Safeguards in Syria

On 6 September 2007, the Israeli Airforce carried out a daring raid, striking a target deep in eastern Syria. Since then, there has been wild speculation in the international press about whether Syria has been attempting to put together a nuclear programme, away from the eyes of the international community. Initial rumors focused on the possibility of nuclear cooperation between Syria and the DPRK. The initial thinking was that the DPRK had transferred uranium enrichment technology to Syria.

Syria has been planning for an expanded use of nuclear power since the mid-1980s. Plans, however, have often been shelved due to a lack of resources. The country presently operates one research reactor called the SSR-1, which went critical in March 1996. This reactor, a Chinese designed Miniature Neutron Source Reactor (MSNR), poses no proliferation risk. It is a pool-type reactor fuelled by 980 grams of Chinese supplied highly enriched uranium (about 90 per cent enriched). This quantity is well below what is needed to build a nuclear explosive device. The MSNR has been a popular third-world reactor. The Chinese have supplied the same design to Ghana, Iran, Pakistan and Nigeria. According to the Syrian Atomic Energy Commission website, the reactor is used for training and scientific work. Also at the site are analytical laboratories for, amongst other things, isotope production. According to the IAEA department of safeguards, this is the only declared and safeguarded facility in the country.

So what then was the target of the strike?
This is still unknown, but there are some indications that the target could be nuclear. On 23 October, the highly respected Institute for Science and Global Security (ISIS) published satellite pictures of a site that could have been the target of the Israeli air strike. The site is located approximately 400 kilometers northeast of the country's capital, Damascus. The imagery is inconclusive, and the nature of the site need to be investigated through some form of on-site visit. That notwithstanding, ISIS concluded that a building on the site might have been "a small gas-graphite reactor of the type North Korea build at the Yongbyon nuclear site north of Pyongyang" under construction. This conclusion is mainly supported by the dimensions of the building (which perfectly match those of the 5MWe in Yongbyon), and the presence of a pumping station at the bank of the nearby Euphrates river. I've been asking around at various defence research establishments in Europe, and their views are very similar to those of ISIS. They pay particular attention to the dimensions of the building.

However, the facility location raises many questions, some minor and some more substantial. For instance, Mr. Albright notes, "a reactor requires a large volume of water for cooling and this pump station could serve that purpose". If the reactor was imported from North Korea, it would be gas cooled. Some have said that this negates the need for a pumping station nearby. This is a minor point, since most gas-cooled reactors need a heat exchanger through which hot gas is cooled. Cold water usually flows into this exchanger, cools down the gas, and is then dumped in the environment (usually in nearby water - some steam can be released through a cooling tower).

Rather, the site is peculiar for other reasons. ISIS has rightly pointed out the existence of an airfield to the north of the site, and have argued that this might be used for logistics. However, there is no infrastructure at this field. There are neither cargo terminals nor reception buildings. Moreover, the site itself is located close to what seems to be a major road, and very close to various farm buildings. If those farms are not owned by the military, how would the Syrian government be able to maintain secrecy, with farmers and livestock walking about the neighborhood? (Maybe they didn't). Also notable is the complete absence of any form of physical security. The site does not seem to be secured by fences, checkpoints, military barracks or air defenses. There are three possible explanations for this. Either the site is supposed to be hidden in plain sight, its defenses are buried, or they were not installed yet.

However, after the bombing, the reactor construction was completely removed and the ground scraped by heavy-duty bulldozers. In the view of ISIS, "dismantling and removing the building at such a rapid pace dramatically complicates any inspection of the facilities and suggests that Syria may be trying to hide what was there". It is difficult to argue against that view.

What's the safeguards issues?
There are several. The most important question is whether Syria has breached its safeguards agreement with the IAEA. The country was slow in bringing its Comprehensive Safeguards Agreement (CSA) into force (INFCIRC/407).

Syria was one of the original signatories of the 1968 Nuclear Non-Proliferation Treaty (NPT). According to article III.4. of the treaty, negotiations of safeguards should have begun around 5 September 1970, or 180 days after the NPT's entry into force. Negotiations should have been completed and safeguards brought into force by March 1972. However, INFCIRC/407 became legally binding only in 1992, some 20 years too late. Syria has yet to conclude an additional protocol to its safeguards agreement, arguing that it will do so only after Israel joins the NPT. Since the additional protocol is not the agreed safeguards standard, there is very little pressure that can be brought upon Syria to make her change her mind.

