Did Stuxnet work?

Did Stuxnet work? Not as well as we think, says Ivanka Barzashka in a new RUSI Journal article, Are Cyber-Weapons Effective? Assessing Stuxnet’s Impact on the Iranian Enrichment Programme:

[C]entrifuge numbers unexpectedly dropped after the first Stuxnet attack wave in June 2009. Inspectors’ records show that between May and August that year, while the total number of operational centrifuges decreased by more than 300, the amount of machines being installed grew by almost twice that amount – indicating either extreme confidence or a desperate attempt to try to keep enrichment production steady. This indicates that Iran was bringing in additional centrifuges, but fewer machines were actually working [...] Consecutive Stuxnet attacks showed no net effect on centrifuge numbers at Natanz; in fact, the number of running centrifuges slightly increased after the malware’s second and third attack waves in March and April 2010 [...]

In fact, uranium-enrichment capacity grew during the time that Stuxnet was said to have been destroying Iranian centrifuges. Iran produced more enriched uranium, more efficiently: the entire plant’s separative capacity per day increased by about 40 per cent, despite the fluctuations in centrifuge numbers, and the IR-1 centrifuges’ annual separative capacity jumped by about 60 per cent [...]

Analysis of trends in centrifuge numbers shows a correlation between an unexplained drop in machines and the first Stuxnet attack in 2009, but not consecutive attacks – contrary to reports that the malware was wrecking Iranian machines in 2010. If sabotage did occur, it was short-lived and most likely happened between May and November 2009. The situation appears to have been under control by 2010. More significantly, Iran’s ability to successfully operate new machines was not hindered.

See, also, my book review in the FT of Thomas Rid’s Cyber War Will Not Take Place.

Richard Betts on deterrence, Iran, and ambiguity

Richard Betts, in a long Foreign Affairs essay on deterrence, talks about Iran:

Nevertheless, rather than planning to deter a prospective Iranian nuclear arsenal, the United States and Israel have preferred preventive war. Although many still hope to turn Iran away from nuclear weapons through sanctions and diplomacy, the debate within and between the United States and Israel over what to do if Iran moves to produce a bomb is about not whether to attack but when. U.S. President Barack Obama has firmly declared that he has not a “policy of containment” but rather “a policy to prevent Iran from obtaining a nuclear weapon,” and other administration officials have repeatedly emphasized this point. As promises in foreign policy go, this one is chiseled in stone. Backing down from it when the time comes would be the right thing to do but would represent an embarrassing retreat.

The logic behind rejecting deterrence is that Tehran might decide to use nuclear weapons despite facing devastating retaliation. The risk can never be reduced to zero, but there is no reason to believe that Iran poses more danger than other nasty regimes that have already developed nuclear weapons. The most telling example is North Korea. Although the American public has not paid nearly as much attention to North Korea, Pyongyang’s record of fanatical belligerence and terrorist behavior over the years has been far worse than Tehran’s.

Refusing to accept an iota of risk from Iran ignores the massive risks of the alternative of initiating war. Leaving aside the danger of being blind-sided by unanticipated forms of Iranian reprisal — for example, the use of biological weapons — the obvious risks include Iranian retaliation by overt or covert military means against U.S. assets. The results of the initially successful assault on Iraq in 2003 are a reminder that wars the United States starts do not necessarily end when and how it wants them to. Indeed, the records of the United States and Israel suggest that both countries tend to underestimate the prospective costs of the wars they enter. Washington paid fewer costs than expected during the Gulf War but faced a far higher bill than anticipated in Korea, Vietnam, Kosovo, Afghanistan, and the second war against Iraq. Israel suffered less in the 1967 Six-Day War than expected but was badly surprised by the costs of the 1973 Yom Kippur War, the 1982 Lebanon war, and the 2006 war against Hezbollah.

Launching a war against Iran would also have negative spillover effects. First and foremost, short of an accompanying ground invasion and occupation, an air attack could not guarantee an end to Iran’s pursuit of nuclear weapons; it could guarantee only a delay and would almost certainly drive the Iranians to commit more fervently to building a bomb. If Iran’s capabilities were only temporarily degraded but its intentions were inflamed, the threat might become worse. Striking first would also fracture the international coalition that now stands behind sanctions against Iran, undercut opposition to the regime inside the country, and be seen throughout the world as another case of arrogant American aggression against Muslims.

Those costs might seem justifiable if launching a war against Iran dissuaded other countries from attempting to get their own nuclear deterrents. But it might just as well energize such efforts. George W. Bush’s war to prevent Iraq from getting nuclear weapons did not dissuade North Korea, which went on to test its own weapons a few years later, nor did it turn Iran away. It may have induced Libyan leader Muammar al-Qaddafi to surrender his nuclear program, but a few years later, his reward from Washington turned out to be overthrow and death — hardly an encouraging lesson for U.S. adversaries about the wisdom of renouncing nuclear weapons.

One reason U.S. leaders might be reluctant to apply deterrence these days is that the strategy’s most potent form — the threat to annihilate an enemy’s economy and population in retaliation — is no longer deemed legitimate [...] But that inhibition should hardly be a reason to prefer starting a war, nor does it cripple deterrence. An acceptable variant would be to threaten not to annihilate Iran’s population but to annihilate its regime — the leaders, security agencies, and assets of the Iranian government — if it used nuclear weapons. Although in practice, even a discriminating counterattack of that kind would result in plenty of collateral damage, U.S. planners could credibly make the threat and could reinforce it by pledging to invade Iran as well — a step that would be far more reasonable to take after an Iranian nuclear strike than it was against Iraq in 2003. And even if legal concerns constrained the United States from massively retaliating against Iranian civilians, Israeli leaders would surely be willing to do so if Iran attacked Israel with nuclear weapons, since Israel’s national existence would be at stake. Those mutually reinforcing threats — that the fruits of the Iranian Revolution and even Iranian society itself would cease to exist — would be an overwhelming restraint on Tehran.

