how to contain underground explosions.146 Use of a mine

would provide a cover story for human activity and would hide much of that activity. Material

removed during excavation could be placed, unseen by satellites, in unused tunnels. Access to

IMS data, a right of all states signatory to the treaty, would help a would-be evader improve

evasion techniques and gather data on some types of evasive tests.

CTBT supporters see evasion as difficult. Containment, while harder for a nation with no test

experience, can fail nonetheless because of unknown aspects of test site geology, as the U.S.

“Baneberry” test showed.147 Satellite photography might reveal suspicious human activity. An

evader would not know capabilities of U.S. monitoring systems. Technical progress in monitoring

and a growing archive of background noise, it is argued, reduce the threshold below which an

evader could feel confident of success. An evader with little nuclear test experience would not

have a precise estimate of weapon yield, forcing it to lower the yield, and value, of a test. Human

intelligence might reveal a test. Supporters assert that the treaty would make evasion harder.

Secretary Albright argued that, while the United States cannot be absolutely certain to detect very

low yield tests with or without the treaty, “by improving our capacity to monitor, we are much

more likely under the treaty to detect such tests and consequently to deter them.”148

ȱȱȱȱ¢ȱǵȱ

A concern that arose in the 1999 CTBT debate was that clandestine testing could increase the

threat to the United States. As Senator John Warner said,

I am also concerned that the treaty’s zero yield test ban is not verifiable. It is difficult, if not

impossible, to detect tests below a certain level. If a nation is determined to conceal their

non-compliance with this treaty, there are certain levels below which we simply cannot

detect. The equipment is not there.

Testing at yields below detection levels may allow certain countries, such as Russia, to

develop a new class of nuclear weapons.149

Some argued then that undetected testing, even at low yield levels, would confer military

advantages. Six former Secretaries of Defense said, “it is impossible to verify a ban that extends

to very low yields.... Tests with yields below 1 kiloton can both go undetected and be militarily

useful to the testing state.”150 C. Paul Robinson, Director of Sandia National Laboratories, said, “I

believe that nuclear testing in the subkiloton range could have utility for certain types of nuclear

designs.”151 A 1995 report by the JASON defense advisory group noted the value of half-kiloton

tests: “For the U.S. stockpile, testing under a 500 ton yield limit would allow studies of boost gas

146

See, for example, Office of Technology Assessment, The Containment of Underground Nuclear Explosions.

“Baneberry,” a 10-kiloton test, spewed a radioactive cloud that was tracked to the Canadian border. See Office of

Technology Assessment, The Containment of Underground Nuclear Explosions, pp. 31-33; and U.S. Department of

Energy. National Nuclear Security Administration. Photo and description of Baneberry test, at http://www.nv.doe.gov/

library/photos/photodetails.aspx?ID=846.

148

SFRC CTBT hearing, 1999, p. 76.

149

SASC CTBT hearings, 1999, p. 5.

150

Letter from James Schlesinger, Richard Cheney, Frank Carlucci, Caspar Weinberger, Donald Rumsfeld, and Melvin

Laird to The Honorable Trent Lott, Majority Leader, United States Senate, and The Honorable Tom Daschle,

Democratic Leader, United States Senate, in SASC CTBT hearings, 1999, p. 58.

151

“Prepared Statement by Dr. C. Paul Robinson,” in SASC CTBT hearings, 1999, p. 132.

147

ȱȱȱ

řŜȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

ignition and initial burn, which is a critical step in achieving full primary design yield.”152 Bruce

Tarter, then Director of Lawrence Livermore National Laboratory, stated in 1997, “If additional

tests were to be allowed, then 500 tons would be the minimum nuclear test yield that would be of

value for validating experimental and computational tools used to assess weapon performance.

For purposes of helping to validate models for assessing weapon safety, nuclear test yields of a

few pounds would be of value.”153

CTBT opponents hold that low-yield weapons can have much more value for new or current

nuclear powers now than was the case decades ago, even within the 1 to 2 kilotons that the NAS

report uses as the upper limit on effective decoupling. Kathleen Bailey and Robert Barker write,

“One to two kilotons can be militarily and politically significant to any proliferator; with today’s

commercially available guidance technology one to two kilotons accurately delivered against a

major city or a major military installation will create massive damage. Today, proliferators don’t

need high-yield, thermonuclear weapons to threaten their neighbors.”154 John Foster writes,

Low yield underground tests of devices with yields of tons to hundreds of tons can provide

high confidence that such devices can be scaled up to strategic yields. Right now we have

little confidence that we could detect such low yield tests with high confidence if evasive

techniques were used. Such tests, if conducted by potential adversaries and not by the United

States, could adversely affect our overall security posture. For example, the US has provided

a nuclear umbrella to a number of its allies, such as South Korea, Japan, and Turkey, to deter

attacks by hostile nations and to reduce their need to develop their own nuclear capabilities.

