ARTICLES
Attacking the Atom: Does
Bombing Nuclear Facilities Affect
Proliferation?
SARAH E. KREPS* AND MATTHEW FUHRMANN**
*Department of Government, Cornell University, USA, **Department of
Political Science, University of South Carolina and Council on
Foreign Relations, USA
ABSTRACT What are the consequences of military strikes against nuclear
facilities? In particular, do they ‘work’ by delaying the target states ability to
build the bomb? This article addresses these questions by conducting an analysis
of 16 attacks against nuclear facilities from 1942 to 2007. We analyze strikes
that occurred during peacetime and raids that took place in the context of an
ongoing interstate war. The findings indicate that strikes are neither as uniformly
fruitless as the skeptics would suggest, nor as productive as advocates have
claimed. There is evidence that the peacetime attacks delayed the target’s nuclear
program, although the size of this effect is rather modest. The wartime cases were
less successful, as attacks often missed their targets either due to operational
failure or limited intelligence on the location of critical targets. In our concluding
section we show that many of the conditions that were conducive to past success
are not present in the contemporary Iran case. Overall, our findings reveal an
interesting paradox. The historical cases that have successfully delayed
proliferation are those when the attacking state struck well before a nuclear
threat was imminent. Yet, this also happens to be when strikes are the least
legitimate under international law, meaning that attacking under these
conditions is most likely to elicit international censure.
KEY WORDS: Preventive war, Counterproliferation, Nuclear proliferation, Iran
Introduction
What are the consequences of military strikes against nuclear facilities?
In particular, do they ‘work’ by delaying the target states ability to
build the bomb? Policymakers in the United States, Israel, and even
The Journal of Strategic Studies
Vol. 34, No. 2, 161–187, April 2011
ISSN 0140-2390 Print/ISSN 1743-937X Online/11/020161-27 Ó 2011 Taylor & Francis
DOI: 10.1080/01402390.2011.559021
Downloadedby[89.177.143.45]at15:1025November2014
Arab countries such as Saudi Arabia1
have implied an affirmative
response, indicating that military force might frustrate Iran’s current
nuclear program. Yet, this perspective is at odds with concerns in the
scholarly literature that the use of force is ineffective and should be
avoided at all costs.2
If attacks against nuclear programs significantly
delay proliferation, then policymakers may be correct to keep military
force in the nonproliferation toolkit. On the other hand, if strikes
are ineffective tools of nonproliferation, then countries are wise to heed
the cautionary advice of scholars who urge them to pursue other
strategies.
This article speaks to the debate on the efficacy of military force by
analyzing the historical record of attacks against nuclear programs. We
identify four theoretical mechanisms for how strikes may affect nuclear
weapons’ production capacity. First, and most directly, attacks can
delay the target’s ability to build nuclear weapons by destroying
chokepoint facilities that are critical for bomb development. Strikes
may also delay the target’s program through three indirect mechanisms.
Raids could produce a change in the target’s fissile material production
strategy, make foreign suppliers less willing to provide nuclear
assistance, and lead to enhanced international inspections. These
outcomes can be thought of as externalities of strikes because they are
unrelated to the original aims of the attacker but can nevertheless
frustrate the target’s ability to proliferate.
To assess these mechanisms, we conduct a comprehensive analysis of
all 16 attacks against nuclear programs that have occurred from 1942
to 2007.3
We analyze strikes that occurred during peacetime as well as
those that took place in the context of an ongoing interstate war. ‘Bolt
from the blue’ attacks are qualitatively different from strikes against
nuclear facilities during wartime, but both types of cases are useful in
1
Ian Black and Simon Tisdall, ‘Saudi Arabia Urges US Attack on Iran to Stop Nuclear
Programme,’ The Guardian, 28 Nov. 2010.
2
See, for example, Dan Reiter, ‘Preventive Attacks against Nuclear Programs and the
‘‘Success’’ at Osiraq,’ Nonproliferation Review, 12 (July 2005), 355–371; Dan Reiter,
‘Preventive Attacks against Nuclear, Biological, and Chemical Weapons Programs: The
Track Record,’ in William Walton Keller and Gordon R. Mitchell (eds), Hitting First,
Preventive Force in US Security Strategy (Pittsburgh, PA: Pittsburgh UP 2006); Richard
K. Betts, ‘The Osirak Fallacy,’ The National Interest 83 (2006), 22–5; Bennett
Ramberg, ‘Preemption Paradox,’ Bulletin of Atomic Scientists, Jul./Aug. 2006, 48–56.
3
We define attacks against nuclear programs as the state-sanctioned use of force against
facilities related to a nuclear weapons program that has the intention of delaying a
country’s acquisition of nuclear bombs. For further details, see Matthew Fuhrmann and
Sarah E. Kreps, ‘Targeting Nuclear Programs in War and Peace: A Quantitative
Empirical Analysis, 1941–2000,’ Journal of Conflict Resolution 54/6 (2010), 831–59.
162 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
understanding this issue.4
Peacetime cases, such as Israel’s 1981 attack
against Iraq and its 2007 strike against Syria, are ideal for evaluating
both the direct and indirect mechanisms because we can isolate the
effects of limited raids from the broader effects of interstate conflict.
This is more difficult for wartime cases, including World War II, the
Iran–Iraq War, and the 1990–91 Persian Gulf War. For example, the
post-Persian Gulf War inspections regime influenced Iraq’s ability to
reconstitute its weapons program in the 1990s, but it is hard to know
whether strikes against nuclear infrastructure during the campaign
contributed to this outcome independent of Iraq’s military defeat. We
therefore study wartime cases to evaluate the direct mechanism, that is,
whether the strike removed past progress by destroying relevant nuclear
facilities.
As we show, the use of force did not significantly delay the target’s
nuclear weapons program in many of the wartime cases. Strikes failed
in large part because there was limited intelligence on the location of
targets. Further, targets were not always effectively destroyed even
when their location was known. On the other hand, the peacetime
attacks tended to delay the target’s nuclear program, providing some
support for both direct and indirect mechanisms. The size of this effect
was rather modest, however, since neither Iraq nor Syria was on the
verge of building nuclear weapons at the time of the raid.
Our findings challenge both sides of the debate on whether force
works and suggest that neither perspective is as clear cut as its
proponents would have us believe. The view that strikes ‘are generally
ineffective, costly, unnecessary, and potentially even counterproduc-
tive’5
downplays evidence of prior strikes that delayed the target state’s
nuclear program. The competing view that strikes might be a panacea
for international proliferation does not take into account the number of
instances in which attackers failed to destroy key nuclear facilities in
the target country. We offer a more nuanced picture; we show that
there have been instances of both success and failure and explain why
there is variation.
While we conclude that some cases bought time for the attacker, this
finding should be seen in a qualified light when it comes to predicting
the consequences of future events. In his study of why countries build
nuclear weapons, Scott Sagan aptly pointed out that ‘predicting the
4
As Reiter notes, ‘intrawar cases are important, both because they offer lessons . . . and
because intrawar preventive strikes against [nuclear, biological, or chemical] programs
are possible in the future’. See Reiter ‘Preventive Attacks against Nuclear, Biological,
and Chemical Weapons Programs,’ 29.
5
Dan Reiter, Preventive War and its Alternatives: The Lessons of History (Carlisle, PA:
Strategic Studies Institute 2006), 2.
Bombing Nuclear Facilities and Proliferation 163
Downloadedby[89.177.143.45]at15:1025November2014
future based on such an understanding of the past . . . [is] problematic,
since the conditions that produced the past proliferation outcomes may
themselves be subject to change.’6
As we show in the discussion of this
paper, the same is likely true for attacks on nuclear facilities. This has
important implications for contemporary debates on how to respond to
Iran and other proliferators.
This article proceeds in four parts. First, we outline four mechanisms
by which the use of military force could affect the target state’s nuclear
program. Second, we evaluate all four theoretical mechanisms by
analyzing the two Israeli peacetime raids. Third, we analyze the direct
mechanism by considering strikes undertaken in the context of
interstate war. The final section discusses the findings, assesses the
conditions under which strikes might be useful in delaying a
proliferator’s nuclear program, and evaluates the likely effects of
strikes against Iranian nuclear facilities.
How Could Attacks Affect Proliferators’ Weapons Programs?
We begin with the straightforward observation that the acquisition of
nuclear weapons requires both political willingness and technical
capacity. Security threats or being insulated from the global economy
often motivate states to pursue the bomb.7
Yet states cannot cross the
nuclear threshold without the requisite nuclear technology, materials,
and knowledge. A growing number of quantitative studies show that
supply-side considerations, particularly whether a country has the
requisite nuclear infrastructure, are salient in explaining who acquires
nuclear weapons.8
This indicates that political will is not a sufficient
condition for going nuclear – even if a country is determined to get the
bomb. The historical record supports this assertion. Although 22
countries have had nuclear weapons programs since 1942, only 10 have
6
Scott Sagan, ‘Why Do States Build Nuclear Weapons? Three Models in Search of the
Bomb,’ International Security 21/3 (1996/97), 85.