A nuclear-armed Iran is an alarming prospect. But there is no sure solution to some dangers, and this challenge presents a strategic choice between different risks. There is simply no real evidence that war with Iran would yield any more safety than handling the problem with good old deterrence.

Fine as far as it goes, but it misses what is now the more relevant debate over what will happen if Iran stays below the American red line in perpetuity.

Last week, The Economist reported that “[Obama] has told [Netanyahu] that America is now closer to the threshold for taking such a step” – that step being “an ultimatum to Iran that it must reach a deal on halting enrichment within months or face military action” – and “that he [Obama] is not prepared to allow diplomatic negotiations to run beyond this year without a big change in Iran’s attitude”.

If that is so, and I am somewhat sceptical that Obama would have been so committal, then we run into a problem identified by Betts earlier in his essay:  ”the deterrent warning must be loud and clear, so the target cannot misread it. Deterrence should be ambiguous only if it is a bluff”.

Iran is unlikely to take any obvious steps towards weaponisation – say, enrichment beyond 20% or the expulsion of inspectors. So what will that ultimatum look like? There will be considerable ambiguity over where other red lines should be drawn if the talks fail and Iran continues its nuclear advances: introduction of IR2s at Fordow? A certain number of cascades?

This ambiguity may even be realistically irresolvable. Setting a specific date, as Israel wants and The Economist suggests might have occurred, is deeply problematic: true, it is unambiguous (2014 comes when it comes), but it is also entirely non-credible, because Iran could temporarily trim down capabilities, undercut the objective basis for any military action, and slowly build back up to those levels.

More likely is a Schelling-esque approach: that the trigger for any future red line(s) is unambiguous (a specific date) but that the response is left ambiguous. But this still leaves open the possibility that, as Betts warns, Iran interprets this as a bluff.

Iranian nuclear weapons: dubious reporting, dubious red lines

In an interesting if thinly sourced piece on how Iran is withdrawing IRGC special forces from Syria, The Sunday Times threw in this nugget at the end, as if it were no big deal:

Western intelligence officials see the next few months as a race between Iran’s programme to build a nuclear bomb and sanctions intended to ruin its economy.

This is a remarkable and, if untrue, irresponsible claim.

Both the United States and Israel continue to believe that Iran has not yet decided to build a nuclear bomb (which is distinct from conducting activities relevant to the fabrication of a bomb). This is a clear and widely-sourced claim. See Jeffrey Lewis here, Micah Zenko here, the chief of staff of the IDF here, Haaretz reporting here, and DNI James Clapper, here (“we don’t believe they have actually made the decision to go ahead with a nuclear weapon”).

There is a debate over when Iran would acquire enough uranium enriched to 20% required for one bomb, and whether and how quickly Iran could securely and/or secretly further enrich that to weapons-grade. But these debates are are all hypotheticals. In fact, Iran’s conversion of 20% enriched uranium to fuel plates actually makes it harder and slower to produce a bomb (for some strange reason, the media has only picked up this story a few days ago despite it being obvious from the IAEA’s August report).

The Sunday Times is saying is that Iran is racing for a bomb, which directly contradicts this very large and diverse mass of reporting and analysis.

My guess is that they did not mean to make this claim. But it is staggeringly poor journalism, and a very dangerous claim in the current environment. American newspapers were chastened by their experience of interpreting and reporting on WMD intelligence prior to the Iraq War, and have responded impressively and maturely. British newspapers – with a few honourable exceptions – have shown poor judgment, repeatedly made alarmist claims, and failed to convey the nuances of this technically complex subject.

Moreover, if Iran were racing for a bomb, which it is very obviously not, then the constraint would not be sanctions. Sanctions, however severe, are not going to bring down the regime in the space of the (at most) eight month period in which Iran could, if it wanted, have more than adequate 20% enriched uranium for one or more bombs. The constraint is instead  the risk of detection and vulnerability during breakout. A dash for a bomb can easily outrun sanctions, which operate only over a longer period. Saddam hung on for 12 years, even as sanctions inflicted a devastating toll on civilians (“near apocalyptic” as early as 1991). But it would still run into the other obstacles.

***

Red lines:

On a related note: as it’s a day of the week ending in “y”, Israel has thrown out a new red line for Iran’s nuclear programme. Netanyahu’s UNGA speech appeared to place his desired red line at the point at which Iran obtains enough 20% enriched uranium for one bomb, a point he estimated would arrive during the spring or summer of 2013 (this seems reasonable, but it ignores the possibility that Iran will keep converting some to fuel). This Netanyahu line differs from the presumptive American red line(s) insofar as Iran would cross this one even before making any decision to enrich beyond 20%.

Now,  according to Special Envoy for the Israeli Prime Minister Zalman Shoval, “the red line for Israel is when the Iranians have produced enough fissionable material from which they can produce at least a dirty bomb within a short time”.

This is an even more incoherent red line than many of those suggested earlier, for the simple reason that enriched uranium is not used in dirty bombs:  ”these materials, perhaps counterintuitively, are not radioactive enough. Their radioactive emissions don’t travel far and are blocked by simple barriers, including skin and clothing. A dirty bomb would use small amounts of highly radioactive materials such as cesium or cobalt, not uranium”. If Iran did use uranium in a dirty bomb, it wold need “some 1,400 tons of uranium – when all Iran has now is 6 tons total”.

