IAEA Report on Iran

I have acquired the latest Agency report on Iran from a friend. I am not going to hold it back from the arms control community, so please enjoy. Rest assured that you'll see more analysis here in due time. Right now, however, I am deeply engaged in preparing stuff for my new boss.

Let’s talk modalities

On 27 August 2007, the IAEA secretariat circulated INFCIRC/711, a document outlining some "understandings" of the IAEA and Iran on how to resolve outstanding verification issues. The document was circulated on the request of Iran. Its language is quite cumbersome, and its content is difficult to fully understand without awareness of the proper context. I attempted to give some background in my post of 6 August 2007; most of it was informed guesswork. Now, it turns out that the Agency and Iran was indeed discussion a facility attachment for the Fuel Enrichment Plant (the FEP) in Natanz. One could be tempted to conclude that this attachment was agreed on after intense negotiations in good faith, but this temptation should be resisted.

What's not mentioned in Iran's information circular is that a facility attachment for the Pilot Fuel Enrichment Plant (the PFEP) was presented to the Iranians as early as in May 2003. In other words, a blueprint on how to go about verifying activities at the site has been in the making for years. But progress in finalizing it was slow. It is reasonable to assume that discussions on a facility attachment in respect to the larger facility have been going on for years as well. Facility attachments are drafted by the Department of Safeguards and is then cleared by the Agency's Office of Legal Affairs and other departments within the Agency before it's sent to the state. The state usually replies with suggested changes to the attachment, and these changes are debated for a while. Once the IAEA and the state are happy with proposed changes, the final text is attached to a letter which basically asks whether the state is happy with the agreement. The facility attachment enters into force on the day the state replies in the affirmative. It's an informal affair. The cover letter is usually signed by the Director of Operations for the appropriate division of safeguards, and it does not require an okay from the Board of Governors. In respect to Iran, the facility attachment is due to be finalized at the end of September 2007, after which it would need to enter into force.

The rest of the first part of the Information Circular, entitled "Present Issues" is practically a wash-up of some past technicalities. For instance, Iran promised to issue more multiple-entry visas (something they are already required to do under their Subsidiary Arrangement to its Safeguards Agreement). It also accepted a few inspector designations to compensate for the massive purge of western inspectors which occurred earlier this year (see this post). However, as I explained in January, the designation and rejection of inspectors is a routine affair, as long as it does not impede inspections in the country. So the designation carries largely symbolical value. Other issues are more interesting.

A question of plutonium
Shortly after the end of the Iran's bloody war with Iraq, the country seemed to take interest in plutonium reprocessing. Between 1988 and 1992, Iran irradiated 7 kilograms of uranium dioxide targets and extracted small quantities of plutonium. Uranium targets had been produced at the Esfahan facility. These targets were then transported to Tehran where they were inserted into the Nuclear Research Centre's 5 megawatt-thermal pool-type research reactor. The Iranians are believed to have separated some amount of plutonium at the centre (in the milligram range). NTI holds that the maximum production capacity of the centre is 600 grams of plutonium per year. This figure reportedly originates in a 1977 report by the US Defence Intelligence Agency.

Ever since Agency investigations in Iran started, there have been uncertainties surrounding the plutonium separation experiments. For instance, the Iranians originally said that they had produced around 200 milligrams of plutonium, but Agency calculations indicated that more should have been produced. This was more or less resolved later on. But other issues remained, for instance the Agency analysis also indicated that experiments were conducted after 1992, which the Iranians disputed. The Agency has therefore not discounted the possibility "that the plutonium analyzed by the Agency has been supplied from source(s) other than the ones declared by Iran" (GOV/2006/27, para. 17). The Agency got hold of the notebook of the researcher in charge of the experiments, but that did not help clear up the issue.

Now, however, it's claimed that the issue is resolved. But the "modalities plan" does not offer any explanation as to how the outstanding questions were resolved. We have to wait and see.

