By all accounts, the US has succeeded in getting North Korea to agree to inspections at undeclared facilities as part of the verification process—another impressive diplomatic victory for Chris Hill. There’s little prospect the US will try to use this provision imminently as it appears to have agreed to a “plutonium first” approach. But, when (or rather if) we ever get to the point of worrying about whether North Korea has an enrichment programme the US hopes inspections at undeclared sites will be useful in proving or disproving its existence.
The problem is (as I’ve noted in passing before) it may very well not turn out to be as simple as that.
Let’s assume (and I strongly suspect this is correct) that, based on intelligence information, the US is pretty certain North Korea has at least a small clandestine gas centrifuge enrichment plant (CGEP), but no idea where it might be. I can see one of two scenarios unfolding.
Scenario 1: The US decides to play the inspection lottery
On the basis of sketchy intelligence, the US picks a handful of the literally thousands of buildings in North Korea that could house a GCEP and asks to inspect them. North Korea agrees. The inspectors show up and find nothing (which they knew would happen as soon as North Korea agreed to the inspections).
The US doesn’t alter its opinion about the existence of an enrichment programme. It argues that the GCEP is located somewhere else (an entirely logical conclusion based on the available information). North Korea, China and Russia have little sympathy and tell the US that it has had its inspections and found nothing so can’t it just get over the whole centrifuge issue already. The US has little success in explaining probability theory to an audience that has no interest in hearing it and everyone gets very cross with one another.
Sound implausible?
Well, it’s pretty much happened before. In 1998/99 at Kumchang-ri. Interestingly, when you google “Kumchang Korea” the top hit is an article of Jeffrey’s from 2005 where he recounts the story for a different purpose (this was in his less temperate days when his posts had titles like “David Sanger: Two Time Loser on Kilju and Kumchang-ri?”).
The post contains a nice article by Daniel Sneider of the San Jose Mercury-News summarizing what happened:
In 1998, satellite images and other intelligence gathered at a site called Kumchang-ri, near North Korea’s border with China, seemed to indicate that the North Koreans were building a secret nuclear reactor and reprocessing facility to replace ones then under international control. For months, intelligence agencies followed developments at the site, observing tunnels being dug, watching concrete being poured, looking for air shafts and cooling ponds for evidence it was a nuclear facility.
…
It [The Clinton administration] demanded access to the site, going to the brink of renewed confrontation.
For a time, the North Koreans pushed the confrontation. “They recognized it was a useful thing to have us spun up for a period of time,’’ the former analyst said. Finally, they agreed to allow access.
Two visits by American inspectors, using sophisticated technology, revealed that while this was a sensitive defense facility of an undetermined nature, “there was no way that it was nuclear,’’ said Pritchard, a conclusion he said was reaffirmed in a 2003 review of the incident.
Bottom line: Visits to undeclared sites are only as effective as the intelligence to guide them.
Scenario 2: The US decides not to play the inspection lottery
Fearful of scenario 1 and a replay of the Kumchang-ri embarrassment, the US still accuses North Korea of harbouring a clandestine GCEP, but doesn’t ask for any unannounced inspections. North Korea, China and Russia tell the US that it should put up or shut up (i.e. use the inspection provisions at undeclared facilities it insisted upon or stop making the accusations). The US doesn’t take the bait and, again, the tension created does no favours to the process.
In short, having negotiated inspection provisions at undeclared facilities the US is in a potentially awkward position whether it uses them or not, unless it knows where to look. And that is a big “unless”. North Korea, as they say, is a hard intelligence target.
“By all accounts, the US has succeeded in getting North Korea to agree to inspections at undeclared facilities as part of the verification process—another impressive diplomatic victory for Chris Hill.”
Why do you state this? Hill got the Norks to agree on inspections of sites outside the declared facilities only on the basis of mutual consent — a provision that certainly must wait the test of time before we can assert that it is a done deed. Most see the provision as papering over the dispute and leaving non-Yongbyon facilities for a later time.
A couple observations:
1) there will not be any verification of the NORKs except by mutual consent; and
2) an important component of what Hill obtained was the participation of the IAEA: the Agency can and often does cool down the tempers that inevitably rise from inspections which are a compromise of national sovereignty, whether you are North Korea, Iran, or Spain.