Israel, Iran and the Entebbe/Rambo option

It turns out Centcom hasn’t the slightest idea how Israel might strike Iran’s nuclear programme. And where facts are light, the imagination takes over:

In this scenario, the Israelis would forego a massed air attack and instead mount a high-risk but high-payoff commando raid that would land an elite Sayeret Matkal (special forces) unit outside of Iran’s enrichment facility at Fordow, near Qom. The unit — or other elite units like it — consisting of perhaps as many as 400 soldiers, would seize Iran’s enriched uranium for transport to Israel.

The operation’s success would depend on speed, secrecy, simplicity, and the credibility of Israeli intelligence [...]

The Israeli unit would be transported on as few as three and perhaps as many as six C-130 aircraft (which can carry a maximum of 70 troops) that would be protected by a “swarm” of well-armed F16Is, according to the scenario being considered by U.S. military officers. The C-130s would land in the desert near Fordow. The Israeli commandos would then defeat the heavily armed security personnel at the complex, penetrate its barriers and interdict any enemy units nearby, and seize the complex’s uranium for transport back to Israel. Prior to its departure, the commando unit would destroy the complex, obviating the need for any high-level bombing attack.

As it happens, the US has some people who know a thing or two about Entebbe:

In the mid-1990s, William McRaven, then a U.S. Navy SEAL [now commander of U.S. Special Operations Command], wrote a book about commando operations. Entitled “Spec Ops: Case Studies in Special Operations Warfare: Theory and Practice” (Presidio Press), the book featured six case studies. One chapter was devoted to Entebbe, beginning with the lessons learned in the Israel Defense Forces as a whole, and in the Sayeret Matkal special operations unit in particular, after the failure to save the lives of 25 hostages in Ma’alot two years earlier. It included a discussion of Israeli intelligence gathering, decision-making processes, creation of the command and control system, personnel conflicts and the actual rescue operation in Entebbe Airport in Uganda, on July 4, 1976.

Except, here’s the thing: Entebbe Airport (and Bin Laden’s Abbotabad compound, for that matter) were not designed to detect and resist a major armed assault. Entebbe’s was an ordinary civilian airport, and Bin Laden’s compound was protected through secrecy, not fortification.

Neither of these facilities was 26 feet underground (the main enrichment hall at Fordow is under 295 feet of rock). Bin Laden’s compound was about 30,000 square feet, whereas Fordow is about 270,000 square feet.

And neither the Israeli units at Entebbe nor the SEALs at Abbotabad had to obliterate a huge facility on their way out, or take ~100kg of reactor-grade uranium out with them.

Even if special forces may be used for intelligence gathering (as Israel did before the 2007 strike on Syria’s reactor), this scenario just doesn’t seem plausible.

Oh, and then there’s this:

That said, “In some scenarios,” the U.S. military planner who told me of the potential operation said, “there would be very high Israeli casualties because of nearby Republican Guard divisions. This operation could be quite bloody.

Iran, which does not have a Republican Guard, will be alarmed to learn that Saddam-era shock troops are still hanging around on Iranian soil, all those years later.

What is the zone of immunity?

In the zone

During the Iranian nuclear crisis, the concept of a “zone of immunity” has been thrown around a lot. Take, for instance, Ronen Bergman’s widely-read January piece in the NYT:

[Ehud Barak] warned that no more than one year remains to stop Iran from obtaining nuclear weaponry. This is because it is close to entering its “immunity zone” — a term coined by Barak that refers to the point when Iran’s accumulated know-how, raw materials, experience and equipment (as well as the distribution of materials among its underground facilities) — will be such that an attack could not derail the nuclear project.

Yet, it’s not clear that the concept is quite coherent.

Derail and destroy

First of all, consider that “derail” has no specific meaning. If it means “terminate”, then Iran is already in the zone of immunity. As General Martin Dempsey explained the other day, an Israeli attack would “clearly delay but probably not destroy Iran’s nuclear programme”. The Bush-era CIA director, Michael Hayden, said much the same thing in January: “they [Israel] can’t do it, it’s beyond their capacity”. So what length of delay would constitute a derailment? Without answering this, the zone of immunity blurs into a nebulous concept with little analytical utility.

The second question is why the US or Israeli ability to “derail the nuclear project” should be declining over time. The answers to this usually pertain to three, frequently conflated, activities:

1. The hardening of Iran’s underground Fordow enrichment facility
2. The accumulation of enriched uranium, and particularly that enriched to 20 percent i.e., nine-tenths of the way to weapons-grade
3. The increase in the number of installed centrifuges (e.g. at Fordow, over 1,000 added since May) and therefore latent enrichment capacity

The first of these – the hardening of Fordow – obviously increases Iran’s ability to withstand attack. It is frequently noted that the zone of immunity comes later for the US than for Israel, because of the former’s greater military capabilities. This reinforces the idea that the primary determinant of a zone of immunity is the physical fortification of Fordow. Yet there are vanishingly few news reports which show Iran is actively increasing Fordow’s defenses. Why, then, should Iran be any closer to the zone of immunity today than it was months ago?

This is where the second and third criteria come in. It is usually reports connected to enriched uranium stockpiles and centrifuge installation that drive concerns over the zone of immunity. But things are less clear in this regard. An increase in nuclear materials and installed centrifuges does shorten Iran’s breakout timeline i.e., the time it would take Iran to produce a bomb were the decision to be taken to do so. But what is the relationship between the breakout timeline and a zone of immunity?

Breakout timeline and the zone of immunity

The first argument goes as follows: if an attack only destroys a certain proportion of Fordow’s contents, then more pre-attack materials and centrifuges might mean more surviving, post-attack quantities with which to break out.