However, recently a number of Russian sources have stated that Russia has developed and is

deploying low yield “clean” (that is, with reduced fission to reduce residual radiation)

nuclear weapons, including some “clean” earth penetrator weapons. Russian development of

clean weapons draws on the past Soviet development and demonstration of clean nuclear

devices for peaceful uses, similar to the U.S. “Plowshare” program of the 1960s and 1970s.

China may also be developing new low-yield weapons. In contrast, current U.S. nuclear

weapons, which date from the Cold War, are largely high yield, high fission, dirty weapons.

If a crisis were to develop between Russia and a U.S. ally, a Russian inventory of low yield

tactical nuclear weapons, and the asymmetry with U.S. weapons, could call into question the

credibility of the U.S. nuclear umbrella. Even without explicit threats, the asymmetry could

lead to nuclear nonproliferation by pressuring U.S. allies to develop their own nuclear

weapons.155

The treaty’s supporters reject the idea that low-yield weapons would make much difference to the

strategic balance, given the many nuclear weapons, of various yields, that this nation has. They

point to an article reporting on an interview with General James Cartwright, USMC, then

Commander of U.S. Strategic Command:

152

Sidney Drell, Chair, et al., Nuclear Testing: Summary and Conclusions, JASON report JSR-95-320, the MITRE

Corporation, August 3, 1995, p. 3.

153

“Response to Questions Regarding Comprehensive Test Ban Treaty (CTBT) for Senator Jon Kyl from C. Bruce

Tarter, Director, University of California, Lawrence Livermore National Laboratory,” September 29, 1997, in U.S.

Congress. Senate. Committee on Governmental Affairs. Subcommittee on International Security, Proliferation, and

Federal Services, Safety and Reliability of the U.S. Nuclear Deterrent, Senate hearing 105-267, 105th Congress, 1st

Session, 1997, p. 75.

154

Bailey and Barker, “Why the United States Should Unsign the Comprehensive Test Ban Treaty and Resume Nuclear

Testing,” p. 135.

155

Personal communication, November 2, 2007.

ȱȱȱ

řŝȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

Theoretically, if a “grave” threat to the United States emerged that could be deterred only by

a low-yield nuclear weapon, the general might be persuaded to support its development,

[Cartwright] said. However, to date, “I haven’t seen anything that approaches that,”

Cartwright said. ... “My priority is not reduced yield,” Cartwright told a reporter in April

2005. “It’s to take the accuracy to the point where conventional can substitute for nuclear.

That’s my first priority.”156

Various evasion scenarios, opponents argue, might be linked into a weapons development

program, with each step providing data and experience for the next step. Extremely low yield

tests could provide data on nuclear physics, nuclear testing, test containment, instrumentation,

and data retrieval; the data could be used to develop and validate computer models for weapons

design. Decoupled tests could provide data for design of an unboosted fission weapon, or perhaps

a boosted fission weapon.157 One or a few atmospheric tests conducted in a remote ocean area

might suffice to develop a higher-yield weapon while arguably avoiding attribution.158

Alternatively, the NAS report states, if a nation were given the design of a weapon, “A single fullyield test would validate both the legitimacy of a blueprint and success in reproducing the object,

but that test might be of yield too high to be concealed.”159

Supporters argue that nations could not develop thermonuclear weapons under a CTBT, and see

very low yield tests as of little value for weapons development. According to Richard Garwin,

hydronuclear tests “will provide little useful knowledge,” and tests of 0.1 kiloton “would have

little value in the development of nuclear weapons.”160 According to the NAS report, tests up to 1

to 2 kilotons are concealable in some circumstances, and could be used to improve unboosted

fission weapons or, with difficulty, for proof tests of weapons of 1 to 2 kilotons Tests up to 20

kilotons are unlikely to be concealable; they could be used to proof-test 20-kiloton fission

weapons, or for “eventual development & full testing of some primaries & low-yield

thermonuclear weapons.” Finally, tests above 20 kilotons could not be concealed; they could be

used to develop and test boosted fission weapons and thermonuclear weapons.161 Thus both the

value of tests and the risk of being caught are thought to increase with yield. At the same time,

there is general agreement that a nation could develop a simple gun-type or implosion weapon,

with a yield of perhaps 10 to 20 kilotons, without testing, thereby avoiding the need for evasion.

The NAS report observes that nations with more test experience could make more progress in a

weapons program through covert testing, but that “the threats these countries can pose to U.S.

156

Elaine Grossman, “Senior U.S. General Sees High Nuclear Threshold,” Global Security Newswire, October 22,

2007.

157

In “boosting,” a mixture of deuterium and tritium gases are injected into a hollow pit (typically made of plutonium).

When the pit is imploded, the heat and pressure cause the deuterium-tritium gas mixture to undergo fusion, thereby

releasing a great many neutrons that fission more plutonium. The significance, according to a Livermore report, is that

“[b]oosting increases the yield by a large amount.” R.E. Kidder, Maintaining the U.S. Stockpile of Nuclear Weapons

During a Low-Threshold or Comprehensive Test Ban, Lawrence Livermore National Laboratory, UCRL-53820,

October 1987, p. 5.