7
See, for example, Sagan ‘Why Do States Build Nuclear Weapons?’; Sonali Singh and
Christopher Way, ‘The Correlates of Nuclear Proliferation: A Quantitative Test,’
Journal of Conflict Resolution 48/6 (2004), 859–95; Dong-Joon Jo and Erik Gartzke,
‘The Determinants of Nuclear Weapons Proliferation,’ Journal of Conflict Resolution
51 (2007), 167–94; Etel Solingen, Nuclear Logics: Contrasting Paths in East Asia and
the Middle East (Princeton, NJ: Princeton UP 2007).
8
See especially Matthew Fuhrmann, ‘Spreading Temptation: Proliferation and Peaceful
Nuclear Cooperation Agreements,’ International Security 34/1 (2009), 7–41; Matthew
Fuhrmann, ‘Taking a Walk on the Supply Side: The Determinants of Civilian Nuclear
Cooperation,’ Journal of Conflict Resolution 53/2 (2009), 181–208; Matthew Kroenig,
‘Importing the Bomb: Sensitive Nuclear Assistance and Nuclear Proliferation,’ Journal
of Conflict Resolution 53/2 (2009), 161–180.
164 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
successfully produced the bomb.9
Iran, for instance, began a nuclear
weapons program in the 1980s but it has not yet acquired the bomb in
part because of technical challenges.
The most significant and technically difficult factor affecting a state’s
opportunity to develop nuclear weapons is its ability to produce
adequate quantities of fissile material. There are two paths countries
can take to acquire fissile nuclear material. One involves enrichment
technology designed to produce highly enriched uranium (HEU). A
second path toward fissile material production involves reactor-based
technology oriented toward plutonium production.10
This means that
key chokepoints in a nuclear weapons program are: (1) uranium
enrichment facilities; (2) plutonium reprocessing facilities; and (3)
reactors.11
Attacks can delay a target state’s ability to produce nuclear weapons
if they make it more difficult for it to possess these chokepoint facilities
and, in turn, produce fissile material. Below, we identify four possible
mechanisms that might produce this outcome. The first affects the
target’s past progress; the others limit its future potential to produce
fissile material.
The Direct Effects of Attacks against Nuclear Facilities
The most direct way that an attack can affect the target state’s nuclear
program is through the destruction of facilities crucial to weapons
development. An attack could delay the target’s nuclear ambitions if
any of the chokepoint facilities we identify above were destroyed. The
magnitude of this effect depends on how many of the target’s
chokepoint facilities are destroyed relative to those continuing to
operate. If the target possesses numerous chokepoint facilities and the
attack destroys all of them, the raid would have a comparatively large
effect on the nuclear program. A raid would have a more modest
impact if some chokepoint facilities are razed but others are left intact.
It is difficult to determine exactly how many years an attack could set
back a program in the event that chokepoint facilities are destroyed.
Such a calculation would depend on the types of facilities countries
9
Jo and Gartzke, ‘The Determinants of Nuclear Weapons Proliferation’.
10
See Gary Gardner, Nuclear Nonproliferation: A Primer (London: Lynne Rienner
1994), 16–17.
11
Some scholars do not classify reactors as sensitive nuclear technology. We include it
on our list since reactors produce plutonium in spent nuclear fuel that could be
reprocessed into bomb grade material. See M.D. Zentner et al., Nuclear Proliferation
Technology Trends Analysis, 2005 (Richland, WA: United States Department of
Energy, Pacific Northwest National Lab 2005).
Bombing Nuclear Facilities and Proliferation 165
Downloadedby[89.177.143.45]at15:1025November2014
possessed, how much progress they had made toward building the
bomb, and their level of indigenous knowledge. If a country does not
possess any chokepoint facilities prior to an attack, it would be hard to
classify a strike as ‘effective’ from a counterproliferation standpoint
even if it razed nuclear infrastructure. In the absence of chokepoint
facilities, the target would not have been able to produce fissile material
at any point in the near future in the absence of a strike. On the other
hand, if a country possesses numerous chokepoint facilities and the
attack destroys all of them, the raid has a comparatively large effect on
the nuclear program since the target would likely otherwise have been
able to produce fissile material for a bomb in the near future.12
In a best
case scenario where a strike razed all chokepoint facilities an attacking
country possessed, it could set the program back five to ten years, if we
assume that the target country possessed chokepoint facilities that were
near completion and continued its pursuit of the bomb at a rate similar
to what it did prior to the attack. This estimate is lower than the
amount of time it generally takes to construct chokepoint facilities13
because diminishing marginal costs enable countries to build a second
facility quicker. For example, it took India more than a decade to
develop its first uranium enrichment facility but it built the second such
plant in only five years.14
Conversely, attempts to hit chokepoints can fail. An obvious cause of
a failed strike would be poor intelligence.15
In other cases the attempt
could end in operational failure due to an accident or the attackers
coming under enemy fire. In the event that the attacker cannot locate or
destroy targets, attacks would obviously not delay the target’s nuclear
program and could actually accelerate it by increasing the state’s
willingness to build nuclear weapons. Failed attacks could also lead to
measures that make future strikes more difficult, for example by
distributing the chokepoints so that they cannot be hit in one strike.
The Indirect Effects of Attacks against Nuclear Facilities
The mechanism we described above is based on the notion that an
attack can directly delay a nuclear program by reversing past progress.
An attack could also impact a target’s program more indirectly by
affecting its future behavior in one of three ways.
12
For a similar discussion, see Reiter, ‘Preventive Attacks against Nuclear Programs.’
13
See M.D. Zentner et al., Nuclear Proliferation Technology Trends Analysis.
14
Ibid.
15
Reiter, ‘Preventive Attacks against Nuclear, Biological, and Chemical Weapons
Programs.’
166 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
Shift in the approach to fissile material production. A raid might alter
a target state’s priorities such that it values keeping its program covert
above all else. This, in turn, could cause a target country to alter its
approach to acquiring fissile material. The most likely such shift is from
plutonium production to uranium enrichment. Proliferators pursuing
the plutonium route may choose to focus on the uranium path
following an attack because it is comparatively more difficult to conceal
reactors and reprocessing facilities due to their sheer size. Target states
might also perceive that some enrichment plants may be more difficult
to keep covert than others. For instance, electromagnetic isotope
separation (EMIS) facilities might arouse less suspicion because the
technologies involved are less tightly controlled. Consequently, they
might abandon plans to develop centrifuge or gaseous diffusion plants
after an attack in favor of an EMIS plant that relies on less conspicuous
technology.
While a target state may correctly perceive that changing its
approach to acquiring fissile material provides greater secrecy, this
shift can also delay its nuclear program if it chooses to pursue a
technology with which it has little experience.16
Under such circumstances,
the target would need to develop indigenous knowledge and
procure or develop new technologies. This would take comparatively
more time because it would not benefit from the favorable effects of
learning. For example, rebuilding a reactor might take less than three
years but building a centrifuge enrichment facility without having
previously done so could take at least 14 years.17
Moreover, there is no
guarantee that the target could successfully develop this facility. Of the
18 countries that have attempted to enrich uranium using the centrifuge
method since the 1940s, only seven (39%) have successfully done
so.18
This indicates that a program could be delayed even further if the
target chose to pursue a technology that was easier to conceal but
inefficient or difficult to master.
Reduction in willingness of foreign suppliers to provide assistance. The
use of force is typically an instrument of last resort because it is
potentially risky and expensive. Military force, therefore, represents a
costly signal that the attacking country is committed to ending or
delaying the target’s nuclear program. This might make third parties
less inclined to supply nuclear technology, materials, or know-how to
the suspected proliferating state for two reasons. First, there are
16
Jeremy Tamsett, ‘The Israeli Bombing of Osiraq Reconsidered: Successful Counterproliferation?’
The Nonproliferation Review, 11/3 (2004), 70–85.
17
Zentner et al., Nuclear Proliferation Technology Trends Analysis.
18
Ibid.
Bombing Nuclear Facilities and Proliferation 167
Downloadedby[89.177.143.45]at15:1025November2014
practical risks associated with constructing facilities that could be
targeted. To build a nuclear facility such as a reactor, personnel from
the supplier country would have to spend years on the ground in the
recipient country.19
Many of these personnel could be killed if the facility
they were constructing was attacked again. The prospect of military
force against the same program might discourage states from engaging in
the perilous business of nuclear supply.
Second, the use of force reveals information about the proliferating
state. It signals that at least one state (i.e., the attacker) had reason to
believe that the target was using, or planning to use, nuclear
infrastructure not to develop energy, but to develop the bomb. In
exposing these dangers, attacks affect the way that third parties –
especially nuclear supplier countries – view the target’s development of
nuclear facilities. Helping a country acquire the bomb could increase
the risk of nuclear war, instigate regional instability, raise the
possibility of nonstate actors getting their hands on nuclear weapons,
and reduce the supplier’s ability to exert influence against the target
state.20
Supplying to a suspected proliferating state could also damage
the supplier’s relations with the attacking state and other powerful
states that champion nonproliferation. Each of these outcomes would
harm the supplier’s interests and create incentives to discontinue its
nuclear commerce.
An inability to obtain foreign assistance would have serious
consequences because of its contribution to a target state’s nuclear
program. Foreign assistance is typically supplied exclusively for
peaceful purposes, but dual-use technology can also be used to build
nuclear weapons.21
Additionally, nuclear assistance helps establish
an indigenous infrastructure that can be drawn on to build facilities
dedicated to a military program. For these reasons, nuclear aid lowers
important barriers to proliferation, whereas the withdrawal of such
assistance would increase the time necessary to develop a nuclear
weapon.