This may very well be a factor in assessing immunity, but it would seem to require an unrealistic degree of foresight and precision. It seems reasonable to postulate that any attack would probably have to be premised on an ability to neutralise almost all of Fordow’s contents, not just a middling proportion. Otherwise, Israel would be at high risk of engendering a bombed, embittered Iran with sufficient (or near-sufficient) short-term capability to produce weapons-grade uranium. And if that’s so, does it make a substantial difference whether Iran has 1,000 or 2,000 centrifuges?

The second argument is this: a shorter breakout timeline might allow Iran to pre-empt (note: not withstand) an attack, either by diverting LEU to a clandestine site for further enrichment or by enriching existing LEU at Fordow to weapons-grade, and then diverting the resultant fissile material.

In each of these cases, Iran could only be said to be “immune” from attack if it could undertake and complete these steps before being caught and attacked. Yet, as David Albright explains in reference to the latter scenario: “if only Fordow is used, breakout times are expected to exceed two months as long as the plant utilizes IR-1 centrifuges and current cascade designs”. According to the latest IAEA report, dated 30 August, Iran is indeed only using IR-1 centrifuges (see p5).

Moreover, Iran has taken steps that likely lengthen its breakout timeline. It is Iran’s stockpile of 20% enriched uranium - material that is nine-tenths of the way to weapons-grade – that would be necessary for the most rapid breakout possible. Yet Iran has converted or is converting 97.9kg of this, over half its entire stockpile of the stuff, for fuel plates for the Tehran Research Reactor (IAEA report, p.13). That makes it much harder (and, after irradiation, harder still) to use for weapons purposes. That leaves Iran with just over 91kg, which is not enough for even one bomb. The overarching point is that a timeline measured in months, rather than weeks, is far in excess of the time that it would plausibly take for detection.

That will change over time, particularly if Iran stops diverting 20% enriched uranium for fuel plates and starts using more of its latent enrichment capacity. But, for now, the time it’d take Iran to actually use its spare enrichment capacity to make weapons-grade uranium is far in excess of the time it would take the IAEA (and intelligence agencies) to detect this, and respond (which is not to say that any armed response would be successful over the longer-term).

And all this is without factoring in the considerable amount of time it would take Iran to convert fissile material into a deliverable nuclear device: “Iran would need an estimated 8-12 months to build a crude nuclear explosive that could be tested underground or delivered by non-missile systems”, and even longer for putting a warhead on a ballistic missile (see this, p2).

Why, then, has the New York Times reported on the new IAEA findings by declaring that “Iran is close to crossing what Israel has long said is its red line: the capability to produce nuclear weapons in a location invulnerable to Israeli attack”? Unless we define “close” with much more precision, this seems to be an unnecessarily alarmist judgment. Unless specified with greater detail, the zone of immunity should be understood as a rhetorical device rather than a clearly defined condition.

Final note:

The Wall Street Journal reports that that Iran’s Mohsen Fakhrizadeh, head of Iran’s alleged clandestine weapons programme until 2003 (outlined in some detail in the IAEA’s November report)  is back at work in the northern suburbs of Tehran. The story doesn’t clarify whether Fakhrizadeh is adding much to the activities of what became, after 2003, a “dormant” programme of less sensitive, dual-use activities. Iran, in its typically helpful fashion, has repeatedly blocked the IAEA from interviewing him.

Fakhrizadeh’s activities could be reasonably categorised under  ”accumulated know-how”, as Bergman puts it in that NYT piece referenced above. To the extent that these activities accumulate know-how and  shorten the time it would take to eventually fabricate a bomb from fissile material, this does reduce the breakout time. If the breakout time is eventually shortened below the time it takes to detect and respond to breakout, then clandestine research would be taking Iran into the zone of immunity.

But there’s no evidence suggesting that the breakout time is being shortened to that degree. Moreover, the fact that Fakhrizadeh’s return has been clocked (even if you add the usual caveats about judging intelligence leaked by interested parties) perhaps indicates just how difficult it would be for Iran to break out without being detected, given the scale of intelligence efforts against its programme.

More thoughts on Israeli airstrikes and IAEA inspectors

Mark Hibbs has a terrific post at Arms Control Wonk, asking whether IAEA inspectors in Iran are at risk of harm from Israeli or American airstrikes:

Were the IAEA to want to carry out a short-notice inspection at Fordow or Natanz, it wouldn’t tip its hand in advance and would want to be prepared. That would imply that, in order to reap the advantage of surprise, inspectors would have to be pretty close to the enrichment plants at the time Iran would be informed for the inspection to have any value. If inspectors were, say, in Tehran, it would take several hours for them to truck out to Natanz. The farther away they are, the less surprise.

We might conjecture that, absent perfect foresight or guidance, if the IAEA takes seriously the threat that Israel sometime during the rest of this year might launch surprise air strikes against Iran’s enrichment plants, it would not want to put its inspectors at risk by requesting from Iran short-notice inspections at those installations. If the IAEA were to undertake an inspection at Fordow or Natanz, it might move safeguards personnel into locations which were about to be bombed in an airstrike the IAEA didn’t know was about to happen.

There are also some excellent comments under the post, including one from former IAEA inspector Robert Kelley:

In June 1993 I was awakened by the sound of cruise missiles striking Baghdad while we of IAEA Action Team mission 21 were sleeping in the Sheraton hotel. The noise was caused by Bill Clinton’s cruise missiles striking the Iraqi intelligence headquarters in retribution for the alleged plot to kill George Bush 41 on a visit to Kuwait. Several Iraqi civilians were killed. Naturally I called Vienna as chief inspector and reported our situation to Maurizio Ziffereo, the Action Team Leader. He called the US representative to the IAEA and woke her up in the middle of the night to tell her that IAEA inspectors with three Americans on the team were in harm’s way. She replied that she had no idea of the attack and no way of warning us.