According to paragraph 42 of the model CSA, design information in respect of new facilities should be given in accordance with the time limits specified in the State's Subsidiary Arrangements. The general rule is that such information shall be provided "as early as possible before nuclear material is introduced". Before the 1991 Persian Gulf War, the arrangement was that design plans should be given 180 days before material is introduced into the facility. After the discovery of Saddam Hussein's clandestine nuclear weapons program, however, the IAEA Board of Governors decided that design information should be given at the drawing board stage. This understanding entered into force through an exchange of letters between the IAEA and the individual states throughout the 1990s. My understanding from talking to Agency officials is that only Iran amongst all IAEA member states remained to sign up to this scheme in 2003. Syria would then have been bound by its agreement to supply design information well before construction starts.

How to prove this?
The Agency's primary responsibility under a CSA is, according to paragraph 2 of the agreement, to "ensure that safeguards will be applied ... On all source or special fissionable material in all peaceful nuclear activities within the territory of the State, under its jurisdiction or carried out under its control anywhere". The exclusive purpose of such safeguards is to verify that "such material is not diverted to nuclear weapons or other nuclear explosive devices".

The Agency regularly receives information from the state on its declared activities. However, if such information "is not adequate for the Agency to fulfill its responsibilities" under the agreement, the IAEA has the right to call a special inspection (see paragraph 73 (b)). A number of conditions must first be met. The state and the IAEA need to consult about the need for access (paragraph 77). Any disagreement should be resolved by consideration by the Board of Governors or, in the last resort, by referral to an arbitral tribunal (paragraphs 21-22). There is one exception to this rule. If the Board of Governors, "upon report of the Director General", decides that an action by the state is "essential and urgent" in order to ensure that nuclear material is not diverted to weapons it "shall be able to call upon the State to take the required action without delay, irrespective of whether procedures for the settlement of a dispute have been invoked" (paragraph 18).

Special inspections have been invoked twice in the history of the IAEA. First in order to investigate the extent of Romania's cold war nuclear programme ("a warrantless and silly request" according to a good friend of mine who used to work in Romania's regulatory authority), and second in order to investigate the extent of North Korean obfuscation in the early 1990s.

While the tool has been used sparingly, it remains the only viable legal option to get in-country to investigate the suspected site. Naturally, other solutions exist. If Syria so requests, the Agency can deploy inspectors based on article XII.A. of the Statue. These inspectors "shall have access at all times to all places and data and to any person who by reason of his occupation deals with materials, equipment, or facilities which are required by the Statute to be safeguarded". This exceptionally broad mandate was recently invoked as the legal basis for inspections in the DPRK. It is unlikely, however, to be applied in the case of Syria.

Stay tuned.

Travels make me feel ashamed

I am aware that my posting frequency has been dreadful lately, and for that I apologize. It is due to me travelling around trying desperately to do good deeds and to return the favors of various conference-organizers by attempting to say smart things. Last week I travelled to Moscow, Russia, to participate in a meeting discussing Iran, and this week I am in Washington D.C. at the Stanley Foundation's meeting on the Fissile Material Cut-Off Treaty. I am very interested in this meeting, since the issue of FMCT verifiability seems to be back on the table. Matt Martin, one of the programme officers at the foundation, has managed to get a good group of people together. Some of which I know only by name, but whose work I deeply respect. Discussions are bound to be interesting, and I expect to upset one or two of my American colleagues by pestering them with remarks about verification. Especially since I am increasingly of the opinion that a lot of the thinking around verification is old, stale, and that new approaches may be called for.

Whenever I get time, I try to finish off a paper on on-site inspections and managed access that I am writing together with a Norwegian colleague. Be on the lookout for that. It is supposed to be a delivery in trilateral process involving the Atomic Weapons Establishment, VERTIC, and a number of Norwegian institutes. So I hope it will interest some.

Finally, the final report of the UN Verification Panel is supposed to be released today. Three generations of VERTIC directors (Patricia Lewis, Trevor Findlay and Angela Woodward) contributed to that process as experts, and yours truly wrote a power-point presentation that introduced Google Earth to the UN. That was a long time ago now.