The Pakistani centrifuge connection
The document is vague here, but claims that outstanding issues can be resolved by November 2007. I am very skeptical. Iran received drawings for the P-2 centrifuge in 1995, but claims that it did not carry out any work on the centrifuge type until 2002. Moreover, it holds that it did not discuss the design with the supplier, and that it did not ask for any deliveries of components. The Agency has been skeptical of that explanation, and Iran maintains that it's given all the information it can give, and that no more information exists. Moreover, the issue of where traces of highly enriched uranium found in areas where centrifuges had been stored came from is not fully resolved, although the results of the Agency's investigations seem – on balance – to support Iran's explanation that the contamination was imported with the machines. Thus, while the P-1 contamination issue is almost resolved, it remains to be seen whether good verifiable explanations can be offered in respect to the P-2 programme.

The "general understandings"
At the end of the document, there are some "general understandings" that seem odd. First, the document claims to cover "all remaining issues". David Albright and Jacqueline Shire have quite rightly described this as "a sweeping statement that sets an unfortunate precedent regarding Iran's past nuclear activities, about which little is known in important areas". I just find it odd that the Agency could have agreed to this language in the first place (which makes me suspect that the modalities plan is just "the modalities plan as seen by Iran"). It is obvious that very little confidence will be gained if the Agency is prohibited from looking into certain areas. Moreover, how can the Agency "confirm that there are no other remaining issues and ambiguities" when they're not even sure that they asked all relevant questions yet? Also, the modalities plan seems to suggest that the Agency has one shot at asking the relevant questions. Once Iran has received those questions, they will reply. That reply is definite, and after that "no other questions are left". Obviously, it doesn't work like that. Finally, there is not a single word as to how this is supposed to play out practically. Not a hint as to whether the Agency will get the kind of broad access which has been called for by the Agency itself, and so many others, for such a long time.

Other stuff: let's talk about that later…
There are other issues where discussions do not seem to have gone that far. Rather, the information circular stipulates that these will be solved down the line, when other issues have been resolved.

Okay, the plan is a positive development insofar that it signals an Iranian willingness to reengage in verification. And, as the old Chinese proverb reminds us, every journey begins with a single step. It's just that this journey could have begun years ago, and under better circumstances. The plan doesn't offer anything that goes beyond what Iran is already legally required to do, which is fine as long as what's legally required will help provide the answers the international community seek. I remain very skeptical that it will.

Six months on – where now for the DPRK?

The 13 February 2007 denuclearization action plan (DAP) is today six months old, and it is fair to say that most of the initial actions are in the processes of being completed. That is, some nuclear activities at the Yongbyon Nuclear Complex have been shut down, and the facilities are sealed up. The large IAEA team that initially swarmed over the site has now reportedly been reduced to two persons who now stay in the regime supplied guesthouse (pictured). These brave international civil servants are in all likelihood experiencing what Michael Collins felt during his day of flying alone around the Moon. Complete loneliness. I have heard that a portable field generator was amongst some of the equipment brought along for the expedition. I hope that other equipment included a DVD player and some decent films, because it's highly unlikely that something good is showing on state controlled television (unless they show the adorable Soviet cartoon Nu, pagodi! of course).

The interesting part of the North Korean verification drama is about to unfold. North Korea is now, amongst other things, supposed to provide what's known as a "complete declaration" of all their nuclear programs. North Korea's nuclear envoy, Mr. Kim Kye Gwan, has reportedly stated that it intends to report and disable all their nuclear assets in a very short period of time. The South Korean envoy recalled that Mr. Kim said that it could be done "within five or six months, or by the end of the year". This statement makes me extremely sceptical. First, I would have preferred to hear this from the North Koreans themselves. Second, unless the North Koreans intend to bring in a veritable army of skilled nuclear workers, I just cannot see a complete disablement happen within such a ridiculously short time.