Inspectors are simply pawns in any great war game and their lives will be measured against the perceived good. It goes with the job that you may get a free tombstone.

Kelley points out that, not only might Israel not give a fig about killing IAEA personnel, but that, “in the current case Iran bars most western inspectors (US, UK, France, Germany etc.) so an attacker would know that the inspectors at risk are mostly from non-aligned states and not allies”. Israel isn’t going to want to wipe out a whole team of inspectors, but it sure as hell isn’t going to put this high up the list of mission priorities.

Kelley also observes that inspectors are at risk of hostage-taking or retribution. He notes that, after cruise missile strikes in the 1990s, ”Iraq identified inspection information as being the source of the very accurate targeting of equipment that was only known from the inside of buildings”. It seems not unlikely that intelligence gleaned from inspections may be used as targeting information in any strike on Iran – and, even if it’s not, that Iran may take this view, and react accordingly.

I had three thoughts in response to the post.

First, would a night-time strike mitigate these risks? Maybe, though not necessarily. Although inspections do seem to be concentrated in daylight hours, inspectors can sometimes work late at Iranian facilities, start early, and – as you would expect – keep an unpredictable schedule. Recall, also, that in 1981 the first Israeli bombs on Iraq’s Osirak reactor were released at 6.35pm.

Second, why should short-notice inspections put inspectors at greater risk than those announced a week beforehand? This isn’t clear to me. One argument is that short-notice inspections require that inspectors reside closer to the facility in question, so that excessive travel time doesn’t negate the surprise. Inspectors’ proximity to sites therefore puts them at risk. But this seems to contain an assumption that airstrikes will have a damage radius of two-hours travel time from a target. Even though airstrikes would surely extend well beyond the boundaries of a suspect site, to encompass things like nearby power supplies and potential support sites, a radius of two-hours travel time seems really quite large.

Third, does this problem wash out if you assume – not unreasonably – that Israel has access to the inspections schedules of IAEA inspectors in Iran? While short-notice inspections may require pre-positioning inspectors uncomfortably close to sites, the decision to pre-position them to that end could have been taken sometime earlier, in a way that allows would-be bombers to take care with their targeting. Of course, this logic is more problematic if (1) the IAEA is worried about Iranian penetration of parts of their inspections team or (2) if it wishes to flexibly react to new bits of intelligence from Member States, both of which would preclude privately pre-set inspections schedules.

 

Iran and the Mujahedin e Khalq (MEK)

Writing in the London Review of Books, Owen Bennett-Jones has a nice overview of the Iranian terrorist group MEK. He argues, quite persuasively, that the United States is making the same mistake with the MEK as it did with the Iraqi National Congress a decade ago, “demonstrat[ing] that once again it has failed to learn its lesson”.

The story of the People’s Mujahedin of Iran, also known as the Mujahedin e Khalq (MEK), is all about the way image management can enable a diehard enemy to become a cherished ally. The MEK is currently campaigning to be officially delisted in the US as a terrorist organisation. Once off the list it will be free to make use of its support on Capitol Hill in order to become America’s most favoured, and no doubt best funded, Iranian opposition group.

Jones’ essay is a review of a new book:

Raymond Tanter’s book is part of the MEK’s image management campaign, a briefing document for advocates of delisting. Tanter, a long-time supporter of the group, has produced a compact guide, complete with colour pictures and transcripts of speeches by paid MEK advocates

Back in April, Seymour Hersh reported in the New Yorker that the United States had recently trained MEK member at a Department of Energy facility in Nevada, and that the group was responsible for killing Iranian nuclear scientists.

On the history of MEK’s links to the West:

The M.E.K.’s ties with Western intelligence deepened after the fall of the Iraqi regime in 2003, and JSOC [Joint Special Operations Command] began operating inside Iran in an effort to substantiate the Bush Administration’s fears that Iran was building the bomb at one or more secret underground locations. Funds were covertly passed to a number of dissident organizations, for intelligence collection and, ultimately, for anti-regime terrorist activities. Directly, or indirectly, the M.E.K. ended up with resources like arms and intelligence. Some American-supported covert operations continue in Iran today, according to past and present intelligence officials and military consultants.

The training was, allegedly, extensive:

The training ended sometime before President Obama took office, the former official said. In a separate interview, a retired four-star general, who has advised the Bush and Obama Administrations on national-security issues, said that he had been privately briefed in 2005 about the training of Iranians associated with the M.E.K. in Nevada by an American involved in the program. They got “the standard training,” he said, “in commo, crypto [cryptography], small-unit tactics, and weaponry—that went on for six months,” the retired general said. “They were kept in little pods.” He also was told, he said, that the men doing the training were from JSOC, which, by 2005, had become a major instrument in the Bush Administration’s global war on terror. “The JSOC trainers were not front-line guys who had been in the field, but second- and third-tier guys—trainers and the like—and they started going off the reservation. ‘If we’re going to teach you tactics, let me show you some really sexy stuff…’ ”

The MEK was a useful conduit for Iranian intercepts:

Khodabandeh told me that he had heard from more recent defectors about the training in Nevada. He was told that the communications training in Nevada involved more than teaching how to keep in contact during attacks—it also involved communication intercepts. The United States, he said, at one point found a way to penetrate some major Iranian communications systems. At the time, he said, the U.S. provided M.E.K. operatives with the ability to intercept telephone calls and text messages inside Iran—which M.E.K. operatives translated and shared with American signals intelligence experts. He does not know whether this activity is ongoing.