158

For example, disagreement remains even about whether an event in the South Atlantic in September 1979 that

registered a certain nuclear test signature on a U.S. satellite designed to detect nuclear explosions was caused by a

nuclear test or a meteoroid, let alone which nation conducted the possible test. See Jeffrey Richelson, ed., “The Vela

Incident: Nuclear Test or Meteoroid?”, National Security Archive, Electronic Briefing Book No. 190, May 5, 2006,

http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB190/index.htm.

159

NAS report, p. 66.

160

SFRC CTBT hearing, 1999, p. 117.

161

NAS report, p. 68.

ȱȱȱ

řŞȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

interests with the types of nuclear weapons they already have tested are large. What they could

achieve with the very limited nuclear testing they could plausibly conceal would not add

significantly to this.”162 CTBT supporters hold that the United States has lived with the prospect

that Russia or China could gain an advantage through clandestine testing since the U.S.

moratorium began in 1992. Russia, at least, has apparently taken a different approach to its

nuclear weapons program than has the United States, so that the programs are not strictly

comparable. For example, the NAS report states, “Russian nuclear weapons are remanufactured

on a 10-year cycle,”163 which contrasts with the current U.S. policy of extending the service lives

of existing warheads for many years. Nonetheless, CTBT supporters argue that the United States

has advanced in its nuclear capability significantly through SSP. In their view, it would be

instructive to ask the current directors of the three U.S. nuclear weapons laboratories if they

would rather be in the position of the Russian or Chinese nuclear weapons programs, even

including the possibility of testing at very low yields, or the U.S. enterprise with the scientific

tools made available by SSP but without testing. The NAS report summarizes the value of

clandestine testing as follows:

Very little of the benefit of a scrupulously observed CTBT regime would be lost in the case

of clandestine testing within the considerable constraints imposed by the available

monitoring capabilities.... The worst-case scenario under a no-CTBT regime poses far bigger

threats to U.S. security interests—sophisticated nuclear weapons in the hands of many more

adversaries—than the worst-case scenario of clandestine testing in a CTBT regime, within

the constraints posed by the monitoring system.164

ȱȱȱȱȱȱȱ ȱ ȱȱǵȱ

Any nation that ratified the CTBT, and then sought to cheat, would have to evaluate the risks and

benefits of clandestine testing. In the 1999 debate, attention focused on the feasibility of

successful cheating and the military gains that such tests might or might not confer, but virtually

no attention was paid to the risks of being caught. Nonetheless, the question merits consideration

because the answer could be crucial to a would-be evader’s calculus. Possible alternative cases

are sketched here; further research would be of use. It could be argued that there would be few

consequences. In this view, the Conference of States Parties to the CTBT, pursuant to Article V of

the treaty, “may recommend to States Parties collective measures which are in conformity with

international law.” Further, “[t]he Conference, or alternatively, if the case is urgent, the Executive

Council, may bring the issue, including relevant information and conclusions, to the attention of

the United Nations.” The U.N. might take little action, or it might delay. In particular, if evidence

of clandestine testing were not conclusive, there might be few or no penalties. Another possibility

is that the U.N., fearful that an unpunished violation could lead to the unraveling not only of the

CTBT, but also of U.S. willingness to take further steps toward nuclear disarmament, could

impose meaningful sanctions. Yet another possibility is that some nations could take actions

outside the U.N. framework.

Instead of attempting to conduct clandestine tests, a nation wishing to conduct one or more

nuclear tests might simply withdraw from the treaty. The case for so doing is that it might want to

conduct a test with a yield that could not be hidden; it might believe that even a low-yield test

162

NAS report, pp. 68, 77.

NAS report, p. 70.

164

NAS report, p. 78.

163

ȱȱȱ

řşȱ

ȱȬȬȱ¢Ǳȱ ȱȱȱ

could be detected, especially if it had little or no experience with nuclear testing and test

containment; and it might want to announce its nuclear capabilities to the world. But clandestine

testing offers advantages: an open weapons development program could spur rivals to launch

their own program, so a nation wanting to develop nuclear weapons might prefer to keep its intent

unknown; a nation that withdrew from the treaty would lose access to IMS data, which could help

it evade detection; a nation that withdrew from the treaty to conduct nuclear tests might face the

same penalties as one that conducted clandestine tests and was caught cheating; and a nation

might prefer to stay in good standing with the international community for as long as possible in

order to delay any sanctions.

ȱǰȱȱǰȱȱȱ

There is widespread agreement among experts within and outside the government that nuclear

proliferation, especially if it leads to terrorists obtaining nuclear weapons, is one of the greatest

security threats facing the United States.

The nuclear nonproliferation regime is a decades-long attempt to hold nuclear proliferation in

check. This regime is an array of treaties, agreements, nuclear weapon free zones, restrictions on

exports of nuclear-related equipment, controls of nuclear materials, and national laws, with the

Nuclear Nonproliferation Treaty (NPT) at its core.165 The NPT enter

This text is long and has been trimmed here. Open the source document for the complete record.

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.