From a verification theorist's point of view, the phrase "complete declaration" brings back bad memories about proving a negative. Few have forgotten about the 13,000 page Iraqi declaration handed over to the United Nations in December 2002, a few days later dismissed by then UK foreign minister Jack Straw as "obvious falsehood". Now like then, some are very likely to say that North Korea will not declare its entire nuclear program, that they cannot be trusted, and that there is no way to verify that their declaration is complete.

Denouncing a complete declaration is like levelling a good conspiracy theory against an established chain of events. The believer of the established truth is forced to prove a negative ("but you don't know that there wasn't someone on that grassy knoll"). A good verification scheme can alleviate some (but not all) of the problems associated with such a situation. However, no verification schemes have been envisioned beyond the initial steps (see my previous post about this), and it would not get around the fact, as any logician will testify, that proving a negative is impossible. A certain degree of trust must be mixed into the denuclearization process. Trust that, I would argue, simply aren't there at the moment.

However, there are hurdles to get over before one should start to worry about the authenticity of any North Korean declaration. Xinhua reported a while back that the latest round of six-party talks failed to agree on a deadline for the DPRK to declare all its nuclear activities, and to disable the ones that are currently frozen. A new progress report is expected when the six-party group reconvenes in Beijing in September.

Agency activities in Iran

Representatives of the International Atomic Energy Agency are supposed to meet with their Iranian counterparts in Iran today, to discuss, in the words of the Washington Post, "an 'action plan' to give the International Atomic Energy Agency (IAEA) more access to Iran's nuclear facilities". The team is reportedly led by Michio Hosoya, who used to be a unit head at Operations C, but that is now apparently doing the job that Christian Charlier used to do before he was reassigned. Readers may recall that Iran refused to issue a visa to Mr. Charlier back in February 2007, and that he since then has been working in another section of the Agency (see my previous post of 15 February 2007). His counterpart on the Iranian side will be Mr. Mohammad Saeedi, the deputy head of the AEOI.

An action plan sounds good. But my reading of the news is that it's mostly spin. While it sounds like Iran and the IAEA is sketching on some grand transparency plan, perhaps along the lines outlined by James Acton and yours truly, the reality is something more mundane. Recent reports strongly suggest that what they are discussing is facility attachments for a number of key Iranian installations, Natanz in particular.

Facility attachments forms part of a states subsidiary arrangement, which is a contract designed to give detail to a country's comprehensive safeguards agreement. It contains technical and administrative procedures which specify how the safeguards agreement shall be applied on the ground. These agreements contain a general part, applicable to all common nuclear activities in the state, and facility attachments, which is prepared for each facility. This document is important, since it outlines the safeguards approach to be taken at the facility.

The general part of Iran's subsidiary arrangement has been in force since 1976 (see GOV/2003/40), but getting Iran to sign up to facility attachments have been trickier, at least in respect to newly build complexed. Prior to 2006, the IAEA published a list of all safeguarded facilities, which contained information on whether the facility had a subsidiary arrangement in force. This very enlightening list has, for some unexplained reason, been substituted with a non-descript statistical statement. Anyway, as of 31 December 2005, the situation in Iran was as follows:

Facilities with subsidiary arrangements:
Tehran Research Reactor, Tehran
Heavy Water Zero Power Reactor, Esfahan
Miniature Neutron Source Reactor, Esfahan
Light Water Subcritical Reactor, Esfahan

Facilities without subsidiary arrangements:
Uranium Chemistry Laboratory, Esfahan
Uranium Conversion Facility, Esfahan
Fuel Fabrication Laboratory, Esfahan
Pilot Fuel Enrichment Plant, Esfahan
Karaj Waste Storage Facility, Karaj
Jabr Ibn Hayan Laboratories, Tehran

I haven't seen anything that would suggest that this situation has changed. Lately, there has been obfuscation enough to make the father of maskirovka, the great marshal (and hero of the Soviet Union) Georgy Konstantinovich Zhukov, burst with pride. The press seems extremely confused as to what's going on, and even Jeffrey Lewis admits that "making sense of the press reports is tough … without some context". In my mind, it seems like the Agency is attempting to make up for some of the ground that was lost in the wake of three Security Council Resolutions, and that the Iranian side is trying to squeeze out as much propaganda value as they can from this present round of Agency-Iranian interactions. Well, as long as something good comes out of it, why not?