And it worked:

“The M.E.K. was a total joke,” the senior Pentagon consultant said, “and now it’s a real network inside Iran. How did the M.E.K. get so much more efficient?” he asked rhetorically. “Part of it is the training in Nevada. Part of it is logistical support in Kurdistan, and part of it is inside Iran. M.E.K. now has a capacity for efficient operations that it never had before.

On MEK’s role in killing Iranian scientists and engaging in sabotage:

[E]arly last month NBC News quoted two senior Obama Administration officials as confirming that the attacks were carried out by M.E.K. units that were financed and trained by Mossad, the Israeli secret service. NBC further quoted the Administration officials as denying any American involvement in the M.E.K. activities. The former senior intelligence official I spoke with seconded the NBC report that the Israelis were working with the M.E.K., adding that the operations benefitted from American intelligence. He said that the targets were not “Einsteins”; “The goal is to affect Iranian psychology and morale,” he said, and to “demoralize the whole system—nuclear delivery vehicles, nuclear enrichment facilities, power plants.” Attacks have also been carried out on pipelines. He added that the operations are “primarily being done by M.E.K. through liaison with the Israelis, but the United States is now providing the intelligence.” An adviser to the special-operations community told me that the links between the United States and M.E.K. activities inside Iran had been long-standing. “Everything being done inside Iran now is being done with surrogates,” he said.

Also see: Owen Bennett-Jones’ From Our Own Correspondent on the MEK; a 2009 RAND study [PDF] on the MEK; Foreign Policy on the MEK’s supporters in Congress; the New York Times on the same issue; and a defense of the group from the Weekly Standard.

Iran on the brink (of what?)

Note: post amended with a correction at the bottom

Dan Senor, a foreign policy adviser to Mitt Romney, has declared that “”Iran is on the brink of getting a nuclear weapon”.

This is hardly unprecedented language. A few months ago, The Times (of London, that is) splashed “Iran close to Bomb after nuclear breakthrough” on its front page. That headline was changed for the online edition – but it still pronounced Iran “one step away from the Bomb”. The Telegraph was just as injudicious: “Iran on brink of nuclear weapon, warns watchdog”. Ditto The Daily Beast, which also put Iran “on the brink of nuclear weapons”.

This language might not surprise many people. After all, the proverbial “brink” is becoming a bit of a foreign policy joke. Yemen and Pakistan both seem to approach it asymptotically. But it does convey the requisite urgency and alarm.

Danielle Pletka, vice-president at the American Enterprise Institute (AEI), argued that though Senor’s claim may be wrong, Iran was “on the brink of a weapon’s capability”. She explained as follows:

Iran has 73+ kg of 20% LEU, needs 85[kg] for a bomb. [M]aking the bomb’s easy. I call that “brink”.

These figures don’t appear to be correct.

This seems like a good time to quote three chunks of an Arms Control Wonk post from two months ago:

All too often, pundits make serious mistakes when converting a real measurement like kilograms into the more evocative “bombs worth.” Few pundits make clear their assumptions on how much nuclear material is needed for a weapon in which context. Invariably, the result is to exaggerate either the danger faced or averted. The problem is especially bad when the topic is an amount of highly enriched uranium.

The concept of “significant quantity” of uranium is important in this regard:

IAEA safeguards standards define a significant quantity as “the approximate amount of nuclear material for which the possibility of manufacturing a nuclear explosive device cannot be excluded. Significant quantities take into account unavoidable losses due to conversion and manufacturing processes and should not be confused with critical masses.” And the current IAEA glossary defines that threshold amount for HEU as 25kg of U-235 in HEU.    (U235 is the stuff that goes boom.)

So the claim that Iran needs 85kg of 20% LEU is clearly untenable. [1]  In fact, even

[o]ne hundred kilograms of 20 percent HEU contains 20 kg of U-235 — less than one significant quantity.

How much does Iran actually have, and how much does it need? On the basis of the February IAEA report, Daryl Kimball estimates that

[a]lthough Iran has now produced about 110 kilograms of 20% enriched uranium, it has dedicated 8 kilograms to make fuel assemblies for the TRR. That material would no longer be part of a ready stockpile of 20% enriched uranium that can be rapidly converted to weapons-grade. Iran would need at least 120 kilograms of 20% material in order to make enough weapons grade uranium for a single weapon. It would want sufficient material for more than just one weapon if it were to decide to produce them. [emphasis added]

Some estimates are more conservative. Iran Watch, part of the Wisconsin Project on Nuclear Arms Control, estimates the quantity “to produce a bomb’s worth of weapon-grade uranium metal” to be higher, at 140 kilograms.

Others are more generous. The ScienceWonk blog at the Federation of American Scientists (FAS) judges that:

A hundred kilograms of uranium enriched to 20% U-235 [just under what Iran is estimated to possess at the moment] will have about 20kg of U-235. But a fissionable mass of 94% pure U-235 weighs about 16kg so Iran might have enough uranium with further enrichment to the weapons-grade level to make a single nuclear weapon – maybe two. [emphasis added]

The AEI’s own blog says the following:

If Iran required 25 kg of weapons-grade uranium to fuel one bomb, it would take 4.7 months after September when it is projected to have 141 kg uranium enriched to 20%. [emphasis added]

***

In fact, there are three distinct questions about timelines.

The first is how much time it would take Iran to enrich a “significant quantity” of uranium to 20% (not long).

The second is how long it would take Iran to enrich that stockpile further to weapons-grade (obviously, longer). The answers to both of these questions depend on a complex set of assumptions about how and where they conducted the enrichment.