New Agency Page

The International Atomic Energy Agency has recently posted a new web page, which timelines the history of the Agency. The labeling of the timeline is telling. The 1940s are characterized by "the terrible sword", followed by the "global visions" of the 1950s and the "rising hopes" of the 1960s. Then, during the 1970s, the Agency faces "the dual challenge", followed by the "transboundary stakes" of the 1980s. According to the timeline, these stakes translated into "new realities" in the 1990s, whereas the first decade of the 21st Century is labeled to belong to "testing times". 

I'm relatively certain that it wasn't meant to be come out this way, but it would seem like things started off bad (no point denying that really), then got a bit better during the 1950s and 60s, only to atrophy over the following 40 years or so. Is the reader to believe that the challenges started to appear after the entry into force of the 1968 Nuclear Non-Proliferation Treaty?

A full house: Russia ratifies the additional protocol

On 3 October 2007, Russian President Vladimir Putin signed Russia's additional protocol into law. This follows the approval by the lower and upper house of the Russian legislature on 14 and 19 September 2007 respectively.

The Russian Federation's Voluntary Offer Agreement (INFCIRC/327) entered into force on 10 June 1985. The country signed the Additional Protocol to this agreement on 22 March 2000. US President George W. Bush signed his country's Additional Protocol into law on 8 January 2007.

Nuclear weapon States are not required to accept safeguards under the 1968 Nuclear Non-Proliferation Treaty (NPT). In theory, a consequence of this could be that non-military enterprises in the nuclear weapon states get an advantage over their competitors in the non-nuclear weapon states. After all, these enterprises would not have to safeguard proprietary information from snooping inspectors, nor would they have to endure interruptions in their operations, while an inspection is ongoing. Therefore, under a voluntary offer agreement (VOA), the nuclear-weapon State offers, for selection by the IAEA, some or all of the nuclear material and/or facilities to be put under safeguards. At least that's some of the theory.

Primarily, however, these agreements add nothing but symbolic value. For instance, the preamble of INFCIRC/357, states that Soviet Union entered into its agreement, "for the purpose of promoting widespread adherence to the Treaty, further development of Agency safeguards and encouraging their acceptance by an even greater number of States". The wider membership of the non-proliferation regime seems to accept that the VOA's are there primarily for show. In 2000, for example, the NPT review conference called for the "wider application of safeguards to peaceful nuclear facilities in the nuclear-weapon States" but also stressed that these safeguards be applied "taking into account the availability of IAEA resources" (see paragraph 12 on page 4 of the Final Document). In other words, safeguards in nuclear weapon states are good, but only as long as the nuclear weapon states pay for them.

Practically, the agreement contributes little to non-proliferation. INFCIRC/372 stipulates that Russia shall provide the Agency with "a list of facilities" of the country's choosing where safeguards shall apply (see article 1). The only thing Russia needs to do if it wants to withdraw material or facilities from safeguards is to notify the Agency (article 12). This notification shall be in advance, although the agreement doesn't specify what "in advance" actually means). The notification shall, according to article 35 (b), contain the following information:

• the facility or facilities from which nuclear material is being withdrawn;


• the date of its withdrawal; and


• the quantity and composition of such material.

Unsurprisingly, there has been little enthusiasm to squander money in verifying the peacefulness of nuclear activities in a state which are legally entitled to produce nuclear weapons anyway. Therefore, in 2006, safeguards activities in the Russian Federation were limited to the evaluation of accounting reports on the export and import of nuclear material. According to the latest safeguards statement, no facilities were selected in 2006 from Russia's list of eligible facilities.

Russia's move is potentially significant from a political point of view, however, since it means that all nuclear weapon states now accepts the idea of an additional protocol in force. The US ratified its Additional Protocol earlier this year (see my previous post here). China's agreement entered into force on 28 March 2002, France and the United Kingdom's on 30 April 2004 (together with the rest of the European Union). It is probably naïve to think that the fact that all nuclear weapon states have signed up will, somehow, accelerate the adoption of Additional Protocols elsewhere. However, by ratifying, the nuclear weapon states have removed the argument that they do not live as they learn. And perhaps that's something.