Enhanced international inspections and safeguards. The nuclear
Nonproliferation Treaty (NPT), which entered into force in 1970,
entitles all non-nuclear-weapon states to nuclear technology for
peaceful purposes on the condition that they accept a system of
19
See, for example, Duane Bratt, The Politics of CANDU Exports (Toronto: U of
Toronto P 2006).
20
See, for example, Scott Sagan and Kenneth Waltz, The Spread of Nuclear Weapons:
A Debate Renewed (New York: W.W. Norton 2002); Fuhrmann and Kreps, ‘Targeting
Nuclear Programs in War and Peace.’
21
Fuhrmann, ‘Spreading Temptation.’
168 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
safeguards imposed by the International Atomic Energy Agency
(IAEA). This safeguards system – and the nuclear nonproliferation
regime more generally – is based on the notion that countries can have a
peaceful nuclear infrastructure that does not contribute to a weapons
program if certain technical and legal restraints are imposed. Although
IAEA safeguards do not guarantee that a proliferator will not use
nuclear technology for military purposes, rigorous inspections can
make such diversions more difficult. It was IAEA inspections that
detected irregularities in North Korea in 1992, for instance.22
Inspections, though not flawless, can help clarify the intentions of a
nuclear program and add a level of scrutiny that may make it more
difficult for a state to produce fissile material for bombs.
The use of force is not the only way to trigger enhanced international
inspections, but there are two reasons to expect that it could lead to
that outcome. First, the IAEA, with support from member countries, is
likely to seek a greater presence in a country that has been attacked to
counter the perception that it is incapable of fulfilling its mandate and
to decrease the likelihood that additional attacks will occur. Second,
the targeted country might encourage the IAEA’s presence in order to
demonstrate to the international community that its intentions are
peaceful.
Peacetime Case Studies
Israeli Attacks Against Iraq’s Nuclear Program, 1981
Beginning in the 1970s, Israel pursued a series of covert and later overt
actions designed to delay the Iraqi nuclear program.23
The Iran–Iraq
war provided an opportunity for it to escalate its opposition to
Baghdad’s bomb campaign. Using eight Israeli F-16s flanked by eight F-
15s for cover, the Israeli Air Force raided the Osirak facility in 1981.
The Israeli strikes completely destroyed the reactor and caused minimal
collateral damage.24
Previous research has debated the effect of the 1981 strike on the
Iraqi nuclear program. According to one view, the attack did little to
affect the program because Osirak – a 70 MW light water reactor – was
22
David Albright and Kevin O’Neill (eds), Solving the North Korean Nuclear Puzzle
(Washington DC: Inst for Science and International Security 2000).
23
Amos Perlmutter et al., Two Minutes over Baghdad (London: Routledge 2003).
24
Schlomo Nakdimon, First Strike: The Exclusive Story of How Israel Foiled Iraq’s
Attempt to Get the Bomb (New York: Summit Books 1987); Rodger Claire, Raid on
the Sun: Inside Israel’s Secret Campaign that Denied Saddam the Bomb (New York:
Broadway 2004).
Bombing Nuclear Facilities and Proliferation 169
Downloadedby[89.177.143.45]at15:1025November2014
not ideal for plutonium production. Skeptics of its ability to generate
plutonium for bombs point to a Congressional Research Service report
indicating that it would have taken 10–30 years to produce enough
plutonium for a bomb.25
Those advocating this position also suggest
that French suppliers would have been ‘highly motivated to report any
illegal weapons activity’ in the event that Iraq attempted to use a
civilian facility to produce plutonium for bombs.26
At the other end
of the spectrum is the view that ‘the Israeli counterproliferation effort
successfully prevented Iraq from building a nuclear weapon,’ destroying
the most critical Iraqi nuclear facility and making it impossible for
Iraq to produce more than six grams of plutonium by 1991.27
Somewhere in between is the argument that the use of military force
set back the Iraqi program at least several years; in 1981, French
nuclear engineers estimated that it would take four and a half years to
rebuild the facility.28
We find evidence supporting the view that the raid both directly and
indirectly delayed the Iraqi nuclear program. A report from French
scientists familiar with the project estimated that the reactor could
produce between three and ten pounds of plutonium annually, at about
seven pounds per bomb; ‘the risk is self-evident,’ reported one of the
scientists involved with the report.29
Jeremy Tamsett argues that Osirak
could have produced plutonium for 28 nuclear weapons by the end
of the decade.30
Yet another estimate suggested a production of about
8–10 kg of plutonium annually, enough for about one bomb a year.31
Moreover, the argument that Iraq could not have used Osirak for
military purposes rests on some questionable assumptions. For
example, the view that France would be motivated to report illegal
weapons activity and therefore that Iraq would have been unable to
produce enough plutonium is unconvincing. This assumes that France,
and the international community more generally, would be aware of
illicit activities in the event that they occurred. It is not clear that this
would have been the case, however. Although Osirak was under IAEA
25
Reiter, ‘Preventive Attacks against Nuclear Programs,’ 358.
26
Reiter, Preventive War.
27
Tamsett, ‘The Israeli Bombing of Osiraq Reconsidered’, 71–72.
28
Shai Feldman, ‘The Bombing of Osiraq – Revisited,’ International Security 7/2
(Autumn 1982), 126; Whitney Raas and Austin Long, ‘Osirak Redux? Assessing Israeli
Capabilities to Destroy Iranian Nuclear Facilities,’ International Security, 31/4 (2007),
10–11; Tim Keeley, ‘Angry French Contend Osirak Couldn’t Have Served Weapons
Purposes,’ Nucleonics Week, 18 June 1981, 3.
29
‘Warn Mitterrand about Iraqi Nuclear Reactor,’ United Press International, 18 Mar.
1982.
30
Tamsett, ‘The Israeli Bombing of Osiraq Reconsidered’.
31
Keeley, ‘Angry French.’
170 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
safeguards, Iraq had devised crafty ways of misleading inspectors that
verified compliance with the NPT.32
Inspectors did not have
a permanent presence in Iraq, making it possible for Baghdad to elude
detection. A final problem with this argument is that France was hardly
the poster child for nonproliferation during this era. Paris knowingly
helped Israel build nuclear weapons and refused to ratify the NPT until
the 1990s. It is by no means obvious that it would have been sufficiently
motivated to take action against Iraq in the name of nonproliferation.
Thus, by destroying a facility suited to plutonium production, Israel
removed Iraq’s past nuclear progress, supporting the direct mechanism
outlined above.
There is also evidence in favor of two of the indirect mechanisms
specified. The attacks provoked a shift in Iraq’s path toward the bomb.
Iraq had considered uranium enrichment before the attacks, but
accelerated those plans after the attacks, both because France did not
rebuild the reactor, but also because the plutonium path would have
been an easier target for subsequent attacks. Iraq did not completely
abandon its plutonium program, but focused the majority of its efforts
on the uranium path, with EMIS and to a far lesser extent gaseous
diffusion emerging as the top candidates.33
The problem with shifting courses is that Iraq lacked indigenous
knowledge necessary to master the complexities of enrichment
technology. The EMIS program faced technical challenges that limited
its ability to produce sufficient enriched uranium for a bomb.34
In part
because of these challenges, Iraq began working on gas centrifuge
technology. The centrifuge program required a sophisticated, foreign
technology with which Iraqis were not familiar; they encountered many
problems because of complexities of rotor dynamics that the Iraqi
scientists did not understand. As one Tuwaitha engineer suggested, ‘a
centrifuge is like a delicate souffle´ that will fall apart if anything is done
incorrectly, and our chefs were woefully unprepared.’35
The attack did provoke Saddam Hussein to intensify his support for
the Iraqi program, adding additional scientists, increasing financial
investment in the effort to produce the bomb.36
Increased resources
were not sufficient for an accelerated nuclear program, however, since
32
David Kay, ‘Denial and Deception Practices of WMD Proliferators: Iraq and
Beyond,’ Washington Quarterly, 18/1 (Jan. 1995), 85.
33
David Albright and Robert Kelley, ‘Massive Programs, Meager Results,’ Bulletin of
the Atomic Scientists (Nov./Dec. 1995), 56–64.
34
Ibid.
35
Mahdi Obeidi and Kurt Pitzer, The Bomb in My Garden: The Secrets of Saddam’s
Nuclear Mastermind (New York: Wiley 2004), 103.
36
Reiter, ‘Preventive Attacks against Nuclear Programs,’ 362.