More important is the Solana-Larijani process, which still seems to move forward (albeit in a pace similar to that of a tortoise in no rush). However, significantly, construction in Natanz continues at a slower pace than before. I've previously estimated that Iran might finish its first module by September. This looks increasingly uncertain if one is to believe the press. According to a report by Mark Heinrich, an EU diplomat said that the Iranians had 10 cascades running with an additional two operating under vacuum by 27 July 2007 (that's about 2,000 units all in all). If true, the pace of construction has slowed down to a trickle. Why this has happened is manifestly unclear. It could be a political decision, or a technical one, or a combination of both. In the meanwhile, there are rumors, and denials, and renewed rumors, about some sort of pause, or suspension, or whatever you want to call it.

It looks like that the next Agency report on Iran will be a very interesting read.

US-India 123 Agreement

The 123 Agreement between India and the United States became public today, and I have decided to put a local copy of the text on my supporting website, just in case someone decides that secrecy is preferable after all. There is a lot of text to go through and ponder, but some newspapers, such as the Economist, remains very critical of the agreement and its implications. Naturally, Indian reactions to the agreement has been ecstatic, whereas the Pakistanis have ominously referred to the agreement as a grave mistake which will only serve to enrich India's nuclear arsenal, not its economic development.

There is language in this agreement that should set of the non-proliferation alarm bells, like, for instance the scope of Indian - US cooperation in article 2.2 of the agreement and the high enrichment threshold in article 6.i. The article covering safeguards looks remarkably weak, especially since it does not contain fall-back (bilateral) safeguards, although a lot in respect to safeguards are still to be agreed between India and the IAEA.

It's very interesting that a lot of safeguards implementation and negotiation is left to the Agency. The safeguards agreement shall explicitly be "India-specific", but there is very little in there in terms of guidance for negotiators, except that no nuclear material is allowed to flow between any party's civilian and nuclear fuel cycle (article 10.1).

Expect more about this. The talks in the NSG are about to commence, and some I've spoken to foresee prolonged and bloody hand to hand combat in the US legislative process.

Revisiting Vela 6911

Introduction
Resistance against the 1996 Comprehensive Nuclear Test Ban Treaty is in the United States driven by an infected ideological debate, which really is about the future of America's vast nuclear arsenal than whether the treaty is verifiable. The consequence of this debate on the treaty itself is considerable. Today, the organization tasked with setting up, testing, certifying and running the treaty's verification regime is in dire need of funding. It has taken the unprecedented step of putting a document on its website detailing its collection-rate, which just a couple of years ago was refreshingly high. The document shows that the organization is suffering from a deep hole in its finances, since its member states collectivly owe some US$81,5m. The United States owe some US$28,3m, or one third of the total debt. Daryl Kimball and his colleauges at the Arms Control Association (an organization many of us takes very seriously) is the persons to turn to for more information on the budget situation.

Recently, the CTBT has resurfaced in the US debate. Quite unfortunately, its been linked to the debate on reliable warhead replacement (see the Economist on RRW here, and on the CTBT here). It is unfortunate, since the CTBT should not, at least not ideally, be placed in the centre of domestic horsetrading: the idea of trading nukes for the ban is fraught with complexities and difficulties. It is good in the sense that many institutes are sharpening their arguments for or against US ratification, which gives the treaty some publicity. And as an old friend of mine in the music business says: "to do some face-marketing is always a good thing".

