David Sanger and Bill Broad have a gossamer story in the New York Times about whether North Korea might have more centrifuge facilities.

Unfortunately, the story sort of dissolves when you start to pick it apart.  On the first read, Sanger and Broad seem to suggest that the United States has completed a new intelligence assessment and that US officials are rolling it out.

On closer inspection, however, the story is based on a pair of public statements — an extemporaneous comment by Gary Samore, answering audience questions after a presentation on Iran before the Foundation for the Defense of Democracies and Ambassador Glyn Davies’s statement before the IAEA Board of Governors — and a cryptic remark by an anonymous South Korean intelligence official in the Chosun Ilbo.

The statements don’t appear all that “carefully worded” to me.  And, as far as I can tell, they are not part of a coordinated effort.  Still, it’s an interesting question.  How many enrichment-related sites does North Korea have?

Gary Samore on 12/08

Gary Samore spoke at a Foundation for the Defense of Democracies event on Countering the Iranian Threat.  Although his talk dealt exclusively with Iran, he was asked a question about North Korea.  Gary downplayed any connection between Iran and North Korea — which, by the way, was something David Sanger mentioned in a November 21 story.

QUESTION: Two-part question. Can you rule out that the uranium facility that Sig Hecker reported on in North Korea is in any way connected with Iran? Could you rule that out?

And if you have no concrete information, could you tell us a little bit — give us some of your thoughts on this apparent business of nuclear outsourcing that was going on with the Syrian reactor facility, the discussion of what might be happening in Venezuela and so forth?

Thank you.

SAMORE: Well, on the first question, I would say that the — we can’t confirm, of course what Sig Hecker saw. We’re just going by what he said.

Assuming that what he saw and what the North Koreans told him is accurate, then there’s a very big discrepancy between the North Korean program, which appears to be much more advanced and efficient, and the North — and — and the program in Iran, which appears — which is — it’s a different technology and appears to be having some pretty significant technical problems. So that would suggest no connection.

But in response to your second question, I’m very concerned about the risk of North Korea transferring technology or even nuclear materials.

And of course we have examples in the past where they have done that, you know, apparently providing some nuclear material to Libya, certainly helping the Syrians build a reactor which was destroyed by Israel in 2007.

So I think in the future one of the most important elements of our diplomacy with North Korea and with the other countries in the six-party talks has to be to ensure that North Korea does not sell or transfer nuclear technology or materials to countries in the Middle East.

Because I think that could fundamentally change the pace of the nuclear clock that I talked about, and to the extent that that — in the case in Iran — to the extent that that clock has accelerated, then we lose time for this dual-track strategy that we’re pursuing, which will take months, you know, if it’s going to be successful.

Gary’s response doesn’t seem planned — he asserts that the technology is different when, in fact, the North Koreans appear to be spinning Pakistan-derived P2s, which Iran is also developing.

From the context, it seems clear that Gary is just trying to explain that there is no reason to assume Iran is the culprit — remember, that he is dealing with a slightly disturbed audience.  People who say things like “Just to throw this in at the very last, what about shifting the focus? Cut off the head of the snake in North Korea, and — which is going through a dicey transition right now, and let the people in Iran look at that.”

That was a real comment, by the way. Gary starts his talk by awkwardly joking that he caught the “tail end of the last panel discussion” — when the “head of the snake” comment was made –  “which I thought was very high-spirited.” Gary obviously was aware that he was dealing with a room of lunatics. (The full-text of the Q and A is in the comments.)  I don’t blame him for not executing a letter perfect dissertation on the design heritage of North Korea’s centrifuges.  He was just hoping to escape the mob with the coiffure intact.

Glyn Davies on 12/02

Glyn Davies, the US Ambassador to the IAEA, on the other hand, did make a statement about North Korea’s nuclear program at the IAEA Board of Governors meeting (along with statements about Iran and the assurance of nuclear fuel supply).  The entire thing is worth a read, but here is the graf that interests us:

Based on the apparent scale of the facility, the fast progress the DPRK has made in its construction, and the evidence going back years that North Korea has tried to procure enrichment-related material from abroad, it is likely that North Korea had been pursuing an enrichment capability long before the April 2009 date it now claims.  If so, there is a clear likelihood that DPRK has built other uranium enrichment-related facilities in its territory.

Fair enough.