Bombing Nuclear Facilities and Proliferation 171
Downloadedby[89.177.143.45]at15:1025November2014
the strikes led to insurmountable technical impediments. On the
contrary, the attempt to fast-track the bomb was counterproductive, as
the Iraqi scientists ‘tried to shortcut the difficult science of rotor
dynamics’ and burned out the centrifuges, with one scientist concluding
that ‘a little knowledge is dangerous indeed.’37
Thus, despite Saddam
attributing ‘a high value to the nuclear progress and talent that had
been developed to the 1991 war,’ technical challenges prevented him
from acquiring sufficient quantities of fissile material by the time of the
1991 Persian Gulf War.38
The Israeli raid also made France – Iraq’s most important nuclear
supplier – less likely to assist the program. France appears to have
considered rebuilding the reactor or resupplying Iraq with nuclear fuel
that posed less of a proliferation risk.39
One French official, however,
suggested that declarations about French willingness to assist Iraq in
resuscitating its program were ‘‘‘living-room hypotheses’’ designed to
save face for the Iraqis’ and that the reactor would never be rebuilt.40
In
any case, years passed and neither Mitterrand nor Chirac – despite
allegations that the latter had confidentially promised Saddam that he
would rebuild the facility – ever followed through.41
Mahdi Obeidi, a
high-ranking Iraqi nuclear scientist, regretted that ‘months passed, and
the promised French cooperation never materialized. For those of us
who had once envisioned an Iraqi nuclear program . . . the dream died
on the vine.’42
Iraq signed bilateral civilian nuclear cooperation
agreements with many countries including Brazil, France, Italy, and
the Soviet Union prior to 1981 but it had incredible difficulty securing
atomic assistance after the Israeli strike.43
There are no indications that the strike delayed Iraq’s nuclear
program by producing enhanced international inspections. IAEA
officials certainly opposed the raid, as they viewed it as an indictment
of the safeguards regime.44
Inspectors did not necessarily have greater
access to Iraqi officials following the strike, however.
37
Obeidi and Pitzer, The Bomb in My Garden, 104.
38
Iraq Survey Group, ‘Final Report’ 2004, 5http://www.globalsecurity.org/wmd/
library/report/2004/isg-final-report4.
39
Edward Cody, ‘France Plans to Bar Weapons-Grade Fuel for Iraq’s Reactor,’
Washington Post, 13 Jan. 1982.
40
Tim Keeley, ‘France Has Told Iraq it is Fully Willing to Rebuild,’ Nucleonics Week,
27 Aug. 1981, 2; Ann MacLachlan, ‘French-Iraqi Accord Makes No Provision for
Rebuilding Reactor,’ Nucleonics Week, 24/35 (1983), 1.’
41
Ann MacLachlan, ‘French-Iraqi Accord’.
42
Obeidi and Pitzer, The Bomb in My Garden, 52.
43
Fuhrmann, ‘Taking a Walk.’
44
See, for example, IAEA Director General Sigvard Edlund’s statement in ‘Peaceful
Nuclear Development Must Continue,’ IAEA Bulletin 23/3 (Dec. 1981), 4–5.
172 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
Israeli Attack against Syria’s Nuclear Program, 2007
Israel’s September 2007 strike on a nuclear facility in Syria was
undertaken under a shroud of secrecy. The attack destroyed a Syrian
reactor at Al Kibar that was in the early phases of development, likely
with assistance from North Korea.45
Unlike the attention and censure
surrounding the 1981 Osirak strike, the international reaction was
comparatively silent and weeks passed before Israeli officials acknowledged
that it occurred.46
A US intelligence briefing in April 2008
confirmed suspicions that the facility had been a nuclear reactor
camouflaged in order to minimize attention, but nonetheless had been
‘irreparably damaged’ by the September 2007 Israeli raid.47
The Israeli strike destroyed a facility similar to the North Korean
reactor at Yongbyon, which is well suited to plutonium production.
Operating at full power, the Syrian reactor could have produced about
one weapon’s worth of plutonium annually.48
Without a reprocessing
facility, which has not been located, Syria would have been unable to
extract plutonium from spent nuclear fuel, however. Moreover, while
the reactor was nearing operational capacity at the time of the attack,
full-scale operations would have been impossible in the absence of fuel
to operate the reactor; such fuel was missing and would have ‘required
weeks or months of testing once inside the reactor.’49
By destroying the
physical plant, however, Israel negated about six years of progress
toward nuclear development, the average time states have taken to
build a gas-cooled graphite-moderated reactor.50
In addition to the attack having the direct effect in terms of removing
past progress, there is also evidence supporting two of the indirect
theoretical mechanisms. The Israeli raid complicated Syria’s efforts by
triggering international investigations. Prior to the Al Kibar attack,
Syria’s program was largely unidentified and thus uninspected. After
the strikes, the IAEA solicited information on Syria’s program from
NPT member states. Several months later, the United States responded
45
Bill Gertz and Sara Carter, ‘US: Syria Hid N. Korea–Aided Nukes Plant,’ Washington
Times, 24 Apr. 2008.
46
Leonard Spector and Avner Cohen, ‘Israel’s Airstrike on Syria’s Reactor: Implications
for the Nonproliferation Regime,’ Arms Control Today, 38/6 (2008), 15–21.
47
United States Director of National Intelligence, ‘Background Briefing with Senior US
Officials on Syria’s Covert Nuclear Reactor and North Korea’s Involvement,’ 24 Apr.
2008, 5www.dni.gov/interviews/20080424_interview.pdf4 .
48
David Albright and Paul Brannan, ‘The Al Kibar Reactor: Extraordinary Camouflage,
Troubling Implications,’ Inst. for Science and International Security, 2008.
49
Peter Crail, ‘US Shares Information on NK-Syrian Nuclear Ties,’ Arms Control
Today (May 2008).
50
Zentner et al., Nuclear Proliferation Technology Trends Analysis.
Bombing Nuclear Facilities and Proliferation 173
Downloadedby[89.177.143.45]at15:1025November2014
with a detailed display of satellite images and other previously classified
evidence documenting the development of Syria’s nuclear reactor over a
period of years.51
The IAEA took several other steps. It demanded a visit to inspect
Syria’s suspected nuclear site. According to Mohamed ElBaradei, Syria
had ‘an obligation to report the planning and construction of any
nuclear facility to the agency . . . we are treating this information with
the seriousness it deserves.’52
In May 2008, the IAEA stated its
commitment to its safeguards responsibilities and informed Syria of its
intentions to send inspectors to review information and inspect the site
at Al Kibar. Syria responded with a letter that same month agreeing to
the visit. According to the IAEA, Syria ‘provided unrestricted access to
all of the buildings on the site’ during the June 2008 visit.53
On-site
inspections and imagery allowed the IAEA to conclude that the facility
was ‘similar to what may be found in connection with a reactor
site.’54
Additional inspections produced evidence of uranium particles
from a second site, leading to questions about why ‘material that was
not previously declared to the IAEA was detected at two facilities in
Syria, one of which was being constructed clandestinely.’55
The results
were sufficiently suspicious to land Syria on the IAEA’s official meeting
agenda and to keep the pressure on Syria after the attacks. As one IAEA
diplomat anonymously indicated, ‘the agency clearly thinks it has
something significant enough to report to put Syria on the [nuclear
safeguards] agenda right after North Korea and Iran.’56
Prior to the
inspection and investigation, information on Syria’s nuclear program
had been ‘inconclusive’ and Syria had remained off the official IAEA
meeting agenda.57
Syria has not been forthcoming in answering
questions uncovered during inspections, but the additional intelligence
51
Greg Miller and Paul Richter, ‘US Offers Evidence of North Korea–Syria Nuclear
Plant,’ New York Times, 25 Apr. 2008; Yossi Melman, ‘IAEA Asks for Information on
any Syrian Plans,’ Haaretz, 15 Oct. 2007.
52
Borzou Daragahi, ‘IAEA to Send Inspectors to Syria’s Alleged Nuclear Site,’ Los
Angeles Times, 3 June 2008.
53
International Atomic Energy Agency, Implementation of the NPT Safeguards
Agreement in the Syrian Arab Republic (Vienna, Austria: IAEA Board of Governors
2008).
54
Ibid.
55
Peter Crail, ‘IAEA Finds Uranium at Second Syrian Site,’ Arms Control Today, Jul./
Aug. 2009.
56
Damien McElroy, ‘Uranium Found in Syria by UN Nuclear Inspectors,’ Telegraph,
10 Nov. 2008.
57
Joby Warrick, ‘Syrian Facility Looked like Reactor, UN Says,’ Washington Post, 20
Nov. 2008; Reuters, ‘Report: IAEA Finds Traces of Uranium at Suspected Syrian
Nuclear Site,’ 10 Nov. 2008.
174 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
from member states, the first independent investigation of the reactor in
June 2008, and placement on the IAEA’s meeting agenda all indicate
that the IAEA is far more involved in scrutinizing Syria’s program
following the raid.
It is difficult to obtain comprehensive information on North Korea’s
post-attack intentions, but the raid appears to have made it less tenable
for Pyongyang to assist Syria’s nuclear program. Since Syria’s
indigenous capabilities are insufficient to build sophisticated nuclear
facilities at this point in time, the withdrawal of North Korean
assistance has frustrated the progress of its nuclear program.
Wartime Case Studies
Allied Attacks against Germany’s Nuclear Program, 1942–1945
Between 1942 and 1944, the allies waged four separate attacks on the
Norsk-Hydro heavy water facility in German-occupied Norway. In
October 1942, a 34-person British sabotage team in two Horsa gliders
crashed as it attempted to destroy stockpiles of heavy water at the
facility.58
This raid was a dismal operational failure and actually
prompted Germany to defend the facility more heavily, mining all
access points. In February 1943, skiers from the Royal Norwegian
Army dressed in British uniforms parachuted into Rjukan, the site of
the heavy water facility, and destroyed the heavy water.59
This act of
sabotage destroyed 18 electrolysis cells in the heavy water facility’s
chambers, flushed 500 kg of heavy water, and took the facility out of
commission for about two months.