The third question is how long it would then take to fabricate a nuclear device, if they chose to do so (and, it bears repeating, that they are not thought to have made any such decision). There is a range of views here:

Greg Jones claims it’s easy:

the viewpoint that it will take Iran years to develop the non-nuclear components required for a nuclear weapon is hard to square with the actual historical experience of the nuclear weapon states. It is well-known that for past nuclear weapon programs, the key impediment was the need to acquire the fissile material (HEU or plutonium) for the weapon.

Jacques Hymans, writing in Foreign Affairs, says it’s hard:

The Iranians had to work for 25 years just to start accumulating uranium enriched to 20 percent, which is not even weapons grade. The slow pace of Iranian nuclear progress to date strongly suggests that Iran could still need a very long time to actually build a bomb — or could even ultimately fail to do so. Indeed, global trends in proliferation suggest that either of those outcomes might be more likely than Iranian success in the near future.

This is because of “Iran’s long-standing authoritarian management culture”:

In a study of Iranian human-resource practices, the management analysts Pari Namazie and Monir Tayeb concluded that the Iranian regime has historically shown a marked preference for political loyalty over professional qualifications. “The belief,” they wrote, “is that a loyal person can learn new skills, but it is much more difficult to teach loyalty to a skilled person.” This is the classic attitude of authoritarian managers. And according to the Iranian political scientist Hossein Bashiriyeh, in recent years, Iran’s “irregular and erratic economic policies and practices, political nepotism and general mismanagement” have greatly accelerated. It is hard to imagine that the politically charged Iranian nuclear program is sheltered from these tendencies.

Further reading: a presentation (PDF) by Maseh Zarif for AEI, some questions by ISIS about the intelligence on the Iranian program, former ambassador Thomas Pickering writing on negotiations, a wrap-up of Senate testimony on Iran, and a useful NYT timeline of US-Iran interaction on nuclear issues.

***

[1] Note, however: “Tehran claims it has enriched to only 19.75%, thereby avoiding the 20% level, which is notionally the divide between low-enriched uranium and HEU” [Op-Ed, Olli Heinonen and Simon Henderson]

***

Update/correction: Maseh Zarif correctly notes that I confused solid and gas figures for enriched uranium, and that I omitted to note that a lower-bound for one bombs-worth of enriched uranium was as little as 15kg (much lower than the IAEA’s 25kg “significant quantity).

Bombing Iran (peer-reviewed edition)

The subject of an Israeli or American attack on Iran has been debated and discussed with unusual intensity over the past few months. Only a few of these accounts have taken much stock of the academic literature pertaining to the issue (one exception, from Colin Kahl, here). Here are a few relevant studies.

First, consider the oft-cited precedent of Israel’s bombing of Iraq’s Osirak reactor in 1981. In last summer’s International Security, Målfrid Braut-Hegghammer, an Assistant Professor at the Norwegian Defence University College, published Revisiting Osirak: Preventive Attacks and Nuclear Proliferation Risks. It has a fairly simple argument:

The Israeli attack had mixed effects: it triggered a nuclear weapons program where one did not previously exist, while forcing Iraq to pursue a more difficult and time-consuming technological route. Despite these challenges and added delays resulting from inefficient management, within a decade Iraq stood on the threshold of a nuclear weapons capability. Ultimately, I conclude that the Israeli attack was counterproductive. (p102)

The Iraqi programme in the 1970s was “directionless”:

Iraq’s efforts to move toward a nuclear weapons capability during the late 1970s were informal and incremental. They lacked the institutional foundations and dedicated resources that constitute a nuclear weapons program in any meaningful sense of the word …  Iraq’s nuclear efforts during this period can be characterized as a form of “drift”—an exploration of the technical foundations for a nuclear weapons program without an explicit political mandate guiding these efforts. (p109).

But then, the Israeli attack on Osirak

effectively forged an alliance between Iraqi nuclear entrepreneurs and the Iraqi leadership. This alliance produced a more determined and organized effort to acquire a weapons capability … Saddam’s decision to start the program in September 1981 came with the offer of a “blank check”: in other words, abundant and consistent funding. From 1983 until 1991, the program’s staff increased by 60 percent annually. According to Jafar, the Iraqi nuclear establishment spent 792,899,913 Iraqi dinars on the weapons program from 1982 to 1988, and an additional 669,446,170 dinars during 1989–90. (p117)

That Iraq took so long to get a weapon (and was overtaken by the First Gulf War – something that just isn’t going to happen to Iran) was down to its own stupidity rather than the bombing:

 It seems unlikely—but not impossible—that another targeted state would make as many ill-advised decisions in their efforts to develop a nuclear weapons capability following an attack. (p130)

Several years ago, Dan Reiter, also a political scientist, had reached a similar conclusion [PDF] about Osirak:

It must therefore serve as a cautionary note for future endeavors. Indeed, even the limited successes at Osiraq are unlikely to be repeated, as many states (including Iraq, North Korea, and Iran) learned the lessons of Osiraq and after 1981 sought to disperse and conceal their nuclear facilities, making future raids even less likely to succeed. (p365)

The second paper, Attacking the Atom: Does Bombing Nuclear Facilities Affect Proliferation? by Sarah E. Kreps and Matthew Fuhrmann, published in the Journal of Strategic Studies last April, takes a broader look at the question:

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 counterproductive” downplays evidence of prior strikes that delayed the target state’snuclear 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 inthe 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. (p163)

But, on the specific question of Iran, the authors are skeptical:

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. (p183)