Anonymous South Korean Intelligence Official on 12/14

Finally, Broad and Sanger refer to South Korean officials, but don’t actually quote any.  Dan Dombey in the Financial Times states that “An unnamed South Korean official told Chosun Ilbo newspaper that Seoul and Washington believed North Korea had ‘three or four’ enrichment sites, not including Yongbyon.”

Actually, that’s not precisely what the anonymous official said.  Here is the quote:

“Yongbyon was not included in the list of three or four locations that Seoul and Washington had previously suspected,” a South Korean intelligence official said Monday. “We understand that the North has long been conducting a uranium enrichment experiment somewhere else.”

If I understand that correctly, the official is saying that there is a list of three or four suspect sites other than Yongbyon.  It isn’t clear to me (can someone check the original Korean?) whether this means there are at least three or four other sites, or whether there are three or four candidates for an R&D site that preceded Yongbyon.  (The story identifies the sites as “a research institute in downtown Pyongyang and a missile base in Yongjori, Yanggang Province, as well as a cave complex in Kumchangri 160 km north of Pyongyang.”)

The South Korean Foreign Minister declined to comment on the intelligence claim — which Reuters characterized as “three to four plants to enrich uranium”– but it seems plausible enough that we had a list of three or four suspect locations.

What Do We Actually Know?

Not very much, it seems.

The long-running issue for the US intelligence community is that the US has never, before Yongbyon, been able to identify a “bricks and mortar” site for centrifuge enrichment.  We knew that the North Koreans were pursuing enrichment, with help from Pakistan, but there were substantial debates about the scope.  Was it a little cheating (a “footnote” to the plutonium program) or a massive, parallel effort?  Size, it appears, matters.

Part of the problem is that gas centrifuges, as I argued at the Wilson Center, are almost ideal for a clandestine nuclear-weapons program.  Centrifuge facilities have virtually no signature.  Policymakers have an understandable tendency to overreact to suggestive procurement data, because that’s all the warning they may ever get. Or, as I put it in the Bulletin, it wasn’t an accident that the mythical Iraqi nuclear weapons program was thought to be based on centrifuges.

The assertion is that North Korea maintains a “sophisticated network of other, secret sites — and perhaps a fully operating uranium enrichment plant — elsewhere in the country” sounds ominous, but it is actually somewhat vague.

As far as I can tell, our analysts still staring at adits in North Korea wondering if there be centrifuges inside.

I take the phrase “network of other, secret sites” to mean what must be several centrifuge workshops (Iran declared 13 to the IAEA) and at least an R&D site (sort of the equivalent of the Pilot FEP at Natanz in Iran) where North Korea could fool around with the two-dozen or so P1 and P2 centrifuges provided by Pakistan, as well as the first cascade or two of North Korean assembled machines.

But that isn’t what we really want to know.  What we want to know is whether or not there is a North Korean equivalent of Qom. Or, rather, three or four Qoms, busy churning out highly enriched uranium for the juche bomb.

On that issue, the Broad and Sanger story suggests that the United States still does not know — “perhaps” another fully operating uranium enrichment plant.  Perhaps not. Sig Hecker suggests that it is “highly likely that a parallel covert facility capable of HEU production exists elsewhere in the country.”  I don’t see any reason to disagree with that assessment.  But where the hell is it?

This is a real problem.  We can’t go looking in every mountainous nook and cranny for centrifuges.  As the whole Kumchang-ri fiasco illustrated, this gets very expensive, very fast.  Moreover, if the IC really missed the construction of a facility, the ritual covering of asses will begin.  In a situation like this, the incentive — if you have to pick a number out of thin air — is to overestimate. If the over/under is 3.5 , go ahead and take 4.

That seems to be what is happening now. But that, of course, is how we ended up destructive fantasies about Iraqi and North Korean centrifuge facilities, fantasies that we are still paying for today.

So what do we do?  The only experience we have with disarming a state that went all the way with the high-tech uranium route is South Africa with jet-nozzle technology.  Although the international community has decided we are satisfied with South Africa’s disarmament, a lot of that confidence is judgment.  The math, as Steve Fetter has pointed out, is still pretty fuzzy.

So, at least this evening, I don’t have any bright ideas.  But I would make the argument, again, that simply accepting worst-case assessments is as dangerous as accepting Pollyannish ones.  The right approach, though not always easy, is to try to construct a policy that is robust to uncertainty.