In November 1943 the allies followed-up by attacking the facility by
air. Two hundred American B-17s dropped over seven hundred 1000pound
bombs on the plant.60
Many of these bombs missed or inflicted
only light damage on their targets, but 12 bombs successfully damaged
the facility. These airstrikes dispensed of more heavy water and shut
down the facility for months.61
Reports suggested that this attack was
58
John S. Craig, Peculiar Liaisons in War, Espionage, and Terrorism in the 20th
Century (New York: Algora Pub 2005); Thomas Powers, Heisenberg’s War: The Secret
History of the German Bomb (New York: Penguin Books 1993).
59
Powers, Heisenberg’s War; Knut Haukelid, Skis against the Atom (London: W.
Kimber 1954); Jeremy Bernstein, Hitler’s Uranium Club (Woodbury, NY: American
Inst. of Physics 1996).
60
Powers, Heisenberg’s War.
61
Per F. Dahl, Heavy Water and the Wartime Race for Nuclear Energy (London:
Taylor and Francis 1999).
Bombing Nuclear Facilities and Proliferation 175
Downloadedby[89.177.143.45]at15:1025November2014
one ‘of the most important and successful undertakings the Allied
saboteurs have carried out as yet during the war.’62
Germany was able to rebuild the facility quicker than the allies had
anticipated but facing the prospect of additional attacks, the Germans
decided to transfer materials involved in the production of heavy water
to the continent in 1944.63
A Norwegian saboteur who had been tipped
off by British intelligence intercepted the ferry Hydro that was
transporting heavy water and sank it into the bottom of Lake Tinnsjo
in Norway.64
This attack sank another 607 kg of heavy water and
reinforced the perils of maintaining a nuclear facility in occupied
territory.65
Iraqi Attacks against Iran’s Nuclear Program, 1984–1988
The Iran–Iraq War provided the backdrop for a series of strikes against
nuclear facilities. In 1980, Iranian F-4 Phantoms attacked Iraq’s Osiraq
plant en route home from a bombing raid. This strike was an
operational failure and it caused little damage to Osirak, necessitating
the Israeli raid one year later.66
Later in the war, Iraq raided Iran’s Bushehr reactors in a series of
attacks. The first strike took place in March 1984, and was followed by
subsequent attacks in each year of the war until a final raid in 1988, a
total of seven strikes over five years.67
Iraq’s initial airstrikes did
minimal damage to the reactors.68
It was not until November 1987 that
Iraqi airstrikes actually caused significant damage.69
According to a
German witness involved in the Iranian nuclear project, the 1987 raids
were ‘very accurate’ and ‘destroyed the entire core area of both units’
62
Lesley Groves, Now it Can be Told: The Story of the Manhattan Project, New Ed.
(Da Capo P 1983), 188.
63
Samuel Goudsmit, Alsos (American Inst of Physics 2006); Powers, Heisenberg’s War.
64
Groves, Now it Can be Told; Bernstein, Hitler’s Uranium Club.
65
The problem is that Germany had chosen heavy water as a moderator for its reactors
in 1939, but the Norsk-Hydro plant, in occupied Norway, was the only source of heavy
water for the German program. Dahl, Heavy Water.
66
‘The Ghosts that Hit Osirak,’ The Economist, 18 Oct. 1980, p.54.
67
Mark Hibbs, ‘Bushehr Construction Now Remote after Three Iraqi Air Strikes,’
Nucleonics Week, 28/48 (1987); Mark Hibbs, ‘Iraqi Attack on Bushehr Kills West
German Nuclear Official,’ Nucleonics Week, 28/47 (1987); Andrew Koch and Jeanette
Wolf, ‘Iran’s Nuclear Facilities: A Profile,’ Center for Nonproliferation Studies,
1998, 5http://cns.miis.edu/pubs/reports/pdfs/iranrpt.pdf4 .
68
Padraic Sweeney, ‘Iran Asks IAEA to Inspect Damage from Alleged Iraqi Attack on
Bushehr,’ Nucleonics Week, 26/8 (1985); Associated Press, ‘Agency Iranian Complaint
of Iraqi Attack on Nuclear Site,’ 30 May 1984.
69
Hibbs, ‘Bushehr Construction.’
176 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
and subsequently exposed them to a hostile climate of salt and extreme
temperatures.70
IAEA assessments found that the reactor was ‘certainly
not completely destroyed’ though considerable damage had been
done.71
Iraqi raids ultimately reversed a substantial amount of progress on
the Bushehr projects. The German contractor Kraftwerk Union began
constructing the facilities in 1974 and at the time of the first attack in
1984, the two reactors at Bushehr were 90 per cent and 50 per cent
complete, respectively.72
The Iraqi strikes – especially the 1987
attacks – necessitated nearly a complete reconstruction of the facilities.
Yet, Iraq needed to strike repeatedly over a period of four years in order
to achieve this result.
US Attacks against Iraq’s Nuclear Program, 1991, 1993
Nuclear facilities were among the high priority targets during the
1990–91 Persian Gulf War.73
In the initial stages of the war, coalition
aircraft struck the Tuwaitha Research Facility near Baghdad and F-
117s repeatedly bombed this plant throughout the campaign. The
United States also struck a suspected uranium feedstock production
facility near Mosul and a uranium extraction facility at Al Qaim. These
attacks were mixed in terms of their damage to Iraq’s nuclear
infrastructure. In 1991, the key chokepoints relevant to Iraq’s weapons
program were the facilities related to the EMIS and gas centrifuge
enrichment programs. The bombing raids destroyed several of the
chokepoint facilities, especially those relevant to Iraq’s EMIS enrichment
program. As the Iraq Survey Group (ISG), also referred to as the
Duelfer Report, concluded in the aftermath of the 2003 Iraq War:
‘Nearly all of the key nuclear facilities . . . were bombed during Desert
Storm . . . Many of the facilities located at Tuwaitha were devastated,
and the EMIS enrichment plants at Tarmiya and Ash Sharqat were
largely destroyed.’74
Other key facilities were not destroyed, however, because the United
States was unaware of their existence or their location. The yellowcake
70
Tony Walker, ‘Tehran Says 10 Die as Iraq Hits Nuclear Site,’ Financial Times, 18
Nov. 1987, 3.
71
Mark Hibbs, ‘Bonn will Decline Teheran Bid to Resuscitate Bushehr Project,’
Nucleonics Week, 32/18 (1987); Nuclear News, ‘German Reactor Expert Killed in
Bushehr Attack,’ Jan. 1998, 61.
72
Koch and Wolf, ‘Iran’s Nuclear Facilities: A Profile’.
73
Thomas A. Keaney and Eliot A. Cohen (eds), Gulf War Air Power Survey: Summary
Report (Washington, DC: USGPO 1993), 80.
74
Iraq Survey Group, ‘Final Report’, 2004.
Bombing Nuclear Facilities and Proliferation 177
Downloadedby[89.177.143.45]at15:1025November2014
facility at Al-Qa’im, feed material plant at Mozul, and high explosives
testing site (Al-Athir) were damaged, but the centrifuge facility at
Rashdiya was ‘neither found nor targeted in the 1991 war.’75
The Gulf
War Air Power Survey underscored the challenges associated with
locating and targeting Iraqi nuclear facilities during the war. It stated,
‘we now know that the Iraqis’ program to amass enough enriched
uranium to begin producing atomic bombs was more extensive, more
redundant, further along, and considerably less vulnerable to air attack
than was realized at the outset of Desert Storm.’76
As inspections
discovered soon after the Gulf War, Iraq had three times more nuclear
facilities than military planners believed during the war.77
The case of
Ash Sharqat is representative of the coalition’s targeting challenges. It
was thought to be a rocket facility rather than one related to Iraq’s
nuclear program; the facility was the subject of a series of attacks and
then dismissed, ‘because intelligence did not suspect Ash Sharqat of
nuclear activities.’78
In the aftermath of the Gulf War, the United States again struck
suspected nuclear facilities. On January 17, 1993, the US Navy used
Tomahawk Land Attack Missiles against facilities that had largely
escaped unscathed from the Gulf War: Facility 409 (Ma’malal’Rabia’)
that manufactured calutrons for the Iraqi EMIS program and Facility
416 (Al-Dijla) that produced power supplies for the EMIS project.
These attacks were reasonably successful at the operational level.
UNSCOM and IAEA teams found that the Navy Tomahawk Land
Attack Missiles had successfully hit the buildings and destroyed
sensitive machine tools in the 1993 raid and could be a considered an
operational success.79
Discussion and Conclusion
The standard debate on whether military force delays proliferation is
typically cast in stark terms. One side of the debate suggests that
attacks offer the prospect of unequivocal success in delaying nuclear
proliferation; the other counters that the use of force can actually
backfire by accelerating the target state’s nuclear programs. History
tells a more complicated story. In this section we discuss the
75
Ibid.
76
Keaney and Cohen (eds), Gulf War Air Power Survey: Summary Report, 82.