Deception and verification

It is sometimes said that "all that glitters is not gold". Proverbs of this kind are increasingly relevant in today's information driven society. Humanity is attempting to manage ever larger flows of information pouring out from an ever increasing number of sources: blogs, websites, printed press, television and radio, to mention but a few. The average citizen living in a post-industrial society receives a staggering amount of information every day, but it is unknown precisely how much. While faculty and students at the University of California at Berkeley have tried to estimate how much information is out there, they've yet to factor in the massive flow of "personal information" (for instance word documents and private e-mails). Still, their conclusion is that "the world's total production of information amounts to about 250 megabytes for each man, woman, and child on earth". In their, rather dramatic view, mankind is "drowning in a sea of information". And this was four years ago.

To some degree, computers can help in sorting out and characterizing this information. However, at the end of the day, the good analysis will always be done by human being. A skilled operator will be able to cut away the noise and the chatter, so that only the important issues remain. The output, however, will be dependent on the quality of the input (the raw data), and it is here where deception comes in. All forms of deception conducted by a state or non-state actor will affect the reliability of any analysis. This is nothing new. According to a 1957 U.S. Special National Intelligence Estimate (SNIE) on "Soviet capabilities for deception" there is several ways in which the actor can deceive its opponent.

(a) by planting false information;

(b) by coloring or distorting otherwise authentic information as to make it convey a false impression;

(c) by selectively releasing some correct information on a subject while withholding essential parts of the total picture; and

(d) by releasing plentiful data, whether true or false, with the object of overshadowing and obscuring certain particular items of paramount importance.

The authors of the U.S. SNEI argue that there are three principal and mutually exclusive objectives for deception. First, the aim can be to cause the target to underestimate the actor's strength, capability or determination. This objective aims to instill a false sense of security in the mind of the target. Second, the aim can be to cause the target to overestimate the actor's strength, capability or determination. This objective could aim to force the target to overextend itself somehow. For instance, if the target is a military commander, the strategy would aim to make him deploy sizable forces in anticipation of an attack which, of course, will never come. The final objective can be to cover a certain activity (attempting to conceal it) from the target.

The SNIE is more than 50 years old but its conclusions still seem valid. The report warns that the effects of deception cannot be eliminated, simply mitigated. It applies a simple logical process to counter the effects of deception somewhat. The report argues that:

as a first step, some of the evidence could conclusively shown to be free from the possibility of hoax. The remainder was then assumed to be the product of deception, and the difficulty and cost of such deception, roughly assessed, was set against a judgment of the probable advantage which the USSR would gain from the deception. Frequently, by this process a convincing estimate of the likelihood of deception could be reached.

There is a discussion brewing within the nuclear safeguards community on how, for instance, intentions can be judged, and whether there is a way in which verification resources can be channeled towards hard-cases rather than being used up in countries considered compliant by the international community. VERTIC, for instance, published a paper last year in which we argued that interpreting data on military research and development would "require more use of inference, the evidence produced would probably be less conclusive and it might prove necessary to try to determine the intention of states". The IAEA Director General's Standing Advisory Group on Safeguards Implementation has also done some useful work on this topic, and has attempted to define criteria on which the "peacefulness" (for lack of better words) of a state's nuclear fuel cycle can be judged.

However, opponents to any such scheme argue that "hard cases make for bad law" and that "an unequal law is not a good law", both very strong arguments. However, equally persuasive is the view that it makes no sense to squander resources in places where they need not be squandered. The IAEA's Director General has strong views on the issue. On 16 October 2006, he argued that:

To determine all the details of a country´s nuclear programme is a very daunting challenge. What level of assurance do you need; how do you draw assessments from the facts; how do you distinguish between technical data and future intentions? ... We do not get into reading future intentions of countries because we are not equipped to do so and also because, as we know, future intentions can change.

In any case, the prospect of manipulated data being fed into the verification process should not be underestimated. Especially so in an information driven safeguards environment.

I've made the 1957 SNIE available here. I would be interested in hearing views on the relevance of this document in today's information driven world. What is especially interesting is the discussion on self-deception in paragraphs 21 and 22 (on page 6). Even back in 1957, the wise men of intelligence concluded that there is a tendency to accept data as credible "for no better reason than that it fits harmoniously into a previously established context or pattern". Sounds familiar to anyone?

3D movie of Yongbyon

Whatever you're doing this week, make sure that you do not miss this stunning QuickTime movie produced by the Satellite Imaging Cooperation. It doesn't really tell us anything new, but it looks great.