Body
As the North Korea community ponders the likelihood of another satellite launch, David Albright has issued a timely new report estimating North Korea’s nuclear capabilities. The opacity of the system and the absence of an on-the-ground presence since the ejection of inspectors in 2008 make any such effort fraught. But Albright does as careful a job as is probably possible—providing ranges and hypothesized distributions—and reminds us that the capability continues to expand.
The report can be summarized by combining several of his tables into the one below.
As is typical for such estimates, there are three headline numbers: total fissile material; the nuclear weapons equivalent of that fissile material; and a total device/weapon estimate that includes a discount factor for the fact that all fissile material is not typically weaponized (for example, Albright estimates that 70 percent of plutonium has gone into actual devices, which seems high although certainly plausible). The estimates of current stocks of fissile material have to make a judgment on the early 1990s discrepancies between North Korean and IAEA claims and estimate the material used during the nuclear tests in 2006, 2009, and 2013. (For a comprehensive timeline of North Korea’s nuclear program, see here). Albright and his team recognize that point predictions are not that helpful so he provides qualitative confidence intervals for some data (plutonium) and distributions (such as the one below, generated by Crystal Ball software) and standard deviations of others (those marked above as reporting the medians).
On plutonium, the ultimate source is the declaration provided by the North Koreans following the decision to freeze the program in 2007 and estimates on what has been extracted and used since. The numbers in the 2008 declaration are not in the public domain, but Albright has talked to enough of the principles to converge on an estimate of 30-34 kilograms, with a high estimate of another 8-10 kilos (but with low confidence). Some additional plutonium was probably separated after the collapse of the Six Party Talks in 2008, but a substantial portion of this was probably used up in the 2009 and 2013 tests. Although the reactor was restarted and has been running periodically again since 2013, it has not been operational long enough to generate additional plutonium in that time period.
Estimates with respect to HEU are looser for the simple reason that we don’t know how many sites have spinning cascades. Thanks to the visit of a Stanford team in 2010 we know with reasonably high certainty that there is enrichment going on at Yongbyon. But the size of the facility sparked strong suspicion that there might be another site, typically thought of as a pilot plant which could be smaller, equivalent to or even larger than Yongbyon. The result is not only widely divergent estimates, but wider confidence intervals around each.
With respect to the weapons equivalent estimates, Albright assumes 2-5 kilos for a plutonium weapon. Again, estimates with respect to devices using weapons-grade uranium (WGU) show wider confidence intervals because of the possibility that WGU could be combined with plutonium or used on its own. In the latter case, the amount of WGU needed could be as high as 27.5 kilos; again, this is the value of providing the distributions given that this number could also be as low as 15 kilos.
What ultimately matters is not whether North Korea has 10 or 15 nuclear weapons but their capacity to deploy them. Albright argues that it is likely that the North Koreans can already build a warhead which can fit atop a Nodong missile with a range of less than 800 miles, but the reliability of such a product is still dubious. As ranges go up, reliability goes down. In fact, in an interesting post at 38North, John Shilling argues that the satellites may be just satellites, and not necessarily a stalking horse for an ICBM capability.
This short summary can’t do justice to the care that Albright puts into walking through the background to the program and the assumptions that undergird the analysis; this is a really useful report not only for the estimates but as a tour of the entire North Korean nuclear effort. In the end, the findings are about what we would expect: that the country had the fissile material to produce a dozen or so devices; it is continuing to work on both routes to fissile material; and is simultaneously working to mount a device on a warhead, which could well be a more substantial constraint than the devices themselves. The bigger uncertainties are not in these estimates of material capabilities, but what North Korea thinks it is buying with the program. A deterrent? A tool for coercive bargaining? A prestige project that bolsters the legitimacy of the Kim dynasty? But one thing is certain: given the stated commitment to maintain their program in the April 2013 byungjin line, the farther down the road they go, the more difficult it will be to roll it back.