77
Ibid., 431.
78
Ibid., 431, 434.
79
Federation of American Scientists, ‘Zaafaraniyah,’ 2000, 5www.fas.org/nuke/guide/
iraq/facility/zaafaraniyah.htm4 .
178 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
conclusions that emerge from our analysis and comment on what the
historical record says about the likely effect of an attack against Iran’s
nuclear program.
We theorized that strikes could delay progress through both direct
and indirect mechanisms. Peacetime cases produced some support for
the general argument that attacks delay states’ acquisition of fissile
material and for the specific mechanisms, but the size of this effect was
generally modest.
The 1981 Osiraq raid offered support for three of the four
mechanisms outlined. First, the Israeli attack destroyed a key
chokepoint for Iraq’s nuclear program (although the Israelis left the
reprocessing facility intact), evidence supporting the direct mechanism.
Second, the attack had an indirect effect by prompting Iraq to switch
from reprocessing technology to centrifuge technology, which it hoped
would be more easily concealed. Centrifuge enrichment proved
technically challenging and ineffective for the Iraqis who knew little
about the technology. Third, compounding matters further, France
withdrew its support from the program after the strike. In the absence
of indigenous know-how or materials, Iraq found uranium enrichment
to be laborious and time-consuming since its scientists did not
understand rotor dynamics and spent years either burning up the
centrifuges or trying to make do with poor quality centrifuges.
Israel’s raid on Syria in September 2007 likewise delayed Damascus’
ability to build nuclear weapons both through the direct mechanism of
removing past progress and through two of the indirect mechanisms.
The Israeli raid destroyed a graphite-moderated reactor modeled after
the North Korean facility at Yongbyon. This facility is a chokepoint
because it could have been used to produce plutonium for nuclear
weapons. That said, Syria does not appear to have been close to
acquiring the bomb at the time of the raid. Future progress toward the
bomb, however, has been made more difficult because the attack
prompted IAEA attention and inspections to a program that had
previously been unidentified and uninspected. Moreover, there is no
evidence that Pyongyang has agreed to rebuild the reactor at Al Kibar.
Paradoxically, these two raids produced delays in part because the
Iraqi and Syrian nuclear programs were in their relative infancy. In
neither instance did the target state possess the means to produce fissile
material for nuclear bombs at the time of the attack. Indeed, the Israelis
struck before the Osirak and Al Kibar reactors had gone critical. But
Baghdad and Damascus had critical chokepoints that were concentrated
in a single area, making it easier for the Israelis to delay progress
with one attack. Thus, it appears that attacking countries can achieve
the most success before a program becomes ‘a train without brakes,’ to
borrow a phrase from Iranian President Mahmoud Ahmadinejad. Yet,
Bombing Nuclear Facilities and Proliferation 179
Downloadedby[89.177.143.45]at15:1025November2014
the timeframe in which strikes might be most effective is also when they
would be considered the least legitimate. Anything other than
preemptive uses of force (i.e., striking to prevent an imminent attack)
are considered illegal under international law and the international
community might be less likely to endorse attacks when it is not
obvious that the target was on the verge of acquiring nuclear
weapons.80
The wartime cases underscore the reasons why using military force to
delay proliferation can encounter challenges. The dual-use nature of
nuclear complexes and the relative inconspicuousness of centrifuge
facilities make it possible for states to maintain a covert military
nuclear program.81
As a result, states seeking to minimize proliferation
might lack timely or clear indicators on the status or whereabouts of a
proliferating state’s nuclear facilities. The 1991 Persian Gulf War case
illustrates this problem. Many key chokepoint facilities, particularly
those relating to Iraq’s gas centrifuge program, were not destroyed
during US airstrikes because their locations were unknown. Importantly,
this problem is not limited to the wartime cases. The 1981
Osirak strike is also suggestive of an intelligence gap because the
reprocessing facility next to the reactor was not targeted. Either
the Israelis thought the reprocessing facility was located beneath the
reactor, or they were unaware that this chokepoint existed at all.
Regardless, while the Israelis successfully destroyed the Osirak reactor,
they altogether neglected the adjacent plutonium reprocessing facility.
What does the historical record suggest about the consequences of a
potential American or Israeli strike against Iran’s nuclear program?
Although military force delayed proliferation in some previous cases,
policymakers must remember that past may not be prologue. In
particular, the three indirect mechanisms we identified are unlikely to
‘work’ in the Iranian case. Tehran received helpful nuclear assistance in
the past, but it does not depend on external support today to sustain its
military program. It currently receives civilian nuclear assistance from
Russia but it is unclear that the withdrawal of this aid would have a
major impact on its ability to produce fissile material for nuclear
weapons. It is also unlikely that an attack would lead to a change in
Iran’s fissile material production strategy. Tehran is already relying
primarily on centrifuge enrichment technology which is easier to
conceal than facilities necessary for plutonium production (e.g.,
reactors and reprocessing centers). Centrifuges are likewise easier to
80
Michael Walzer, Just and Unjust Wars: A Moral Argument with Historical
Illustrations (New York: Basic Books 2000).
81
Michael S. Goodman and Wyn Q. Bowen, ‘Behind Iran’s Nuclear Weapons ‘‘Halt,’’’
Bulletin of the Atomic Scientists, 19 Feb. 2008.
180 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
hide than other enrichment technologies, such as gaseous diffusion
plants. The third indirect mechanism could have a modest effect in
delaying Iran’s nuclear program. Inspectors from the IAEA have been
on the ground in Iran for decades, but they have had only limited
success in detecting transgressions in a timely fashion. If an attack
caused Iran to enter the Additional Protocol (AP), which provides the
IAEA the authority to visit any facility in a country, this could frustrate
weaponization efforts. Potential attackers should not count on this
outcome given that Syria granted the IAEA some additional access after
being attacked but has still not committed to the AP.
This suggests that the direct, physical destruction of Iranian nuclear
facilities would be the main route by which an attack could delay
progress.82
The most critical facilities for Iran’s nuclear program are (1)
the uranium enrichment plants at Natanz and Qom, (2) the Arak heavy
water production center, and (3) the Isfahan uranium conversion
facility. Of these facilities, the most sensitive are the enrichment plants
because they could provide a critical source of fissile material for
nuclear weapons (i.e., HEU). The plants at Arak and Isfahan are
significant but they are alone insufficient to provide Iran with bombgrade
materials. How much time could Israel or the United States buy
by destroying the two uranium enrichment facilities? The history of
nuclear programs reveals that it takes an average of 14 years to go from
the initiation of a gas centrifuge program to the completion of the
first full-scale facility.83
Iran is already well behind the average time
since it initiated its program in 1987 and it did not demonstrate
operational capacity until 2004. It would not take 17 years to
demonstrate operational capacity as it did before, since Iran has
acquired a significant amount of indigenous knowledge that cannot
realistically be taken away in an attack. But in all likelihood, a raid
would still delay the program. Considering that it took India five years
to construct a second centrifuge enrichment facility once it completed a
pilot plant, we could assume that destroying Natanz and other related
enrichment facilities could delay Iran’s ability to produce fissile
material by about the same amount of time. This is a relatively modest
gain in light of the well-known risks associated with striking Iran’s
nuclear facilities.84
Yet, policymakers who adopt short time horizons
may calculate that a delay of up to five years would justify the dangers
of preventive military strikes.
82
For more on when and why states target suspected proliferators’ nuclear facilities, see
Fuhrmann and Kreps, ‘Targeting Nuclear Programs in War and Peace.’
83
Zentner et al., Nuclear Proliferation Technology Trends Analysis.
84
See, for example, Caitlin Talmadge, ‘Closing Time: Assessing the Iranian Threat to
the Strait of Hormuz,’ International Security, 33/1 (Summer 2008), 82–117.
Bombing Nuclear Facilities and Proliferation 181
Downloadedby[89.177.143.45]at15:1025November2014
Either way, it is critical to recognize that this assessment rests on two
fairly ambitious assumptions. The first is that all of Iran’s sensitive
nuclear facilities are known to Israel and/or the United States. History
provides good reason to doubt that this is true. For the last several years
the IAEA has been ‘unable effectively to monitor the R&D activities
being carried out by Iran,’ except at sites with safeguarded materials,
meaning that the agency cannot address concerns about the existence of
covert facilities.85
Revelations of the second enrichment plant at Qom –
also known as the Fordow Fuel Enrichment Plant – did not emerge until
September 2009. It is unclear when Western intelligence agencies
discovered this facility, but construction likely began in 2002.86
The facility is located in an underground tunnel complex at a site
controlled by the Islamic Revolutionary Guards Corps. Given that Iran
managed to keep this facility secret for seven years, it is not implausible
that there are other covert facilities that remain unknown even to
intelligence services. Our analysis of the wartime cases further underscores
this point. The United States was unaware of many critical nuclear
facilities in Iraq prior to the 1990–91 Persian Gulf War, for example.