So, what's coming out from the think-tanks of the enemies of the CTBT? The Heritage Foundation is usually refreshingly clear about its motives, the CTBT is a bad thing because it ultimately aims to give away the country's arsenal of some 5,000 warheads. Others, such as Dr. Kathleen Bailey, still stubbornly maintain that the treaty cannot be verified. In a 22 March 2007 letter to the editor of the Economist, Dr. Bailey argues that:

The CTBT's monitoring centre can detect and identify non-evasive nuclear testing of 1 kiloton or greater, but it cannot detect nuclear tests below this yield nor those above the yield done evasively. Both types of tests could be carried out to assure the reliability and design of new weapons ... using de-coupling, a 1,000-ton explosion can be made to look seismically like a 14-ton explosion. Neither the international monitoring system nor American national technical means could detect this.

Her letter is clearly disingenuous. It is, for instance, more or less established that the DPRK's first nuclear weapon test failed to reach the one kiloton threshold. Yet, its signals were clearly detected by the CTBTO's international monitoring system, as well as seismological research institutes across the world. The release of radionucleid particles were detected, not only by the the CTBTO, but also by national technical means rushed to neighbouring South Korea.

Not even when she's discussing evasion scenarios is she correct. Indeed, she is using the maximum theoretical decoupling factor (70) and throwing it around like it is fact. She chooses to ignore that a decoupling factor that large has never been seen for a nuclear test with a yield greater than about 300 tons. A more sensible decoupling factor, all things considered, is 50. Her argument, therefore, is either uninformed or biased (for more information about decoupling and alleged nuclear weapons tests, see this article by Professor Terry Wallace of the University of Arizona). A fairly safe bet is that political considerations drive her "technical" argument.

In my mind, the real problem with verifying compliance with the 1996 Comprehensive Test Ban Treaty is not if the treaty's extremely sensitive monitoring system can pick up and pinpoint a decoupled nuclear explosion somewhere at the Semipalatinsk or Nevada test sites. Neither is the problem whether the system can pick up the signature of a large explosion somewhere else. I would submit that the real problem lies elsewhere.

Assume that a state has enough nuclear material for a reasonably long test series, and that it would like to verify that its new nuke design is working, how would it go about doing that? Well, there are several environments in which one can test: on the surface (or in the air), in space, on or under water, or underground.

At present no state seems to be willing to conduct a surface test anymore. All nuclear newcomers - including North Korea - have opted for the underground test. In most cases, their test sites have been selected and prepared years, perhaps even decades, in advance. It would have been great if states refrain from testing due to opinio iuris (i.e. the belief by the state that its refraining from practice due to legal obligation), but I think the reluctance to test above ground has more to do with the impact on the environment than any in-depth legal discussions within government (at least as far as North Korea is concerned, India and Pakistan are state parties to the Partial Test Ban Treaty).

In any case, a legally unrestrained nuclear newcomer, such as North Korea, do not really care that much if other states find out about their test. Indeed, in some cases, they'd want other states to know. In these cases, the government usually gives an heads up, announces it post facto, and lets the international media pack verify it (dubbed "verification by CNN" by one of the coolest guys I know at the CTBTO).

But what happens if a nuclear newcomer wants to maintain some form of strategic ambiguity? Wouldn't a nuclear test would be quite counterproductive for a country who wants to be deliberately unclear about its nuclear weapons? Wouldn't a country like that make sure that a weapon detonates on international territory? Perhaps over international waters or on either Pole? Could such a test be traced back to the state?

The ultimate evasion scenario - maybe
Such a testing protocol may already have been successfully tried. On 22 September 1979, an ageing orbiting US satellite (Vela 6911) picked up a flash over the South Atlantic. The characteristics of this bright light indicated that a nuclear device with a yield of approximately three kilotons had gone off on the earth's surface at a location close to South Africa's Prince Edward Island. First, the US was convinced that the flash was a nuclear explosion. According to a memo dated 22 October 1979, the US intelligence community first had "high confidence, after intense technical scrutiny of the satellite date" that a low yield atmospheric explosion had occurred. However, doubt lingered, especially since no nuclear debris could be found. According to a declassified but largely redacted US report, the detonation "set off one of the most extensive air sampling operations in recent years". Sampling aircraft flew 25 sorties totaling over 230 hours of flight-time in order to attempt to capture possible radionuclides. However, the analysis of the data collected indicated that the event was "probably non-nuclear and more likely of an unknown natural origin".