The third and final paper is Targeting Nuclear Programs in War and Peace: A Quantitative Empirical Analysis, 1941–2000 published a couple of years ago by the same authors, Fuhrmann and Kreps, in the Journal of Conflict Resolution. It comes at this from a different angle, asking, “When do states attack or consider attacking nuclear infrastructure in nonnuclear weapons states?”. They find that

states are likely to attack or consider attacking nuclear facilities when they are highly threatened by a particular country’s acquisition of nuclear weapons. Three factors increase the salience of the proliferation threat: (1) prior violent militarized conflict; (2) the presence of a highly autocratic proliferator; and (3) divergent foreign policy interests. (p1)

Such attacks are more common than you might think:

In 1942, British commandos launched an attack against a suspected nuclear facility, targeting the Norsk-Hydro heavy water plant in German-occupied Norway. This raid represents the first use of military force to hinder nuclear proliferation, but it was not the only time that a nuclear program was targeted in the twentieth century. New data collected for this article reveal that fifteen separate attacks against nuclear facilities occurred between 1942 and 2000 and attacks were seriously considered on fifty separate occasions during this period. (p2)

And some interesting examples:

[B]etween 1979 and 1987, Israel requested cooperation from India in attacking Pakistani nuclear facilities because operational success depended on the use of bases in India for launching and refueling … Countries may also request cooperation from another state to blunt the potential consequences of attacks. The United States requested assistance from the Soviet Union in attacking Chinese nuclear installations during the early 1960s because Washington hoped that Moscow’s involvement would deter a violent response from Peking and limit the further deterioration of East–West relations. (p5)

And:

 Not only did Taiwanese officials discuss the possibility of raiding Beijing’s key nuclear facilities but at least one senior official privately advocated for military action. During a visit to Washington in September 1963, General Chiang Ching-kuo—Chiang Kai-shek’s son—lobbied for strikes against nuclear Chinese nuclear facilities in private meetings with U.S. officials, including President John F. Kennedy and National Security Advisor McGeorge Bundy. (p5)

See also Osirak Redux? Assessing Israeli Capabilities to Destroy Iranian Nuclear Facilities in International Security (though this was written in 2007, before the hardened facility near Qom was revealed).

Iranian retaliation and Arab oil

Gary Sick’s writings on Iran have (rightly) been getting a lot of attention recently. In his most recent piece, he warned that, in the event of an Israeli or American attack on Iran,

one might expect disruptions in oil delivery and loading in Arab ports up and down the Gulf, some because of sabotage but others from cyberattacks on the control systems. Iran would attribute these to “the hand of God,” but the more pragmatic effect would be a very substantial portion of the world’s oil suddenly removed from world supply.

That such an attack would sharply drive up oil prices, possibly by a sufficient degree to induce another recession Western economies, seems a reasonable argument. But what are the prospects of Arab oil going offline?

A Crude Threat: The Limits of an Iranian Missile against Saudi Arabian Oil [PDF], a paper published last summer in International Security by Miranda Priebe and Joshua R. Itzkowitz Shifrinson (both of MIT), offers a partial answer.

What would Iran want to attack?

Iran is unlikely to attack Saudi oil fields. To stop production at an oil field, Iran would have to destroy the wells. This would require the destruction of many small targets spread over a large area … Nor is Iran likely to target pipelines, given their small size and ease of repair. Not only are there more than 15,000km of pipeline in the country, but Aramco has taken steps to minimize the effects of pipeline damage … Saudi Arabia’s refined products constitute a small percentage of Saudi exports and are comparatively less important as a percentage of global oil supplies compared with crude oil. We conclude that Iran would target Saudi Arabian stabilization facilities … each stabilization facility is within 300km of Iran and thus within range of most Iranian missiles. As a next-best option, Iran might try to prevent oil from reaching the market by attacking the Saudi export system. (pp177-178)

Stabilization facilities remove natural gas components in crude oil. Saudi Arabia’s main stabilization facility is at Abqaiq (it was subjected to a terrorist attack in 2006). The article finds that:

With current assumptions, more than 1,300 Shahab-type missiles would be needed to target Abqaiq’s towers. With the 400 missiles on hand, Iran would be unlikely to do significant damage. Increasing the desired probability of success raises missile requirements: for example, a 50 percent overall probability of destroying Abqaiq’s towers would require more than 3,300 missiles. Moreover, even if Abqaiq was destroyed, Saudi Arabia would still be able to produce and stabilize 5.6 [million barrels per day] of oil. Therefore, even if Iran has many times the number of missiles we estimate, a significant portion of Saudi Arabian oil is secure. (p192)

In other words, Iran needs more than three times the number of missiles it has just to get 75 percent confidence of destroying any one tower. What if it fired all of its missiles at a single tower?

 Damage would be minimal, because 400 missiles are sufficient to destroy only one tower with a 60 percent chance of success. If the tower were destroyed, Abqaiq’s total capacity would drop from 13 to 12.3 [million barrels per day]. Because, however, Abqaiq runs below half capacity, its destruction would have no long-term impact on Saudi Arabia’s ability to stabilize oil. The facility could still handle its 6.1 [million barrels per day] throughput with capacity to spare. (p193)

Pretty feeble stuff. How long would it take to repair?

If all 1.4 billion barrels in [worldwide] government reserves were employed, the repair window would be nearly fifteen months. A six-to-fifteen-month window is in the midrange of past repair experiences [e.g. after the Gulf War], suggesting that world oil consumption would not be impaired even after a successful Iranian attack. (p195)

All this may have some regional strategic significance, argue the authors:

In showing that the Saudi oil network would be resilient in the face of a concerted attack by one of the most capable actors in the region, our research indicates that threats to regional oil production are overblown. By implication, the United States may be able to reduce its military commitment to the region … in sum, oil is a lucrative target, but it is not universally vulnerable. (p201)