The second assumption deals with the operational feasibility of
an attack, a question that has received excellent scholarly treatment
elsewhere.87
Although the affordability and ubiquity of precision
weapons available means that targeting states are likely to hit known
targets,88
a factor that offsets improvements in military technology is
that potential targets have learned from previous attacks and taken
appropriate defensive measures. Just as Germany learned that it needed
to better defend the Norsk-Hydro facility following the first Allied
attack, Iran has learned from the Osirak and Al Kibar strikes that it
should not concentrate its nuclear facilities in one location. Doing so
85
International Atomic Energy Agency, Implementation of the NPT Safeguards
Agreement in the Islamic Republic of Iran, GOV/2006/8 (Vienna, Austria: IAEA
Board of Governors 27 Feb. 2006). See also Michael Eisenstadt, ‘Iran: The Complex
Calculus of Preventive Military Action,’ Washington Institute for Near East Policy,
Policy Watch #1152, 25 Sept. 2006.
86
International Atomic Energy Agency, ‘Implementation of the NPT Safeguards
Agreement and Relevant Provisions of Security Council Resolutions 1737 (2006), 1747
(2007), 1803 (2008) and 1805 (2008) in the Islamic Republic of Iran,’ Report by the
Director General, International Atomic Energy Agency, 16 Nov. 2009. See also,
Nuclear Threat Initiative, ‘Iran: Nuclear Overview,’ March 2010, 5www.nti.org/
e_research/profiles/iran/Nuclear/index.html4 .
87
Raas and Long, ‘Osirak Redux?’.
88
See John E. Peters et al., ‘Allied Power Projection Capabilities,’ in Persian Gulf
Security: Improving Allied Military Contributions (Arlington, VA: Rand Corp 2001),
75; Ryan Hansen, ‘JDAM Continues to be Warfighter’s Weapon of Choice,’ Air Force
Link, 17 Mar. 2006, 5www.af.mil/news/story.asp?storyID ¼ 1230176134 .
182 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
makes it vulnerable to the possibility of a one-strike success, whereas
disseminating the facilities makes each one less vulnerable. From a
probabilistic standpoint, the more targets that attackers have to hit, the
lower the likelihood of net success.
In sum, given that Iran already possesses the requisite knowledge to
enrich uranium – and this knowledge cannot be taken away – the best
possible outcome of military force would be delaying Tehran’s ability
to build nuclear weapons by around five years. Based on our survey of
the historical record, it is far from obvious that military force would
yield even this modest return. Policymakers should also be aware that
multiple attacks against Iran might be necessary. We now know that
Iraq terminated its nuclear weapons program in the 1990s, but this
happened only after three different countries (Iran, Israel, and the
United States) had attacked its facilities.
With this cautious conclusion in mind, we propose a few next steps
for research. One step is to undertake a systematic study of potential
costs – diplomatic, economic, or military – of using force. This analysis
bracketed the question of costs, since if military force does not delay the
target state’s nuclear program, then the strategy has nothing to
recommend it, even if the costs are negligible. However, the
effectiveness question is just one side of the ledger and the overall
utility of force is best assessed by taking into account the possible costs
to the attacking state. For example, if the target state has the ability to
launch counter strikes, the costs from attacking nuclear facilities might
outweigh the benefits and justifiably deter the attacking state from
using force. Indeed, the fear of high costs in part explains why the
United States refrained from attacking China in the 1960s and North
Korea in the 1990s89
and may be one reason for caution even if
removing the Iranian facilities through force is operationally feasible.
Having identified the reasons why striking nuclear programs can
hinder the target state’s proliferation goals, we also suggest analyzing
whether tools other than force can provoke the same mechanisms we
discuss here. We also urge future research into why countries choose
force to oppose proliferation, since we have confined our focus to the
consequences, not causes of attacks. If raids delay proliferation and
most countries highlight the spread of nuclear weapons as the greatest
threat to their national security, why have strikes occurred relatively
infrequently? At what point – whether relative to the target state’s
nuclear program or to the instruments that have been tried – do states
resort to force? What explains the ‘near misses,’ the cases where states
89
William Burr and Jeffrey Richelson, ‘Whether to Strangle the Baby in the Cradle,’
International Security 25/3 (Winter 2000/1), 54–99; Joel Wit et al., Going Critical: The
First North Korean Nuclear Crisis (Washington DC: Brookings Institution Press 2005).
Bombing Nuclear Facilities and Proliferation 183
Downloadedby[89.177.143.45]at15:1025November2014
considered force but decided against it? How do the perceived high
costs of attacking, normative constraints, or other factors affect the
calculus on using force? Based on the gravity of proliferation and
military force, we conclude that these are all important avenues for
future study.
Acknowledgments
We thank Avner Cohen, Alexander Downes, Bryan Early, Michael
Horowitz, Matthew Kroenig, Austin Long, Sean Lynn-Jones, Martin
Malin, Alexander Montgomery-Amo, Thomas Nichols, Bennett Ramberg,
Etel Solingen, Leonard Spector, and anonymous reviewers for
helpful comments and criticisms. Participants at the Center for Security
Studies’ workshop ‘Uncovering the Sources of Nuclear Proliferation,’ in
Zurich, Switzerland and a research seminar at the James Martin Center
for Nonproliferation Studies in Washington, DC also provided valuable
feedback. Any remaining errors are our own.
Bibliography
Albright, David and Paul Brannan, ‘The Al Kibar Reactor: Extraordinary Camouflage, Troubling
Implications,’ Inst. for Science and International Security, 2008.
Albright, David and Robert Kelley, ‘Massive Programs, Meager Results,’ Bulletin of the Atomic
Scientists (Nov./Dec. 1995), 56–64.
Albright, David and Kevin O’Neill (eds), Solving the North Korean Nuclear Puzzle (Washington
DC: Inst for Science and International Security 2000).
Associated Press, ‘Agency Iranian Complaint of Iraqi Attack on Nuclear Site,’ 30 May 1984.
Bernstein, Jeremy, Hitler’s Uranium Club (Woodbury, NY: American Inst. of Physics 1996).
Betts, Richard K., ‘The Osirak Fallacy,’ The National Interest 83 (2006), 22–5.
Black, Ian and Simon Tisdall, ‘Saudi Arabia Urges US Attack on Iran to Stop Nuclear Programme,’
The Guardian, 28 Nov. 2010.
Bratt, Duane, The Politics of CANDU Exports (Toronto: U of Toronto P 2006).
Burr, William and Jeffrey Richelson, ‘Whether to Strangle the Baby in the Cradle,’ International
Security 25/3 (Winter 2000/1), 54–99.
Claire, Rodger, Raid on the Sun: Inside Israel’s Secret Campaign that Denied Saddam the Bomb
(New York: Broadway 2004).
Cody, Edward, ‘France Plans to Bar Weapons-Grade Fuel for Iraq’s Reactor,’ Washington Post, 13
Jan. 1982.
Craig, John S., Peculiar Liaisons in War, Espionage, and Terrorism in the 20th Century (New
York: Algora Pub 2005).
Crail, Peter, ‘IAEA Finds Uranium at Second Syrian Site,’ Arms Control Today, Jul./Aug. 2009.
Crail, Peter, ‘US Shares Information on NK-Syrian Nuclear Ties,’ Arms Control Today (May
2008).
Dahl, Per F., Heavy Water and the Wartime Race for Nuclear Energy (London: Taylor and Francis
1999).
Daragahi, Borzou, ‘IAEA to Send Inspectors to Syria’s Alleged Nuclear Site,’ Los Angeles Times, 3
June 2008.
184 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
Eisenstadt, Michael, ‘Iran: The Complex Calculus of Preventive Military Action,’ Washington
Institute for Near East Policy, Policy Watch #1152, 25 Sept. 2006.
Federation of American Scientists, ‘Zaafaraniyah,’ 2000, 5www.fas.org/nuke/guide/iraq/facility/
zaafaraniyah.htm4.
Feldman, Shai, ‘The Bombing of Osiraq – Revisited,’ International Security 7/2 (Autumn 1982),
114–42.
Fuhrmann, Matthew, ‘Spreading Temptation: Proliferation and Peaceful Nuclear Cooperation
Agreements,’ International Security 34/1 (2009), 7–41.
Fuhrmann, Matthew, ‘Taking a Walk on the Supply Side: The Determinants of Civilian Nuclear
Cooperation,’ Journal of Conflict Resolution 53/2 (2009), 181–208.
Fuhrmann, Matthew and Sarah E. Kreps, ‘Targeting Nuclear Programs in War and Peace: A
Quantitative Empirical Analysis, 1941–2000,’ Journal of Conflict Resolution 54/6 (2010), 831–
59.
Gardner, Gary, Nuclear Nonproliferation: A Primer (London: Lynne Rienner 1994).
Gertz, Bill and Sara Carter, ‘US: Syria Hid N. Korea-Aided Nukes Plant,’ Washington Times, 24
Apr. 2008.
‘The Ghosts that Hit Osirak,’ The Economist, 18 Oct. 1980, p.54.
Goodman, Michael S. and Wyn Q. Bowen, ‘Behind Iran’s Nuclear Weapons ‘‘Halt,’’’ Bulletin of
the Atomic Scientists, 19 Feb. 2008.
Goudsmit, Samuel, Alsos (American Inst of Physics 2006).
Groves, Lesley, Now it Can be Told: The Story of the Manhattan Project, New Ed. (Da Capo P
1983).
Gulf War Air Power Survey (GWAPS), Vol. 1, ‘Planning and Command and Control,’
Washington DC: Government Printing Office 1993.