The results of these examinations are remarkably inconclusive, and the event itself therefore remains controversial. On balance, a lot of the evidence seems to support the theory that it was a freak natural occurrence, perhaps a falling meteoroid. Vela 6911 had been in orbit longer than it was designed for, and as a result some of its systems had started to fail. For instance, its EMP sensor had reportedly broken down, which meant that it was unable to pick up any electromagnetic pulses emanating from the explosion.

However, the Vela system did successfully detect several other tests. Had not one satellite malfunctioned, the incident would not have remained a mystery for long. So the incident in many ways contains both an evasion scenario and the solution to the problem.

While the CTBTO operates a finely tuned and remarkably effective system to detect nuclear explosions, its capabilities does not include data collected by space based assets. This is, to some degree, offset by the treaty's reference to national technical means. Many years ago, VERTIC brought forward some cutting edge thinking on verification (see this article in Trust & Verify No. 37 from May 1993). Patricia Lewis and her colleagues quite accurately foresaw the present monitoring system, but held that satellite imagery and overflights would also be quite useful components. However, at the end of the day, these useful tools were never incorporated.

A couple of years ago, I tried to talk to people at the CTBTO, and elsewhere, about opening up the debate on whether one could think about supplementing the verification regime with space based means, but I was told that no-one would be interested in that. Instead, I was told that the organization need to focus on the development of existing and treaty stipulated verification technologies and methodologies, and that a renewed discussion would avert attention from the very real, and very important goal of finishing the construction of the monitoring system, and making sure that there is a working on-site inspection protocol. Undeniably, those are strong arguments, but there is also something to be said about being forward looking.

Contrary to what Dr. Bailey and others are suggesting, the International Monitoring System is far better than its designers would ever believe. Undeniably, mankind's thirst for knowledge, better technology, and faster computers has benefited the treaty in ways hard to imagine in 1996. Okay, this must be highlighted, perhaps by another study by the US academy of sciences, or by a reconvening of the CTBT Commission. But at the same time, discussion on how to make the verification regime better, something already foreseen in article IV.A.11 of the treaty, should not be suppressed. Humans are stimulated by progress and put off by stagnation.

DPRK safety standards lacking

Reproduced below is a very interesting piece from RIA Novosti. Apparently, monitoring the 5MWe reactor was not as straight forward as many (including myself) originally thought. The main difficulty, however, is not attributable to plant layout or to the general cooperativeness of the host. No, the main problem is safety. Not that surprising, really.

TOKYO, August 1 (RIA Novosti) - Radioactive contamination could delay the work of experts from the UN nuclear watchdog to seal North Korea's nuclear facilities, shut down by Pyongyang under a disarmament deal agreed in Beijing in February, the Kyodo news agency said Wednesday.

The agency cited sources at the International Atomic Energy Agency (IAEA) as saying that traces of radiation have been detected at a an operational five-megawatt nuclear reactor and a plutonium-extraction plant in Yongbyon, 100 kilometers (60 miles) north of the capital Pyongyang.

IAEA experts arrived in North Korea in July to put seals and install monitoring equipment at five North Korean nuclear facilities by mid-August as part of an international effort to fold Pyongyang's nuclear program.

The IAEA officials said the contamination did not pose any threat to the environment, but would delay their work until the end of August because the inspectors had to decontaminate the facilities before installing the monitoring equipment and seals.

"Their [North Korean] nuclear safety standards differ from our standards," Kyodo quoted an IAEA experts as saying.

The Yongbyon complex consists of an operational five-megawatt nuclear reactor, a plutonium-extraction plant, a nuclear fuel production facility and research labs. The site also contains a 50-megawatt reactor whose construction was suspended under a 1994 nuclear deal with the United States.

[...]