Gulf War Air Power Survey (GWAPS), Vol. 2, ‘Planning and Command and Control,’
Washington DC: Government Printing Office 1993.
Hansen, Ryan, ‘JDAM Continues to be Warfighter’s Weapon of Choice,’ Air Force Link. 17 Mar.
2006, 5www.af.mil/news/story.asp?storyID¼1230176134.
Haukelid, Knut, Skis against the Atom (London: W. Kimber 1954).
Hibbs, Mark, ‘Bonn will Decline Teheran Bid to Resuscitate Bushehr Project,’ Nucleonics Week,
32/18 (1987).
Hibbs, Mark, ‘Bushehr Construction Now Remote after Three Iraqi Air Strikes,’ Nucleonics
Week, 28/48 (1987).
Hibbs, Mark, ‘Iraqi Attack on Bushehr Kills West German Nuclear Official,’ Nucleonics Week,
28/47 (1987).
International Atomic Energy Agency, Implementation of the NPT Safeguards Agreement in the
Islamic Republic of Iran, GOV/2006/8 (Vienna, Austria: IAEA Board of Governors 27 Feb.
2006).
International Atomic Energy Agency, ‘Implementation of the NPT Safeguards Agreement and
Relevant Provisions of Security Council Resolutions 1737 (2006), 1747 (2007), 1803 (2008)
and 1805 (2008) in the Islamic Republic of Iran,’ Report by the Director General, International
Atomic Energy Agency, 16 Nov. 2009.
International Atomic Energy Agency, Implementation of the NPT Safeguards Agreement in the
Syrian Arab Republic (Vienna, Austria: IAEA Board of Governors 2008).
Iraq Survey Group, ‘Final Report’, 2004, 5www.globalsecurity.org/wmd/library/report/2004/isg-
final-report/4.
Jo, Dong-Joon and Erik Gartzke, ‘The Determinants of Nuclear Weapons Proliferation,’ Journal
of Conflict Resolution 51 (2007), 167–94.
Kay, David, ‘Denial and Deception Practices of WMD Proliferators: Iraq and Beyond,’
Washington Quarterly, 18/1 (Jan. 1995), 85.
Keaney, Thomas A. and Eliot A. Cohen (eds), Gulf War Air Power Survey: Summary Report
(Washington, DC: USGPO 1993).
Bombing Nuclear Facilities and Proliferation 185
Downloadedby[89.177.143.45]at15:1025November2014
Keeley, Tim, ‘Angry French Contend Osirak Couldn’t have Served Weapons Purposes,’ Nucleonics
Week, 18 June 1981, 3.
Keeley, Tim, ‘France has Told Iraq it is Fully Willing to Rebuild,’ Nucleonics Week, 27 Aug. 1981, 2.
Koch, Andrew and Jeanette Wolf, ‘Iran’s Nuclear Facilities: A Profile,’ Center for Nonproliferation
Studies. 1998, 5http://cns.miis.edu/pubs/reports/pdfs/iranrpt.pdf4.
Kroenig, Matthew, ‘Importing the Bomb: Sensitive Nuclear Assistance and Nuclear Proliferation,’
Journal of Conflict Resolution 53/2 (2009), 161–180.
MacLachlan, Ann, ‘France’s Prime Minister Denies He Promised to Rebuild the Osirak Reactor,’
Nucleonics Week, 28/33 (1987), 16.
MacLachlan, Ann, ‘French-Iraqi Accord Makes No Provision for Rebuilding Reactor,’ Nucleonics
Week, 24/35 (1983), 1.
McElroy, Damien, ‘Uranium Found in Syria by UN Nuclear Inspectors,’ Telegraph, 10 Nov. 2008.
Melman, Yossi, ‘IAEA Asks for Information on any Syrian Plans,’ Haaretz, 15 Oct. 2007.
Miller, Greg and Paul Richter, ‘US Offers Evidence of North Korea-Syria Nuclear Plant,’ New
York Times, 25 Apr. 2008.
Nakdimon, Schlomo, First Strike: The Exclusive Story of How Israel Foiled Iraq’s Attempt to Get
the Bomb (New York: Summit Books 1987).
Nuclear News, ‘German Reactor Expert Killed in Bushehr Attack,’ Jan. 1998, 61.
Nuclear Threat Initiative, ‘Iran: Nuclear Overview,’ March 2010, 5www.nti.org/e_research/
profiles/Iraq/Nuclear/2117.html4.
Obeidi, Mahdi and Kurt Pitzer, The Bomb in My Garden: The Secrets of Saddam’s Nuclear
Mastermind (New York: Wiley 2004).
‘Peaceful Nuclear Development Must Continue,’ IAEA Bulletin 23/3 (Dec. 1981), 4–5.
Perlmutter, Amos, Michael I. Handel, and Uri Bar-Joseph, Two Minutes over Baghdad (London:
Routledge 2003).
Peters, John E., David Shlapak, and Timothy Liston, ‘Allied Power Projection Capabilities,’ in
Richard Sokolsky, Stuart E. Johnson, F. Stephen Larrabee, Ian O. Lesser, John E. Peters, David
A. Shlapak and Timothy Liston (eds), Persian Gulf Security: Improving Allied Military
Contributions (Arlington, VA: Rand Corp 2001),71–104.
Powers, Thomas, Heisenberg’s War: The Secret History of the German Bomb (New York: Penguin
Books 1993).
Raas, Whitney and Austin Long, ‘Osirak Redux? Assessing Israeli Capabilities to Destroy Iranian
Nuclear Facilities,’ International Security, 31/4 (2007), 7–33.
Ramberg, Bennett, ‘Preemption Paradox,’ Bulletin of Atomic Scientists, Jul./Aug. 2006, 48–56.
Reiter, Dan, ‘Preventive Attacks against Nuclear, Biological, and Chemical Weapons Programs:
The Track Record,’ in William Walton Keller and Gordon R. Mitchell (eds), Hitting First,
Preventive Force in US Security Strategy (Pittsburgh, PA: Pittsburgh UP 2006).
Reiter, Dan, ‘Preventive Attacks against Nuclear Programs and the ‘‘Success’’ at Osiraq,’
Nonproliferation Review, (12 July 2005), 355–371.
Reiter, Dan, Preventive War and its Alternatives: The Lessons of History (Carlisle, PA: Strategic
Studies Institute 2006).
Reuters, ‘Report: IAEA Finds Traces of Uranium at Suspected Syrian Nuclear Site,’ 10 Nov. 2008.
Sagan, Scott, ‘Why Do States Build Nuclear Weapons? Three Models in Search of the Bomb,’
International Security 21/3 (1996/97), 54–86.
Sagan, Scott and Kenneth Waltz, The Spread of Nuclear Weapons: A Debate Renewed (New
York, W.W. Norton 2002).
Singh, Sonali and Christopher Way, ‘The Correlates of Nuclear Proliferation: A Quantitative
Test,’ Journal of Conflict Resolution 48/6 (2004), 859–95.
Solingen, Etel, Nuclear Logics: Contrasting Paths in East Asia and the Middle East (Princeton, NJ:
Princeton UP 2007).
Spector, Leonard S. and Avner Cohen, ‘Israel’s Airstrike on Syria’s Reactor: Implications for the
Nonproliferation Regime,’ Arms Control Today, 38/6 (2008), 15–21.
186 Sarah E. Kreps and Matthew Fuhrmann
Downloadedby[89.177.143.45]at15:1025November2014
Sweeney, Padraic, ‘Iran Asks IAEA to Inspect Damage from Alleged Iraqi Attack on Bushehr,’
Nucleonics Week, 26/8 (1985).
Talmadge, Caitlin, ‘Closing Time: Assessing the Iranian Threat to the Strait of Hormuz,’
International Security, 33/1 (Summer 2008), 82–117.
Tamsett, Jeremy, ‘The Israeli bombing of Osiraq Reconsidered: Successful Counterproliferation?’
The Nonproliferation Review, 11/3 (2004), 70–85.
United States Director of National Intelligence, ‘Background Briefing with Senior US Officials on
Syria’s Covert Nuclear Reactor and North Korea’s Involvement,’ 24 Apr. 2008, 5www.dni.
gov/interviews/20080424_interview.pdf .
Walker, Tony, ‘Tehran Says 10 Die as Iraq Hits Nuclear Site,’Financial Times, 18 Nov. 1987, 3.
Walzer, Michael, Just and Unjust Wars: A Moral Argument with Historical Illustrations (New
York: Basic Books 2000).
‘Warn Mitterrand about Iraqi Nuclear Reactor,’ United Press International, 18 Mar. 1982.
Warrick, Joby, ‘Syrian Facility Looked like Reactor, UN Says,’ Washington Post, 20 Nov. 2008.
Wit, Joel, Daniel Poneman, and Robert Gallucci, Going Critical: The First North Korean Nuclear
Crisis (Washington DC: Brookings Institution P 2005).
Zentner, M.D., G.L Coles, and R.J. Talbert, Nuclear Proliferation Technology Trends Analysis,
2005 (Richland, WA: United States Department of Energy, Pacific Northwest National Lab
2005).
Bombing Nuclear Facilities and Proliferation 187
Downloadedby[89.177.143.45]at15:1025November2014